You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cxf.apache.org by Sergey Beryozkin <sb...@gmail.com> on 2011/02/04 18:55:43 UTC
LogBrowser project is on the trunk
Hi all,
I have just moved the LogBrowser project Tomasz was working upon as part of
GSOC 2010 to the trunk. I've done few minor changes before moving it in.
- rt/management-web has a new package,
org.apache.cxf.management.web.browser* which contains the code for the
browser. It was org.apache.cxf.management.web.logging.logbrowser* in the
sandbox but I thought we might enhance it in the future to work
PersistingIn/Out interceptors in the rt/management so decided to simplify
the package name.
- samples/logbrowser is the demo showing the LogBrowser in action.
It may make sense even move the whole rt/management-web into a CXF
subproject, we'll see. For now, I'd like to have the browser polished and
then we can decide.
The GWT compilation will happen as part of building the rt/management-web,
so the final jar will actually have the generated GWT code and resources
included. We may consider doing it during the demo run, but again, lets try
to polish the browser first.
Tomasz, first of all congrats, I think it is good that your project is
moving to the trunk.
I'd appreciate if you could help me a bit with the initial clean-up, hope
you'll find some time. You are a CXF committer now so you can do changes on
the trunk...
Here is the initial list of issues which I need some help with :
1. LogBrowser has a showstopper bug which I'm sure is due to the fact we
just did not look at it for a while :-). Basically, when I go and generate
the logs, and then click 'refresh' on the endpoint, the browser reports an
exception.
The AtomPullServer is functioning ok, Firefox subscribes to the Atom Feed
with no probs.
2. Please move ManageEndpoints button either immediately above or below the
Filter button
3. Remove the Tasks and Endpoints buttons/entries, lets have it the way you
did it originally. We only need to see the list of endpoints which will be
added via "ManageEndpoints", the explorer style is just too complex.
4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel and
overlaps with the "Settings" entry,
5. In ManageEndpoints please change "back to browse" to "Back to Explorer"
6. We will need to think what to do with the jscript library added to the
demo - we will have to link to its via remore URI or simple do a simpler
plain HTML form - not a priority right now.
Hope it's not a big list :-). We will catch up on the GWT dev a bit and I
also hope other users with the GWT experience contribute in time. What I
will do before 2.4 is look at the browser code and see what I can refactor
but also add some helpers to the AtomPullServer so that users can use it for
reading from the existing file logs...
Thanks, Sergey
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
I had to disable the compilation of the browser subpackage for now due to
Hudson compoling with Java 1.5 and the subpackage requiring 1.6.
Tomasz - when you get a chance, please re-enable the compilation in your
local snapshot and work with 1.6 - once we have things polished then we will
probably need to it all to the sub-project
thanks, Sergey
On Fri, Feb 4, 2011 at 5:55 PM, Sergey Beryozkin <sb...@gmail.com>wrote:
> Hi all,
>
> I have just moved the LogBrowser project Tomasz was working upon as part of
> GSOC 2010 to the trunk. I've done few minor changes before moving it in.
>
> - rt/management-web has a new package,
> org.apache.cxf.management.web.browser* which contains the code for the
> browser. It was org.apache.cxf.management.web.logging.logbrowser* in the
> sandbox but I thought we might enhance it in the future to work
> PersistingIn/Out interceptors in the rt/management so decided to simplify
> the package name.
>
> - samples/logbrowser is the demo showing the LogBrowser in action.
>
> It may make sense even move the whole rt/management-web into a CXF
> subproject, we'll see. For now, I'd like to have the browser polished and
> then we can decide.
>
> The GWT compilation will happen as part of building the rt/management-web,
> so the final jar will actually have the generated GWT code and resources
> included. We may consider doing it during the demo run, but again, lets try
> to polish the browser first.
>
> Tomasz, first of all congrats, I think it is good that your project is
> moving to the trunk.
> I'd appreciate if you could help me a bit with the initial clean-up, hope
> you'll find some time. You are a CXF committer now so you can do changes on
> the trunk...
>
> Here is the initial list of issues which I need some help with :
>
> 1. LogBrowser has a showstopper bug which I'm sure is due to the fact we
> just did not look at it for a while :-). Basically, when I go and generate
> the logs, and then click 'refresh' on the endpoint, the browser reports an
> exception.
> The AtomPullServer is functioning ok, Firefox subscribes to the Atom Feed
> with no probs.
>
> 2. Please move ManageEndpoints button either immediately above or below the
> Filter button
>
> 3. Remove the Tasks and Endpoints buttons/entries, lets have it the way you
> did it originally. We only need to see the list of endpoints which will be
> added via "ManageEndpoints", the explorer style is just too complex.
>
> 4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel and
> overlaps with the "Settings" entry,
>
> 5. In ManageEndpoints please change "back to browse" to "Back to Explorer"
>
>
> 6. We will need to think what to do with the jscript library added to the
> demo - we will have to link to its via remore URI or simple do a simpler
> plain HTML form - not a priority right now.
>
> Hope it's not a big list :-). We will catch up on the GWT dev a bit and I
> also hope other users with the GWT experience contribute in time. What I
> will do before 2.4 is look at the browser code and see what I can refactor
> but also add some helpers to the AtomPullServer so that users can use it for
> reading from the existing file logs...
>
> Thanks, Sergey
>
>
>
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Forgot one thing, the GWT compiler generates this warning:
[INFO] [WARN] g:DecoratedTabPanel is deprecated. Use the
g:TabLayoutPanel instead.
this is just FYI, not critical at all, for now those 2 issues I described
below are a priority...
thanks, Sergey
On Wed, Mar 9, 2011 at 1:28 PM, Sergey Beryozkin <sb...@gmail.com>wrote:
> Hi Tomasz
>
> I'm finishing some work for AtomPullServer be able to connect to the
> existing file logs, I saw you committing some relevant updates to the
> LogBrowser code too, thanks. I'll will experiment with the Filter dialog
> shortly.
>
> What I'd really appreciate if you could help me with the following, hope it
> will take just a fraction of your time:
>
> - Remove the Authentication/SignIn/SignOut related code/resources. We just
> don't have time now to deal with the authentication/security issues and we'd
> need to have another discussion later on about it. I believe you agreed to
> it too.
> - Explain a bit which code/script deals with processing the dates. I know
> you spent a lot of time on trying to reproduce the application exception,
> including testing it on Ubuntu, but I'd like to give it a try and see if I
> can spot something that causes the issue on my particular OS - if I can
> trace it to some bad Java or Firefox installation then I'll get relaxed
> :-)...
>
> Thanks, Sergey
>
>
>
> On Sun, Feb 27, 2011 at 6:20 PM, Sergey Beryozkin <sb...@gmail.com>wrote:
>
>> String text = columnDefinition.getContent(entry);
>>
>> in SelectableTable.setData
>>
>> is causing it,
>>
>> and it's only happening when the Date column is filled in, the Level
>> and Message are fine.
>>
>> Looks like the code co0nverting the Atom/http-format date is brittle a bit
>>
>> cheers, Sergey
>>
>> On Sun, Feb 27, 2011 at 5:10 PM, Sergey Beryozkin <sb...@gmail.com>
>> wrote:
>> > Hi Tomasz
>> >
>> > I've spent some time and it is some sort of the asynchronous exception
>> > which when captured shows the Date of the log entry. It is thrown when
>> > the ViewerViewImpl.setEntries(List<Entry> entries) is called. We'll
>> > track it down soon :-)
>> >
>> > Sergey
>> >
>> > On Sat, Feb 26, 2011 at 1:23 PM, Tomasz Oponowicz
>> > <to...@gmail.com> wrote:
>> >> Hi Sergey,
>> >>
>> >> On Fri, Feb 25, 2011 at 1:19 PM, Sergey Beryozkin <
>> sberyozkin@gmail.com> wrote:
>> >>> Hi Tomasz
>> >>>
>> >>> I've rerun the demo.
>> >>> First of all, the changes you've made recently have definitely made it
>> >>> look much better, thanks.
>> >>> Unfortunately, I'm still hitting this GWT ApplicationException when
>> >>> refreshing the endpoint :
>> >>> "
>> >>> Application Error
>> >>> Class$jcb
>> >>> 2011-02-25T11:46:55.078Z
>> >>> "
>> >>>
>> >>> Can this message, particularly Class$jcb, help somehow to identify the
>> >>> problem ? It does look like it's a platform/browser specific issue,
>> >>> I'm on Ubuntu 9, Firefox 3.6.13, but it would be good to get rid of it
>> >>> somehow.
>> >>> Just tried Chrome and it showed the same error but with the
>> >>> "Class$kbc" - it's probably the some gwt proxy...
>> >>>
>> >>> Can you please point to the code in the logbrowser project where the
>> >>> response from the remote endpoint is processed ? I will investigate...
>> >>
>> >> I will install similar environment as a virtual machine - I hope I
>> >> will reproduce this issue...
>> >>
>> >>> Few more comments. I agree, the way Tasks and Endpoints are currently
>> >>> shown is nice.
>> >>>
>> >>> - Can you please consider having both ManageEndpoints and Filter links
>> >>> located under the "Tasks" ? And have Endpoints shown first, with the
>> >>> "Tasks" in the bottom of the pane ? Ultimately, the user just wants to
>> >>> see the list of endpoints. Creating/deleting the endpoints and
>> >>> applying filters is critical but I'd just prefer the "Tasks" not to
>> >>> feature as the main activity of the LogBrowser users...
>> >>> - Filter dialog can not be closed at the moment, after it has been
>> opened...
>> >>> - Probably makes sense to rename "Explore" to "Explorer" given that
>> >>> the "ManageEndpoints" panel has the "Back to Explorer" link...
>> >>>
>> >>> More comments inline...
>> >>>
>> >>>
>> >>>>> - Authentication: I've noticed there's AuthenticationRequired
>> >>>>> annotation attached to some of the BootstrapStorage methods - we
>> >>>>> really need to remove this annotation and for now just pop-up a
>> login
>> >>>>> window on the browser start-up.
>> >>>>> Users will be configuring the LogBrowser application as part of the
>> >>>>> real deployments. So what would be good is to write the GWT client
>> >>>>> code such that it only pops up a window when the initial GET
>> returns
>> >>>>> 401 - can you use CXF WebClient there and do 'Response r =
>> >>>>> webClient.get()' and if r.getStatus() == 401 then pop-up a login
>> >>>>> dialog ? We can deal with this issue later, when we have more time,
>> >>>>> and then we'll also decide whether to support https in cases when
>> the
>> >>>>> authentication is needed or may be do the UT profile, we'll see...
>> >>>>
>> >>>> According to your list of tasks please consider also fallowing tasks:
>> >>>
>> >>> Thanks for this analysis...
>> >>>
>> >>>>
>> >>>> - Removing "Sign in" feature;
>> >>>> - Pros:
>> >>>> - Simplify implementation;
>> >>>> - Easy configuration for end user;
>> >>>> - Every company has got their own internal user
>> >>>> authentication system (LDAP, OpenID, internal SSO etc.);
>> >>>> - Even if LogBrowser doesn't contain any user authentication
>> >>>> system, it's still very easy to add integration with some
>> >>>> authentication system:
>> >>>> - Simply interceptor before request rich controller;
>> >>>> - Apache directives (of course if user use Apache
>> >>>> before Tomcat);
>> >>>> - Cons:
>> >>>> - I understand that feeds should be secured, but I think we
>> >>>> should rather concentrate on:
>> >>>> - HTTPS connection;
>> >>>> - password per feed (optional);
>> >>>>
>> >>>
>> >>> I think we are in agreement here. I'd like to propose:
>> >>> - remove the initial Sign-In dialog altogether
>> >>> - Remove SighIn and SignOut buttons
>> >>> - Remove AuthenticationRequired annotation from the coode
>> >>> - postpone dealing with the authentication issues - at the moment we
>> >>> just need to focus on making sure
>> >>> the browser is operational.
>> >>>
>> >>> After 2.4.0, we can enhance it for the authentication+HTTPS be
>> supported.
>> >>>
>> >>> If you agree then please remove all the authentication-related
>> code/settings...
>> >>>
>> >>>> - Removing storing user settings remotely on the servers;
>> >>>> - Pros:
>> >>>> - Simplify implementation;
>> >>>> - Easy configuration for end user;
>> >>>> - Very clear message - all settings are stored in browser
>> >>>> local storage. At the moment the logic it's to complicated. It depend
>> >>>> on situation we keep settings in memory, browser local storage or
>> >>>> bootstrap settings;
>> >>>> - Cons:
>> >>>> - When end user clear cache all settings will be removed;
>> >>>> - Settings are stored per browser. When you add something in
>> >>>> Firefox it won't be available in Chrome;
>> >>>>
>> >>>
>> >>> I think it makes sense to keep the list of endpoints and the filter
>> >>> properties the current user has created.
>> >>> *But*, these settings just need to shared across multiple restarts of
>> >>> the browser between all the users.
>> >>> This is because I don't really think it is realistic that one user
>> >>> will want to see EndpointA only, the other one, EndpointB, etc.
>> >>> So lets keep it - I'm not worried about many users using different
>> >>> browsers for checking the logs of the single server :-)
>> >>>
>> >>>> What do you think about these tasks? I'd like to keep LogBrowser
>> minimalistic.
>> >>>>
>> >>> makes sense :-)
>> >>>
>> >>> Cheers, Sergey
>> >>>
>> >>>> --
>> >>>> Best regards,
>> >>>> Tomasz Oponowicz
>> >>>>
>> >>>
>> >>
>> >> --
>> >> Best regards,
>> >> Tomasz Oponowicz
>> >>
>> >
>>
>
>
>
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz
On Thu, Apr 21, 2011 at 12:53 PM, Tomasz Oponowicz
<to...@gmail.com> wrote:
> Hi Sergey,
>
> Fixed.
>
> The schema of log entry has been changed. From 'eventTimestamp' to 'date'.
Sorry, that was my fault I did not think it would affect the browser.
I did update that property name so that
it was easier to express queries with dates:
/logs?_s=date=le=2011-03-04
is a bit simpler than
/logs?_s=eventTimestamp=le=2011-03-04
> Later I will prepare unit test for checking contract between client and server.
>
Please, whenever you get a chance. Using alternate feed/entries will
simplify this issue because a pure HTML will be returned.
However, GWT does give us more flexibility in the way individual
entries are displayed, etc, so lets stick to using GWT for now :-)
thanks, Sergey
> On Wed, Apr 20, 2011 at 11:47 AM, Sergey Beryozkin <sb...@gmail.com> wrote:
>> Hi Tomasz -
>>
>> thanks for looking into it - yes, lets try to get it back to normal.
>> Lets see if we can realistically work with GWT - effectively we need
>> to decide at some point if GWT is what will be used for other similar
>> tasks, for getting the log files downloaded, for having another tab
>> showing the list of managed components, etc.
>>
>> Cheers, Sergey
>>
>> On Wed, Apr 20, 2011 at 10:37 AM, Tomasz Oponowicz
>> <to...@gmail.com> wrote:
>>> Hi,
>>>
>>> I've tried to create couple of screenshots but I figured out that
>>> nothing is working:
>>>
>>> - exceptions in sample project;
>>> - exceptions in development mode;
>>>
>>> I'm fixing this.
>>>
>>> On Mon, Apr 18, 2011 at 11:31 AM, Sergey Beryozkin <sb...@gmail.com> wrote:
>>>> Hi Tomasz
>>>>
>>>> On Sun, Apr 17, 2011 at 9:06 AM, Tomasz Oponowicz
>>>> <to...@gmail.com> wrote:
>>>>>
>>>>> Hi Sergey,
>>>>>
>>>>> On Wed, Apr 13, 2011 at 7:17 PM, Sergey Beryozkin <sb...@gmail.com>
>>>>> wrote:
>>>>> > I've finally managed to spend a day today working on Windows, and I did
>>>>> > get
>>>>> > the exception when using IE too, though, IE itself does subscribe to the
>>>>> > log
>>>>> > feed without problems.
>>>>>
>>>>> Grrrr.... I will investigate this one. Can you provide stack trace and
>>>>> other details (use case, browser version etc.) ?
>>>>>
>>>>
>>>> The exception was something to do with "null", I'll need to run the browser
>>>> in the dev mode to get more details.
>>>> It's Windows XP/IE7
>>>>
>>>>>
>>>>> > There's definitely something unstable there which is a pity, it's been a
>>>>> > great effort from Tomasz and with the ability to search the logs, it's a
>>>>> > feature to miss, it could've been used nicely in conjunction with the
>>>>> > the
>>>>> > push logging, i.e. critical events are reported asap and then an admin
>>>>> > can
>>>>> > do a more through analysis of the logs, rather than having all log MBs
>>>>> > sent
>>>>> > across the network
>>>>> >
>>>>> > But it does look it works for some people, so we just need to document a
>>>>> > bit
>>>>> > how to use it as a Maven dependency, untill we figure out what is
>>>>> > happening,
>>>>> >
>>>>> > Tomasz, can you give me a favor please and update the LogBrowser page
>>>>> > with few lastest images as I can't get them myself ? I will add some
>>>>> > info to
>>>>> > the Migration Guide (Maven deps, how to set things up, etc).
>>>>>
>>>>> Good point. I will update it.
>>>>>
>>>>
>>>>
>>>> Please - a couple of snapshots, with some description on how to add a new
>>>> endpoint, will help.
>>>>
>>>>
>>>>>
>>>>> > One thing I'm thinking about is we basically get rid of GWT code dealing
>>>>> > with showing the lists of log entries (top-right panel) and just use
>>>>> > HTML
>>>>> > returned from the log feed. GWT will only be used to offer the list of
>>>>> > endpoints, filters, etc, and may be in time we can just replace GWT with
>>>>> > something lighter/simpler.
>>>>>
>>>>> If you decide that GWT is too big overweight for such a simple task, I
>>>>> can rewrite this using Javascript/JQuery/HTML (but firstly let's fix
>>>>> this bug).
>>>>
>>>>
>>>> OK. I'll try to get more details. And may be you can reproduce something
>>>>>
>>>>> > Tomasz, AtomPullServer has getAlternateFeed and getAlternateEntry
>>>>> > methods
>>>>> > and they serve HTML as opposed to XML feeds.
>>>>> > RSSBandit reader on Windows is very good in that it shows alternate
>>>>> > feeds
>>>>> > and entries properly.
>>>>> > I think GWT should have some browser support, so if you could make the
>>>>> > top/right panel a browser window which will only show the alternate
>>>>> > content,
>>>>> > as part of the frame, such that whenever log entry links are clicked,
>>>>> > the
>>>>> > output is directed to the right-bottom pane..
>>>>> > Do you think it will make sense ?
>>>>>
>>>>> Firstly I will try to fix the bug in the normal way but if there won't
>>>>> be any other solution I will replace current implementation with
>>>>> "getAlternateFeed" and "getAlternateEntry".
>>>>>
>>>>> I will update documentation and fix the bug on Monday.
>>>>>
>>>>
>>>> Great, please update the docs if you can.
>>>>
>>>> Thanks, Sergey
>>>>
>>>>>
>>>>> Thanks!
>>>>
>>>>
>>>
>>>
>>>
>>> --
>>> Best regards,
>>> Tomasz Oponowicz
>>>
>>
>>
>>
>> --
>> Sergey Beryozkin
>>
>> Application Integration Division of Talend
>> http://sberyozkin.blogspot.com
>>
>
>
>
> --
> Best regards,
> Tomasz Oponowicz
>
--
Sergey Beryozkin
Application Integration Division of Talend
http://sberyozkin.blogspot.com
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi Sergey,
Fixed.
The schema of log entry has been changed. From 'eventTimestamp' to 'date'.
Later I will prepare unit test for checking contract between client and server.
On Wed, Apr 20, 2011 at 11:47 AM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi Tomasz -
>
> thanks for looking into it - yes, lets try to get it back to normal.
> Lets see if we can realistically work with GWT - effectively we need
> to decide at some point if GWT is what will be used for other similar
> tasks, for getting the log files downloaded, for having another tab
> showing the list of managed components, etc.
>
> Cheers, Sergey
>
> On Wed, Apr 20, 2011 at 10:37 AM, Tomasz Oponowicz
> <to...@gmail.com> wrote:
>> Hi,
>>
>> I've tried to create couple of screenshots but I figured out that
>> nothing is working:
>>
>> - exceptions in sample project;
>> - exceptions in development mode;
>>
>> I'm fixing this.
>>
>> On Mon, Apr 18, 2011 at 11:31 AM, Sergey Beryozkin <sb...@gmail.com> wrote:
>>> Hi Tomasz
>>>
>>> On Sun, Apr 17, 2011 at 9:06 AM, Tomasz Oponowicz
>>> <to...@gmail.com> wrote:
>>>>
>>>> Hi Sergey,
>>>>
>>>> On Wed, Apr 13, 2011 at 7:17 PM, Sergey Beryozkin <sb...@gmail.com>
>>>> wrote:
>>>> > I've finally managed to spend a day today working on Windows, and I did
>>>> > get
>>>> > the exception when using IE too, though, IE itself does subscribe to the
>>>> > log
>>>> > feed without problems.
>>>>
>>>> Grrrr.... I will investigate this one. Can you provide stack trace and
>>>> other details (use case, browser version etc.) ?
>>>>
>>>
>>> The exception was something to do with "null", I'll need to run the browser
>>> in the dev mode to get more details.
>>> It's Windows XP/IE7
>>>
>>>>
>>>> > There's definitely something unstable there which is a pity, it's been a
>>>> > great effort from Tomasz and with the ability to search the logs, it's a
>>>> > feature to miss, it could've been used nicely in conjunction with the
>>>> > the
>>>> > push logging, i.e. critical events are reported asap and then an admin
>>>> > can
>>>> > do a more through analysis of the logs, rather than having all log MBs
>>>> > sent
>>>> > across the network
>>>> >
>>>> > But it does look it works for some people, so we just need to document a
>>>> > bit
>>>> > how to use it as a Maven dependency, untill we figure out what is
>>>> > happening,
>>>> >
>>>> > Tomasz, can you give me a favor please and update the LogBrowser page
>>>> > with few lastest images as I can't get them myself ? I will add some
>>>> > info to
>>>> > the Migration Guide (Maven deps, how to set things up, etc).
>>>>
>>>> Good point. I will update it.
>>>>
>>>
>>>
>>> Please - a couple of snapshots, with some description on how to add a new
>>> endpoint, will help.
>>>
>>>
>>>>
>>>> > One thing I'm thinking about is we basically get rid of GWT code dealing
>>>> > with showing the lists of log entries (top-right panel) and just use
>>>> > HTML
>>>> > returned from the log feed. GWT will only be used to offer the list of
>>>> > endpoints, filters, etc, and may be in time we can just replace GWT with
>>>> > something lighter/simpler.
>>>>
>>>> If you decide that GWT is too big overweight for such a simple task, I
>>>> can rewrite this using Javascript/JQuery/HTML (but firstly let's fix
>>>> this bug).
>>>
>>>
>>> OK. I'll try to get more details. And may be you can reproduce something
>>>>
>>>> > Tomasz, AtomPullServer has getAlternateFeed and getAlternateEntry
>>>> > methods
>>>> > and they serve HTML as opposed to XML feeds.
>>>> > RSSBandit reader on Windows is very good in that it shows alternate
>>>> > feeds
>>>> > and entries properly.
>>>> > I think GWT should have some browser support, so if you could make the
>>>> > top/right panel a browser window which will only show the alternate
>>>> > content,
>>>> > as part of the frame, such that whenever log entry links are clicked,
>>>> > the
>>>> > output is directed to the right-bottom pane..
>>>> > Do you think it will make sense ?
>>>>
>>>> Firstly I will try to fix the bug in the normal way but if there won't
>>>> be any other solution I will replace current implementation with
>>>> "getAlternateFeed" and "getAlternateEntry".
>>>>
>>>> I will update documentation and fix the bug on Monday.
>>>>
>>>
>>> Great, please update the docs if you can.
>>>
>>> Thanks, Sergey
>>>
>>>>
>>>> Thanks!
>>>
>>>
>>
>>
>>
>> --
>> Best regards,
>> Tomasz Oponowicz
>>
>
>
>
> --
> Sergey Beryozkin
>
> Application Integration Division of Talend
> http://sberyozkin.blogspot.com
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz -
thanks for looking into it - yes, lets try to get it back to normal.
Lets see if we can realistically work with GWT - effectively we need
to decide at some point if GWT is what will be used for other similar
tasks, for getting the log files downloaded, for having another tab
showing the list of managed components, etc.
Cheers, Sergey
On Wed, Apr 20, 2011 at 10:37 AM, Tomasz Oponowicz
<to...@gmail.com> wrote:
> Hi,
>
> I've tried to create couple of screenshots but I figured out that
> nothing is working:
>
> - exceptions in sample project;
> - exceptions in development mode;
>
> I'm fixing this.
>
> On Mon, Apr 18, 2011 at 11:31 AM, Sergey Beryozkin <sb...@gmail.com> wrote:
>> Hi Tomasz
>>
>> On Sun, Apr 17, 2011 at 9:06 AM, Tomasz Oponowicz
>> <to...@gmail.com> wrote:
>>>
>>> Hi Sergey,
>>>
>>> On Wed, Apr 13, 2011 at 7:17 PM, Sergey Beryozkin <sb...@gmail.com>
>>> wrote:
>>> > I've finally managed to spend a day today working on Windows, and I did
>>> > get
>>> > the exception when using IE too, though, IE itself does subscribe to the
>>> > log
>>> > feed without problems.
>>>
>>> Grrrr.... I will investigate this one. Can you provide stack trace and
>>> other details (use case, browser version etc.) ?
>>>
>>
>> The exception was something to do with "null", I'll need to run the browser
>> in the dev mode to get more details.
>> It's Windows XP/IE7
>>
>>>
>>> > There's definitely something unstable there which is a pity, it's been a
>>> > great effort from Tomasz and with the ability to search the logs, it's a
>>> > feature to miss, it could've been used nicely in conjunction with the
>>> > the
>>> > push logging, i.e. critical events are reported asap and then an admin
>>> > can
>>> > do a more through analysis of the logs, rather than having all log MBs
>>> > sent
>>> > across the network
>>> >
>>> > But it does look it works for some people, so we just need to document a
>>> > bit
>>> > how to use it as a Maven dependency, untill we figure out what is
>>> > happening,
>>> >
>>> > Tomasz, can you give me a favor please and update the LogBrowser page
>>> > with few lastest images as I can't get them myself ? I will add some
>>> > info to
>>> > the Migration Guide (Maven deps, how to set things up, etc).
>>>
>>> Good point. I will update it.
>>>
>>
>>
>> Please - a couple of snapshots, with some description on how to add a new
>> endpoint, will help.
>>
>>
>>>
>>> > One thing I'm thinking about is we basically get rid of GWT code dealing
>>> > with showing the lists of log entries (top-right panel) and just use
>>> > HTML
>>> > returned from the log feed. GWT will only be used to offer the list of
>>> > endpoints, filters, etc, and may be in time we can just replace GWT with
>>> > something lighter/simpler.
>>>
>>> If you decide that GWT is too big overweight for such a simple task, I
>>> can rewrite this using Javascript/JQuery/HTML (but firstly let's fix
>>> this bug).
>>
>>
>> OK. I'll try to get more details. And may be you can reproduce something
>>>
>>> > Tomasz, AtomPullServer has getAlternateFeed and getAlternateEntry
>>> > methods
>>> > and they serve HTML as opposed to XML feeds.
>>> > RSSBandit reader on Windows is very good in that it shows alternate
>>> > feeds
>>> > and entries properly.
>>> > I think GWT should have some browser support, so if you could make the
>>> > top/right panel a browser window which will only show the alternate
>>> > content,
>>> > as part of the frame, such that whenever log entry links are clicked,
>>> > the
>>> > output is directed to the right-bottom pane..
>>> > Do you think it will make sense ?
>>>
>>> Firstly I will try to fix the bug in the normal way but if there won't
>>> be any other solution I will replace current implementation with
>>> "getAlternateFeed" and "getAlternateEntry".
>>>
>>> I will update documentation and fix the bug on Monday.
>>>
>>
>> Great, please update the docs if you can.
>>
>> Thanks, Sergey
>>
>>>
>>> Thanks!
>>
>>
>
>
>
> --
> Best regards,
> Tomasz Oponowicz
>
--
Sergey Beryozkin
Application Integration Division of Talend
http://sberyozkin.blogspot.com
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi,
I've tried to create couple of screenshots but I figured out that
nothing is working:
- exceptions in sample project;
- exceptions in development mode;
I'm fixing this.
On Mon, Apr 18, 2011 at 11:31 AM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi Tomasz
>
> On Sun, Apr 17, 2011 at 9:06 AM, Tomasz Oponowicz
> <to...@gmail.com> wrote:
>>
>> Hi Sergey,
>>
>> On Wed, Apr 13, 2011 at 7:17 PM, Sergey Beryozkin <sb...@gmail.com>
>> wrote:
>> > I've finally managed to spend a day today working on Windows, and I did
>> > get
>> > the exception when using IE too, though, IE itself does subscribe to the
>> > log
>> > feed without problems.
>>
>> Grrrr.... I will investigate this one. Can you provide stack trace and
>> other details (use case, browser version etc.) ?
>>
>
> The exception was something to do with "null", I'll need to run the browser
> in the dev mode to get more details.
> It's Windows XP/IE7
>
>>
>> > There's definitely something unstable there which is a pity, it's been a
>> > great effort from Tomasz and with the ability to search the logs, it's a
>> > feature to miss, it could've been used nicely in conjunction with the
>> > the
>> > push logging, i.e. critical events are reported asap and then an admin
>> > can
>> > do a more through analysis of the logs, rather than having all log MBs
>> > sent
>> > across the network
>> >
>> > But it does look it works for some people, so we just need to document a
>> > bit
>> > how to use it as a Maven dependency, untill we figure out what is
>> > happening,
>> >
>> > Tomasz, can you give me a favor please and update the LogBrowser page
>> > with few lastest images as I can't get them myself ? I will add some
>> > info to
>> > the Migration Guide (Maven deps, how to set things up, etc).
>>
>> Good point. I will update it.
>>
>
>
> Please - a couple of snapshots, with some description on how to add a new
> endpoint, will help.
>
>
>>
>> > One thing I'm thinking about is we basically get rid of GWT code dealing
>> > with showing the lists of log entries (top-right panel) and just use
>> > HTML
>> > returned from the log feed. GWT will only be used to offer the list of
>> > endpoints, filters, etc, and may be in time we can just replace GWT with
>> > something lighter/simpler.
>>
>> If you decide that GWT is too big overweight for such a simple task, I
>> can rewrite this using Javascript/JQuery/HTML (but firstly let's fix
>> this bug).
>
>
> OK. I'll try to get more details. And may be you can reproduce something
>>
>> > Tomasz, AtomPullServer has getAlternateFeed and getAlternateEntry
>> > methods
>> > and they serve HTML as opposed to XML feeds.
>> > RSSBandit reader on Windows is very good in that it shows alternate
>> > feeds
>> > and entries properly.
>> > I think GWT should have some browser support, so if you could make the
>> > top/right panel a browser window which will only show the alternate
>> > content,
>> > as part of the frame, such that whenever log entry links are clicked,
>> > the
>> > output is directed to the right-bottom pane..
>> > Do you think it will make sense ?
>>
>> Firstly I will try to fix the bug in the normal way but if there won't
>> be any other solution I will replace current implementation with
>> "getAlternateFeed" and "getAlternateEntry".
>>
>> I will update documentation and fix the bug on Monday.
>>
>
> Great, please update the docs if you can.
>
> Thanks, Sergey
>
>>
>> Thanks!
>
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz
On Sun, Apr 17, 2011 at 9:06 AM, Tomasz Oponowicz <
tomasz.oponowicz@gmail.com> wrote:
> Hi Sergey,
>
> On Wed, Apr 13, 2011 at 7:17 PM, Sergey Beryozkin <sb...@gmail.com>
> wrote:
> > I've finally managed to spend a day today working on Windows, and I did
> get
> > the exception when using IE too, though, IE itself does subscribe to the
> log
> > feed without problems.
>
> Grrrr.... I will investigate this one. Can you provide stack trace and
> other details (use case, browser version etc.) ?
>
>
The exception was something to do with "null", I'll need to run the browser
in the dev mode to get more details.
It's Windows XP/IE7
> > There's definitely something unstable there which is a pity, it's been a
> > great effort from Tomasz and with the ability to search the logs, it's a
> > feature to miss, it could've been used nicely in conjunction with the the
> > push logging, i.e. critical events are reported asap and then an admin
> can
> > do a more through analysis of the logs, rather than having all log MBs
> sent
> > across the network
> >
> > But it does look it works for some people, so we just need to document a
> bit
> > how to use it as a Maven dependency, untill we figure out what is
> happening,
> >
> > Tomasz, can you give me a favor please and update the LogBrowser page
> > with few lastest images as I can't get them myself ? I will add some info
> to
> > the Migration Guide (Maven deps, how to set things up, etc).
>
> Good point. I will update it.
>
>
Please - a couple of snapshots, with some description on how to add a new
endpoint, will help.
> > One thing I'm thinking about is we basically get rid of GWT code dealing
> > with showing the lists of log entries (top-right panel) and just use HTML
> > returned from the log feed. GWT will only be used to offer the list of
> > endpoints, filters, etc, and may be in time we can just replace GWT with
> > something lighter/simpler.
>
> If you decide that GWT is too big overweight for such a simple task, I
> can rewrite this using Javascript/JQuery/HTML (but firstly let's fix
> this bug).
>
OK. I'll try to get more details. And may be you can reproduce something
>
> > Tomasz, AtomPullServer has getAlternateFeed and getAlternateEntry methods
> > and they serve HTML as opposed to XML feeds.
> > RSSBandit reader on Windows is very good in that it shows alternate feeds
> > and entries properly.
> > I think GWT should have some browser support, so if you could make the
> > top/right panel a browser window which will only show the alternate
> content,
> > as part of the frame, such that whenever log entry links are clicked, the
> > output is directed to the right-bottom pane..
> > Do you think it will make sense ?
>
> Firstly I will try to fix the bug in the normal way but if there won't
> be any other solution I will replace current implementation with
> "getAlternateFeed" and "getAlternateEntry".
>
> I will update documentation and fix the bug on Monday.
>
>
Great, please update the docs if you can.
Thanks, Sergey
> Thanks!
>
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi Sergey,
On Wed, Apr 13, 2011 at 7:17 PM, Sergey Beryozkin <sb...@gmail.com> wrote:
> I've finally managed to spend a day today working on Windows, and I did get
> the exception when using IE too, though, IE itself does subscribe to the log
> feed without problems.
Grrrr.... I will investigate this one. Can you provide stack trace and
other details (use case, browser version etc.) ?
> There's definitely something unstable there which is a pity, it's been a
> great effort from Tomasz and with the ability to search the logs, it's a
> feature to miss, it could've been used nicely in conjunction with the the
> push logging, i.e. critical events are reported asap and then an admin can
> do a more through analysis of the logs, rather than having all log MBs sent
> across the network
>
> But it does look it works for some people, so we just need to document a bit
> how to use it as a Maven dependency, untill we figure out what is happening,
>
> Tomasz, can you give me a favor please and update the LogBrowser page
> with few lastest images as I can't get them myself ? I will add some info to
> the Migration Guide (Maven deps, how to set things up, etc).
Good point. I will update it.
> One thing I'm thinking about is we basically get rid of GWT code dealing
> with showing the lists of log entries (top-right panel) and just use HTML
> returned from the log feed. GWT will only be used to offer the list of
> endpoints, filters, etc, and may be in time we can just replace GWT with
> something lighter/simpler.
If you decide that GWT is too big overweight for such a simple task, I
can rewrite this using Javascript/JQuery/HTML (but firstly let's fix
this bug).
> Tomasz, AtomPullServer has getAlternateFeed and getAlternateEntry methods
> and they serve HTML as opposed to XML feeds.
> RSSBandit reader on Windows is very good in that it shows alternate feeds
> and entries properly.
> I think GWT should have some browser support, so if you could make the
> top/right panel a browser window which will only show the alternate content,
> as part of the frame, such that whenever log entry links are clicked, the
> output is directed to the right-bottom pane..
> Do you think it will make sense ?
Firstly I will try to fix the bug in the normal way but if there won't
be any other solution I will replace current implementation with
"getAlternateFeed" and "getAlternateEntry".
I will update documentation and fix the bug on Monday.
Thanks!
> Cheers, Sergey
>
> On Thu, Mar 31, 2011 at 4:18 PM, Sergey Beryozkin <sb...@gmail.com>
> wrote:
>>
>> I've added NOTICE to the rt/management-web only as I'm still seeing all
>> sort of crashes on Ubuntu, it does look like my environment is somehow
>> broken for GWT to work properly as Tomasz did confirm iy was working ok on
>> Ubuntu for him. I''ll try to confirm that it is all working on Windows and
>> after that I'll update the distribution.
>>
>> It would be good if few more people tried a logbrowser demo in the
>> distribution - it is very easy to run...I think if we can get at least one
>> more crash report then we might want to consider 'distributing' the
>> logbrowser as a Maven dependency only. I hope it's my Ubuntu distribution
>> only (old enough already) that is not stable...
>>
>> Cheers, Sergey
>>
>> On Wed, Mar 16, 2011 at 2:21 AM, Daniel Kulp <dk...@apache.org> wrote:
>>>
>>> On Tuesday 15 March 2011 6:25:14 PM Tomasz Oponowicz wrote:
>>> > > Probably one open issue is to figure out what to do with different
>>> > > .js
>>> > > scripts. ? The scripts which are in the rt/management-web module, can
>>> > > they be Apache licensed ? Dan - what do you think ?
>>> >
>>> > There are:
>>> >
>>> > * json2.min.js - Public Domain;
>>> > * uuid.min.js - Dual licensed under the MIT and GPL licenses;
>>>
>>> Both licenses are fine, but we need to note them in the NOTICE file.
>>>
>>> You'll need to create a
>>> src/main/appended-resources/META-INF/NOTICE
>>> file that contains whatever text would be required.
>>>
>>> For example, in the javascript module, the NOTICE there has:
>>>
>>> Portions of the file cxf-utils.js derives from code marked:
>>> This code was written by Tyler Akins and has been placed in the
>>> public domain. It would be nice if you left this header intact.
>>> Base64 code from Tyler Akins -- http://rumkin.com
>>>
>>> You can use that as a guideline.
>>>
>>> You'll also need to update
>>> distribution/src/main/appended-resources/META-INF/NOTICE
>>> with the same info for the release assemplies to pick up.
>>>
>>>
>>> --
>>> Daniel Kulp
>>> dkulp@apache.org
>>> http://dankulp.com/blog
>>> Talend - http://www.talend.com
>>
>>
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
I've finally managed to spend a day today working on Windows, and I did get
the exception when using IE too, though, IE itself does subscribe to the log
feed without problems.
There's definitely something unstable there which is a pity, it's been a
great effort from Tomasz and with the ability to search the logs, it's a
feature to miss, it could've been used nicely in conjunction with the the
push logging, i.e. critical events are reported asap and then an admin can
do a more through analysis of the logs, rather than having all log MBs sent
across the network
But it does look it works for some people, so we just need to document a bit
how to use it as a Maven dependency, untill we figure out what is happening,
Tomasz, can you give me a favor please and update the LogBrowser page
with few lastest images as I can't get them myself ? I will add some info to
the Migration Guide (Maven deps, how to set things up, etc).
One thing I'm thinking about is we basically get rid of GWT code dealing
with showing the lists of log entries (top-right panel) and just use HTML
returned from the log feed. GWT will only be used to offer the list of
endpoints, filters, etc, and may be in time we can just replace GWT with
something lighter/simpler.
Tomasz, AtomPullServer has getAlternateFeed and getAlternateEntry methods
and they serve HTML as opposed to XML feeds.
RSSBandit reader on Windows is very good in that it shows alternate feeds
and entries properly.
I think GWT should have some browser support, so if you could make the
top/right panel a browser window which will only show the alternate content,
as part of the frame, such that whenever log entry links are clicked, the
output is directed to the right-bottom pane..
Do you think it will make sense ?
Cheers, Sergey
On Thu, Mar 31, 2011 at 4:18 PM, Sergey Beryozkin <sb...@gmail.com>wrote:
> I've added NOTICE to the rt/management-web only as I'm still seeing all
> sort of crashes on Ubuntu, it does look like my environment is somehow
> broken for GWT to work properly as Tomasz did confirm iy was working ok on
> Ubuntu for him. I''ll try to confirm that it is all working on Windows and
> after that I'll update the distribution.
>
> It would be good if few more people tried a logbrowser demo in the
> distribution - it is very easy to run...I think if we can get at least one
> more crash report then we might want to consider 'distributing' the
> logbrowser as a Maven dependency only. I hope it's my Ubuntu distribution
> only (old enough already) that is not stable...
>
> Cheers, Sergey
>
>
> On Wed, Mar 16, 2011 at 2:21 AM, Daniel Kulp <dk...@apache.org> wrote:
>
>> On Tuesday 15 March 2011 6:25:14 PM Tomasz Oponowicz wrote:
>> > > Probably one open issue is to figure out what to do with different .js
>> > > scripts. ? The scripts which are in the rt/management-web module, can
>> > > they be Apache licensed ? Dan - what do you think ?
>> >
>> > There are:
>> >
>> > * json2.min.js - Public Domain;
>> > * uuid.min.js - Dual licensed under the MIT and GPL licenses;
>>
>> Both licenses are fine, but we need to note them in the NOTICE file.
>>
>> You'll need to create a
>> src/main/appended-resources/META-INF/NOTICE
>> file that contains whatever text would be required.
>>
>> For example, in the javascript module, the NOTICE there has:
>>
>> Portions of the file cxf-utils.js derives from code marked:
>> This code was written by Tyler Akins and has been placed in the
>> public domain. It would be nice if you left this header intact.
>> Base64 code from Tyler Akins -- http://rumkin.com
>>
>> You can use that as a guideline.
>>
>> You'll also need to update
>> distribution/src/main/appended-resources/META-INF/NOTICE
>> with the same info for the release assemplies to pick up.
>>
>>
>> --
>> Daniel Kulp
>> dkulp@apache.org
>> http://dankulp.com/blog
>> Talend - http://www.talend.com
>>
>
>
>
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
I've added NOTICE to the rt/management-web only as I'm still seeing all sort
of crashes on Ubuntu, it does look like my environment is somehow broken for
GWT to work properly as Tomasz did confirm iy was working ok on Ubuntu for
him. I''ll try to confirm that it is all working on Windows and after that
I'll update the distribution.
It would be good if few more people tried a logbrowser demo in the
distribution - it is very easy to run...I think if we can get at least one
more crash report then we might want to consider 'distributing' the
logbrowser as a Maven dependency only. I hope it's my Ubuntu distribution
only (old enough already) that is not stable...
Cheers, Sergey
On Wed, Mar 16, 2011 at 2:21 AM, Daniel Kulp <dk...@apache.org> wrote:
> On Tuesday 15 March 2011 6:25:14 PM Tomasz Oponowicz wrote:
> > > Probably one open issue is to figure out what to do with different .js
> > > scripts. ? The scripts which are in the rt/management-web module, can
> > > they be Apache licensed ? Dan - what do you think ?
> >
> > There are:
> >
> > * json2.min.js - Public Domain;
> > * uuid.min.js - Dual licensed under the MIT and GPL licenses;
>
> Both licenses are fine, but we need to note them in the NOTICE file.
>
> You'll need to create a
> src/main/appended-resources/META-INF/NOTICE
> file that contains whatever text would be required.
>
> For example, in the javascript module, the NOTICE there has:
>
> Portions of the file cxf-utils.js derives from code marked:
> This code was written by Tyler Akins and has been placed in the
> public domain. It would be nice if you left this header intact.
> Base64 code from Tyler Akins -- http://rumkin.com
>
> You can use that as a guideline.
>
> You'll also need to update
> distribution/src/main/appended-resources/META-INF/NOTICE
> with the same info for the release assemplies to pick up.
>
>
> --
> Daniel Kulp
> dkulp@apache.org
> http://dankulp.com/blog
> Talend - http://www.talend.com
>
--
Sergey Beryozkin
Application Integration Division of Talend <http://www.talend.com>
http://sberyozkin.blogspot.com
Re: LogBrowser project is on the trunk
Posted by Daniel Kulp <dk...@apache.org>.
On Tuesday 15 March 2011 6:25:14 PM Tomasz Oponowicz wrote:
> > Probably one open issue is to figure out what to do with different .js
> > scripts. ? The scripts which are in the rt/management-web module, can
> > they be Apache licensed ? Dan - what do you think ?
>
> There are:
>
> * json2.min.js - Public Domain;
> * uuid.min.js - Dual licensed under the MIT and GPL licenses;
Both licenses are fine, but we need to note them in the NOTICE file.
You'll need to create a
src/main/appended-resources/META-INF/NOTICE
file that contains whatever text would be required.
For example, in the javascript module, the NOTICE there has:
Portions of the file cxf-utils.js derives from code marked:
This code was written by Tyler Akins and has been placed in the
public domain. It would be nice if you left this header intact.
Base64 code from Tyler Akins -- http://rumkin.com
You can use that as a guideline.
You'll also need to update
distribution/src/main/appended-resources/META-INF/NOTICE
with the same info for the release assemplies to pick up.
--
Daniel Kulp
dkulp@apache.org
http://dankulp.com/blog
Talend - http://www.talend.com
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi
> Probably one open issue is to figure out what to do with different .js
> > scripts. ? The scripts which are in the rt/management-web module, can
> they
> > be Apache licensed ? Dan - what do you think ?
>
> There are:
>
> * json2.min.js - Public Domain;
> * uuid.min.js - Dual licensed under the MIT and GPL licenses;
>
> Are those scripts basic enough for us just to do our own simple scripts ?
> > The script which is shipped with the log browser demo - should we remove
> it
> > and just use a plain client code to do the invocations for the logs be
> > generated ?
>
I guess we will need to remove that one, I can do it and just add the
WebClient based demo code...
thanks, Sergey
>
> --
> Best regards,
> Tomasz Oponowicz
>
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi Sergey,
On Tue, Mar 15, 2011 at 11:13 PM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi Tomasz
>>
>> > Can you do a sanity check please and confirm that only those filter
>> > fields
>> > which are selected are added to the URI ?
>> > Example, if the filter becomes empty then the query is set to null/empty
>> > value, if only the 'from' date field selected then it is
>> >
>> > _s=date=lt=2011-03-07
>> >
>> > only. And yes, the message value should have '*' symbols on both ends of
>> > the
>> > entered value....
>>
>> Done.
>>
>> Some variations from current implementation:
>>
>>
>> /log/logs?_s=phrase==*test%20test*;date=ge=2011-03-17;date=lt=2011-03-07;(level==INFO,level==WARN)
>>
>> /log/logs?_s=phrase==*test*;date=ge=2011-03-17;date=lt=2011-03-07;(level==INFO,level==WARN)
>> /log/logs?_s=level==INFO,level==WARN
>> /log/logs?_s=date=ge=2011-03-23
>>
>
> That is exactly what is needed, nice...A very minor fix is needed - replace
> 'phrase' with 'message' - this is because LogRecord has a property
> 'message', it's really a minor issue...
>
Fixed.
>> > One thing I noticed in your merge email is that it said that storing the
>> > user settings in the local cache was also removed. But we should
>> > actually
>> > keep it ? So that after we start the browser again it remembers the
>> > information about the endpoints ? If you did remove this code then
>> > please
>> > restore it...
>>
>> At the moment it works like:
>>
>> - Start application;
>> - Load settings from BootStrapStorage web service to memory (later we
>> use settings from memory);
>> - There is synchronization with BootStrapStorage when you
>> add/remove/edit endpoint;
>>
>> ...this solution is according to your quote:
>>
>> "
>> I think it makes sense to keep the list of endpoints and the filter
>> properties the current user has created. *But*, these settings just
>> need to shared across multiple restarts of
>> the browser between all the users. This is because I don't really
>> think it is realistic that one user
>> will want to see EndpointA only, the other one, EndpointB, etc.
>> So lets keep it - I'm not worried about many users using different
>> browsers for checking the logs of the single server :-)
>> "
>>
>> ...I think it's a good idea - users can share settings.
>>
> So you only removed per-user specific settings - that is ok, thanks...
>
>>
>> I don't consider filter options as settings. It's something per user
>> and only during the session.
>>
> Sounds good - lets see if users will want to keep them across restarts, good
> for now...
>
>
> I think we're getting close enough to finalizing the current code on the
> trunk. I still have to trace that Date issue, and play a bit with the
> browser checking the files, but I think things are looking good.
>
> Probably one open issue is to figure out what to do with different .js
> scripts. ? The scripts which are in the rt/management-web module, can they
> be Apache licensed ? Dan - what do you think ?
There are:
* json2.min.js - Public Domain;
* uuid.min.js - Dual licensed under the MIT and GPL licenses;
> The script which is shipped with the log browser demo - should we remove it
> and just use a plain client code to do the invocations for the logs be
> generated ?
>
> thanks, Sergey
>
>>
>>
>> --
>> Best regards,
>> Tomasz Oponowicz
>
>
>
> --
> Sergey Beryozkin
>
> Application Integration Division of Talend
> http://sberyozkin.blogspot.com
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz
> > Can you do a sanity check please and confirm that only those filter
> fields
> > which are selected are added to the URI ?
> > Example, if the filter becomes empty then the query is set to null/empty
> > value, if only the 'from' date field selected then it is
> >
> > _s=date=lt=2011-03-07
> >
> > only. And yes, the message value should have '*' symbols on both ends of
> the
> > entered value....
>
> Done.
>
> Some variations from current implementation:
>
>
> /log/logs?_s=phrase==*test%20test*;date=ge=2011-03-17;date=lt=2011-03-07;(level==INFO,level==WARN)
>
> /log/logs?_s=phrase==*test*;date=ge=2011-03-17;date=lt=2011-03-07;(level==INFO,level==WARN)
> /log/logs?_s=level==INFO,level==WARN
> /log/logs?_s=date=ge=2011-03-23
>
>
That is exactly what is needed, nice...A very minor fix is needed - replace
'phrase' with 'message' - this is because LogRecord has a property
'message', it's really a minor issue...
>
> > One thing I noticed in your merge email is that it said that storing the
> > user settings in the local cache was also removed. But we should actually
> > keep it ? So that after we start the browser again it remembers the
> > information about the endpoints ? If you did remove this code then please
> > restore it...
>
> At the moment it works like:
>
> - Start application;
> - Load settings from BootStrapStorage web service to memory (later we
> use settings from memory);
> - There is synchronization with BootStrapStorage when you
> add/remove/edit endpoint;
>
> ...this solution is according to your quote:
>
> "
> I think it makes sense to keep the list of endpoints and the filter
> properties the current user has created. *But*, these settings just
> need to shared across multiple restarts of
> the browser between all the users. This is because I don't really
> think it is realistic that one user
> will want to see EndpointA only, the other one, EndpointB, etc.
> So lets keep it - I'm not worried about many users using different
> browsers for checking the logs of the single server :-)
> "
>
> ...I think it's a good idea - users can share settings.
>
> So you only removed per-user specific settings - that is ok, thanks...
> I don't consider filter options as settings. It's something per user
> and only during the session.
>
> Sounds good - lets see if users will want to keep them across restarts,
good for now...
I think we're getting close enough to finalizing the current code on the
trunk. I still have to trace that Date issue, and play a bit with the
browser checking the files, but I think things are looking good.
Probably one open issue is to figure out what to do with different .js
scripts. ? The scripts which are in the rt/management-web module, can they
be Apache licensed ? Dan - what do you think ?
The script which is shipped with the log browser demo - should we remove it
and just use a plain client code to do the invocations for the logs be
generated ?
thanks, Sergey
>
> --
> Best regards,
> Tomasz Oponowicz
>
--
Sergey Beryozkin
Application Integration Division of Talend <http://www.talend.com>
http://sberyozkin.blogspot.com
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi Sergey,
On Sun, Mar 13, 2011 at 2:50 PM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi Tomasz
>
> thanks for your effort, comments inline,
>
>> At the moment URL looks like:
>>
>>
>> <endpoint_url>s?_s=date=ge=2011-03-07;date=lt=2011-03-07;level==DEBUG,level==INFO,level==WARN,level==ERROR
>>
>
> There is one more update needed:
>
> _s=date=ge=2011-03-07;date=lt=2011-03-07;(level==DEBUG,level==INFO,level==WARN,level==ERROR)
>
> Please note the brackets around the multiple OR level condition. So the
> above says "all the records within the given date range *and* its level
> equal to INFO *or* DEBUG, etc"
>
> Can you do a sanity check please and confirm that only those filter fields
> which are selected are added to the URI ?
> Example, if the filter becomes empty then the query is set to null/empty
> value, if only the 'from' date field selected then it is
>
> _s=date=lt=2011-03-07
>
> only. And yes, the message value should have '*' symbols on both ends of the
> entered value....
Done.
Some variations from current implementation:
/log/logs?_s=phrase==*test%20test*;date=ge=2011-03-17;date=lt=2011-03-07;(level==INFO,level==WARN)
/log/logs?_s=phrase==*test*;date=ge=2011-03-17;date=lt=2011-03-07;(level==INFO,level==WARN)
/log/logs?_s=level==INFO,level==WARN
/log/logs?_s=date=ge=2011-03-23
>>
>> >> > - Remove the Authentication/SignIn/SignOut related code/resources. We
>> >> > just
>> >> > don't have time now to deal with the authentication/security issues
>> >> > and
>> >> > we'd
>> >> > need to have another discussion later on about it. I believe you
>> >> > agreed
>> >> > to
>> >> > it too.
>> >>
>> >> I'm working on it right now. I almost finished.
>>
>> Done.
>>
> Brilliant - please be assured that we will discuss the issue of the
> authentication and the use of HTTPs later on, in due time.
>
> One thing I noticed in your merge email is that it said that storing the
> user settings in the local cache was also removed. But we should actually
> keep it ? So that after we start the browser again it remembers the
> information about the endpoints ? If you did remove this code then please
> restore it...
At the moment it works like:
- Start application;
- Load settings from BootStrapStorage web service to memory (later we
use settings from memory);
- There is synchronization with BootStrapStorage when you
add/remove/edit endpoint;
...this solution is according to your quote:
"
I think it makes sense to keep the list of endpoints and the filter
properties the current user has created. *But*, these settings just
need to shared across multiple restarts of
the browser between all the users. This is because I don't really
think it is realistic that one user
will want to see EndpointA only, the other one, EndpointB, etc.
So lets keep it - I'm not worried about many users using different
browsers for checking the logs of the single server :-)
"
...I think it's a good idea - users can share settings.
I don't consider filter options as settings. It's something per user
and only during the session.
>> >> > - Explain a bit which code/script deals with processing the dates. I
>> >> > know
>> >> > you spent a lot of time on trying to reproduce the application
>> >> > exception,
>> >> > including testing it on Ubuntu, but I'd like to give it a try and see
>> >> > if
>> >> > I
>> >> > can spot something that causes the issue on my particular OS - if I
>> >> > can
>> >> > trace it to some bad Java or Firefox installation then I'll get
>> >> > relaxed
>> >> > :-)...
>>
>> I added helper classes and static resource which support GWT debug mode.
>>
>> To execute application in debug mode please fallow instructions:
>>
>> Run application:
>>
>> 1) To run LogBorwser in the debug mode, you must type in the terminal:
>> mvn clean gwt:run -Pdev
>> 2) Open new browser's window and go to:
>>
>> http://127.0.0.1:8888/logbrowser/LogBrowser.html?gwt.codesvr=127.0.0.1:9997
>>
>> Configure settings:
>>
>> 4) Add new endpoint with URL:
>> http://127.0.0.1:8888/log/logs
>> 5) Generate entry logs by interacting with URL:
>> http://127.0.0.1:8888/generate.html
>
> Great - hopefully that will let me trace the problem with dates.
>
> Thanks, Sergey
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz
thanks for your effort, comments inline,
At the moment URL looks like:
>
>
> <endpoint_url>s?_s=date=ge=2011-03-07;date=lt=2011-03-07;level==DEBUG,level==INFO,level==WARN,level==ERROR
>
>
There is one more update needed:
_s=date=ge=2011-03-07;date=lt=2011-03-07;(level==DEBUG,level==INFO,level==WARN,level==ERROR)
Please note the brackets around the multiple OR level condition. So the
above says "all the records within the given date range *and* its level
equal to INFO *or* DEBUG, etc"
Can you do a sanity check please and confirm that only those filter fields
which are selected are added to the URI ?
Example, if the filter becomes empty then the query is set to null/empty
value, if only the 'from' date field selected then it is
_s=date=lt=2011-03-07
only. And yes, the message value should have '*' symbols on both ends of the
entered value....
> >> > - Remove the Authentication/SignIn/SignOut related code/resources. We
> >> > just
> >> > don't have time now to deal with the authentication/security issues
> and
> >> > we'd
> >> > need to have another discussion later on about it. I believe you
> agreed
> >> > to
> >> > it too.
> >>
> >> I'm working on it right now. I almost finished.
>
> Done.
>
> Brilliant - please be assured that we will discuss the issue of the
authentication and the use of HTTPs later on, in due time.
One thing I noticed in your merge email is that it said that storing the
user settings in the local cache was also removed. But we should actually
keep it ? So that after we start the browser again it remembers the
information about the endpoints ? If you did remove this code then please
restore it...
>> > - Explain a bit which code/script deals with processing the dates. I
> >> > know
> >> > you spent a lot of time on trying to reproduce the application
> >> > exception,
> >> > including testing it on Ubuntu, but I'd like to give it a try and see
> if
> >> > I
> >> > can spot something that causes the issue on my particular OS - if I
> can
> >> > trace it to some bad Java or Firefox installation then I'll get
> relaxed
> >> > :-)...
>
> I added helper classes and static resource which support GWT debug mode.
>
> To execute application in debug mode please fallow instructions:
>
> Run application:
>
> 1) To run LogBorwser in the debug mode, you must type in the terminal:
> mvn clean gwt:run -Pdev
> 2) Open new browser's window and go to:
>
> http://127.0.0.1:8888/logbrowser/LogBrowser.html?gwt.codesvr=127.0.0.1:9997
>
> Configure settings:
>
> 4) Add new endpoint with URL:
> http://127.0.0.1:8888/log/logs
> 5) Generate entry logs by interacting with URL:
> http://127.0.0.1:8888/generate.html
>
Great - hopefully that will let me trace the problem with dates.
Thanks, Sergey
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi Sergey,
On Wed, Mar 9, 2011 at 4:51 PM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi Tomasz
>
> On Wed, Mar 9, 2011 at 3:00 PM, Tomasz Oponowicz
> <to...@gmail.com> wrote:
>>
>> Hi Sergey,
>>
>> On Wed, Mar 9, 2011 at 2:28 PM, Sergey Beryozkin <sb...@gmail.com>
>> wrote:
>> > Hi Tomasz
>> >
>> > I'm finishing some work for AtomPullServer be able to connect to the
>> > existing file logs, I saw you committing some relevant updates to the
>> > LogBrowser code too, thanks. I'll will experiment with the Filter dialog
>> > shortly.
>>
>> I've just added logic to "Filter" dialog box.
>>
>> It doesn't work because we have to talk about how should I pass
>> attributes to AtomPullServer. At the moment I'm building URL:
>>
>>
>> <endpoint_url>?_s=date=ge=1299668400000;date=lt=1300273200000;level==DEBUG;level==WARN
>>
>> How should I pass an array of levels?
>>
>
> ';' means 'and' while ',' means 'or' so in this case we have to use ',' :
>
> <endpoint_url>?_s=level==WARN,level==DEBUG
>
> Dates need to be passed differently, you can probably use the following
> code:
>
> String format = "yyyy-MM-dd";
> String date = new SimpleDateFormat(format).format(date);
> and add it to the query.
>
> I'll need to write a test involving dates as well.
Fixed.
At the moment URL looks like:
<endpoint_url>s?_s=date=ge=2011-03-07;date=lt=2011-03-07;level==DEBUG,level==INFO,level==WARN,level==ERROR
>> I've made assumption that feed's links (i.e. next, previous etc) will
>> contain filtering attributes.
>
> Yes, exactly, I've already written few tests confirming it.
>
>>
>> > What I'd really appreciate if you could help me with the following, hope
>> > it
>> > will take just a fraction of your time:
>> >
>> > - Remove the Authentication/SignIn/SignOut related code/resources. We
>> > just
>> > don't have time now to deal with the authentication/security issues and
>> > we'd
>> > need to have another discussion later on about it. I believe you agreed
>> > to
>> > it too.
>>
>> I'm working on it right now. I almost finished.
Done.
>> > - Explain a bit which code/script deals with processing the dates. I
>> > know
>> > you spent a lot of time on trying to reproduce the application
>> > exception,
>> > including testing it on Ubuntu, but I'd like to give it a try and see if
>> > I
>> > can spot something that causes the issue on my particular OS - if I can
>> > trace it to some bad Java or Firefox installation then I'll get relaxed
>> > :-)...
I added helper classes and static resource which support GWT debug mode.
To execute application in debug mode please fallow instructions:
Run application:
1) To run LogBorwser in the debug mode, you must type in the terminal:
mvn clean gwt:run -Pdev
2) Open new browser's window and go to:
http://127.0.0.1:8888/logbrowser/LogBrowser.html?gwt.codesvr=127.0.0.1:9997
Configure settings:
4) Add new endpoint with URL:
http://127.0.0.1:8888/log/logs
5) Generate entry logs by interacting with URL:
http://127.0.0.1:8888/generate.html
>> I will give more details at this afternoon.
>>
>
> thanks again :-)
>
>>
>> > Thanks, Sergey
>> >
>> >
>> > On Sun, Feb 27, 2011 at 6:20 PM, Sergey Beryozkin <sb...@gmail.com>
>> > wrote:
>> >>
>> >> String text = columnDefinition.getContent(entry);
>> >>
>> >> in SelectableTable.setData
>> >>
>> >> is causing it,
>> >>
>> >> and it's only happening when the Date column is filled in, the Level
>> >> and Message are fine.
>> >>
>> >> Looks like the code co0nverting the Atom/http-format date is brittle a
>> >> bit
>> >>
>> >> cheers, Sergey
>> >>
>> >> On Sun, Feb 27, 2011 at 5:10 PM, Sergey Beryozkin
>> >> <sb...@gmail.com>
>> >> wrote:
>> >> > Hi Tomasz
>> >> >
>> >> > I've spent some time and it is some sort of the asynchronous
>> >> > exception
>> >> > which when captured shows the Date of the log entry. It is thrown
>> >> > when
>> >> > the ViewerViewImpl.setEntries(List<Entry> entries) is called. We'll
>> >> > track it down soon :-)
>> >> >
>> >> > Sergey
>> >> >
>> >> > On Sat, Feb 26, 2011 at 1:23 PM, Tomasz Oponowicz
>> >> > <to...@gmail.com> wrote:
>> >> >> Hi Sergey,
>> >> >>
>> >> >> On Fri, Feb 25, 2011 at 1:19 PM, Sergey Beryozkin
>> >> >> <sb...@gmail.com> wrote:
>> >> >>> Hi Tomasz
>> >> >>>
>> >> >>> I've rerun the demo.
>> >> >>> First of all, the changes you've made recently have definitely made
>> >> >>> it
>> >> >>> look much better, thanks.
>> >> >>> Unfortunately, I'm still hitting this GWT ApplicationException when
>> >> >>> refreshing the endpoint :
>> >> >>> "
>> >> >>> Application Error
>> >> >>> Class$jcb
>> >> >>> 2011-02-25T11:46:55.078Z
>> >> >>> "
>> >> >>>
>> >> >>> Can this message, particularly Class$jcb, help somehow to identify
>> >> >>> the
>> >> >>> problem ? It does look like it's a platform/browser specific issue,
>> >> >>> I'm on Ubuntu 9, Firefox 3.6.13, but it would be good to get rid of
>> >> >>> it
>> >> >>> somehow.
>> >> >>> Just tried Chrome and it showed the same error but with the
>> >> >>> "Class$kbc" - it's probably the some gwt proxy...
>> >> >>>
>> >> >>> Can you please point to the code in the logbrowser project where
>> >> >>> the
>> >> >>> response from the remote endpoint is processed ? I will
>> >> >>> investigate...
>> >> >>
>> >> >> I will install similar environment as a virtual machine - I hope I
>> >> >> will reproduce this issue...
>> >> >>
>> >> >>> Few more comments. I agree, the way Tasks and Endpoints are
>> >> >>> currently
>> >> >>> shown is nice.
>> >> >>>
>> >> >>> - Can you please consider having both ManageEndpoints and Filter
>> >> >>> links
>> >> >>> located under the "Tasks" ? And have Endpoints shown first, with
>> >> >>> the
>> >> >>> "Tasks" in the bottom of the pane ? Ultimately, the user just wants
>> >> >>> to
>> >> >>> see the list of endpoints. Creating/deleting the endpoints and
>> >> >>> applying filters is critical but I'd just prefer the "Tasks" not to
>> >> >>> feature as the main activity of the LogBrowser users...
>> >> >>> - Filter dialog can not be closed at the moment, after it has been
>> >> >>> opened...
>> >> >>> - Probably makes sense to rename "Explore" to "Explorer" given that
>> >> >>> the "ManageEndpoints" panel has the "Back to Explorer" link...
>> >> >>>
>> >> >>> More comments inline...
>> >> >>>
>> >> >>>
>> >> >>>>> - Authentication: I've noticed there's AuthenticationRequired
>> >> >>>>> annotation attached to some of the BootstrapStorage methods - we
>> >> >>>>> really need to remove this annotation and for now just pop-up a
>> >> >>>>> login
>> >> >>>>> window on the browser start-up.
>> >> >>>>> Users will be configuring the LogBrowser application as part of
>> >> >>>>> the
>> >> >>>>> real deployments. So what would be good is to write the GWT
>> >> >>>>> client
>> >> >>>>> code such that it only pops up a window when the initial GET
>> >> >>>>> returns
>> >> >>>>> 401 - can you use CXF WebClient there and do 'Response r =
>> >> >>>>> webClient.get()' and if r.getStatus() == 401 then pop-up a login
>> >> >>>>> dialog ? We can deal with this issue later, when we have more
>> >> >>>>> time,
>> >> >>>>> and then we'll also decide whether to support https in cases when
>> >> >>>>> the
>> >> >>>>> authentication is needed or may be do the UT profile, we'll
>> >> >>>>> see...
>> >> >>>>
>> >> >>>> According to your list of tasks please consider also fallowing
>> >> >>>> tasks:
>> >> >>>
>> >> >>> Thanks for this analysis...
>> >> >>>
>> >> >>>>
>> >> >>>> - Removing "Sign in" feature;
>> >> >>>> - Pros:
>> >> >>>> - Simplify implementation;
>> >> >>>> - Easy configuration for end user;
>> >> >>>> - Every company has got their own internal user
>> >> >>>> authentication system (LDAP, OpenID, internal SSO etc.);
>> >> >>>> - Even if LogBrowser doesn't contain any user
>> >> >>>> authentication
>> >> >>>> system, it's still very easy to add integration with some
>> >> >>>> authentication system:
>> >> >>>> - Simply interceptor before request rich
>> >> >>>> controller;
>> >> >>>> - Apache directives (of course if user use Apache
>> >> >>>> before Tomcat);
>> >> >>>> - Cons:
>> >> >>>> - I understand that feeds should be secured, but I think
>> >> >>>> we
>> >> >>>> should rather concentrate on:
>> >> >>>> - HTTPS connection;
>> >> >>>> - password per feed (optional);
>> >> >>>>
>> >> >>>
>> >> >>> I think we are in agreement here. I'd like to propose:
>> >> >>> - remove the initial Sign-In dialog altogether
>> >> >>> - Remove SighIn and SignOut buttons
>> >> >>> - Remove AuthenticationRequired annotation from the coode
>> >> >>> - postpone dealing with the authentication issues - at the moment
>> >> >>> we
>> >> >>> just need to focus on making sure
>> >> >>> the browser is operational.
>> >> >>>
>> >> >>> After 2.4.0, we can enhance it for the authentication+HTTPS be
>> >> >>> supported.
>> >> >>>
>> >> >>> If you agree then please remove all the authentication-related
>> >> >>> code/settings...
>> >> >>>
>> >> >>>> - Removing storing user settings remotely on the servers;
>> >> >>>> - Pros:
>> >> >>>> - Simplify implementation;
>> >> >>>> - Easy configuration for end user;
>> >> >>>> - Very clear message - all settings are stored in browser
>> >> >>>> local storage. At the moment the logic it's to complicated. It
>> >> >>>> depend
>> >> >>>> on situation we keep settings in memory, browser local storage or
>> >> >>>> bootstrap settings;
>> >> >>>> - Cons:
>> >> >>>> - When end user clear cache all settings will be removed;
>> >> >>>> - Settings are stored per browser. When you add something
>> >> >>>> in
>> >> >>>> Firefox it won't be available in Chrome;
>> >> >>>>
>> >> >>>
>> >> >>> I think it makes sense to keep the list of endpoints and the filter
>> >> >>> properties the current user has created.
>> >> >>> *But*, these settings just need to shared across multiple restarts
>> >> >>> of
>> >> >>> the browser between all the users.
>> >> >>> This is because I don't really think it is realistic that one user
>> >> >>> will want to see EndpointA only, the other one, EndpointB, etc.
>> >> >>> So lets keep it - I'm not worried about many users using different
>> >> >>> browsers for checking the logs of the single server :-)
>> >> >>>
>> >> >>>> What do you think about these tasks? I'd like to keep LogBrowser
>> >> >>>> minimalistic.
>> >> >>>>
>> >> >>> makes sense :-)
>> >> >>>
>> >> >>> Cheers, Sergey
>> >> >>>
>> >> >>>> --
>> >> >>>> Best regards,
>> >> >>>> Tomasz Oponowicz
>> >> >>>>
>> >> >>>
>> >> >>
>> >> >> --
>> >> >> Best regards,
>> >> >> Tomasz Oponowicz
>> >> >>
>> >> >
>> >
>> >
>> >
>> > --
>> > Sergey Beryozkin
>> >
>> > Application Integration Division of Talend
>> > http://sberyozkin.blogspot.com
>> >
>>
>>
>>
>> --
>> Best regards,
>> Tomasz Oponowicz
>
>
>
> --
> Sergey Beryozkin
>
> Application Integration Division of Talend
> http://sberyozkin.blogspot.com
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz
> In org.apache.cxf.management.web.browser.client.service.browser.Entry
> class, line 213:
>
> @Nullable
> public Date getEventTimestamp() {
> String value = getTagValue(EVENT_TIMESTAMP_TAG);
> return value != null ? DATETIME_FORMATTER.parse(value) : null;
> }
>
> And the definition of the DATETIME_FORMATTER is:
>
> private static final DateTimeFormat DATETIME_FORMATTER =
> DateTimeFormat.getFormat("yyyy-MM-ddTHH:mm:ss'.'SSSZ");
>
>
Very good, thanks
> During a merge you removed some test classes which allows you to
> execute full feature debug mode.
> I can provide a patch (with these classes) and some instructions if you
> want.
>
>
>
Please do, sorry about that ...
Cheers, Sergey
>
>
> --
> Best regards,
> Tomasz Oponowicz
>
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi Sergey,
On Fri, Mar 11, 2011 at 11:55 AM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi Tomasz
>>
>> >>
>> >>
>> >> <endpoint_url>?_s=date=ge=1299668400000;date=lt=1300273200000;level==DEBUG;level==WARN
>> >>
>> >> How should I pass an array of levels?
>> >>
>> >
>> > ';' means 'and' while ',' means 'or' so in this case we have to use ','
>> > :
>> >
>> > <endpoint_url>?_s=level==WARN,level==DEBUG
>> >
>> > Dates need to be passed differently, you can probably use the following
>> > code:
>> >
>> > String format = "yyyy-MM-dd";
>> > String date = new SimpleDateFormat(format).format(date);
>> > and add it to the query.
>>
>> Thanks.
>>
>
> We also have a message property there so just wanted to ask you to add
> wildcards on both sides of the typed phrase.
> For example, "find all the log records with messages containing a 'problem'
> word":
>
> _s=message==*problem*
>
> realistically we should let users do partial checks, and we can't expect
> them to type the complete message for a match to succeed, so adding
> wildcards allows for it. "*problem" and "problem*" are more specific
> possibilities which won't apply in our case...
>
> Hope I'm not stressing you too much :-) - I appreciate you are busy with
> your own work,
No problem. It looks like quick fix. I will look at this on Monday.
> Thanks, Sergey
>
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz
> >>
> >>
> <endpoint_url>?_s=date=ge=1299668400000;date=lt=1300273200000;level==DEBUG;level==WARN
> >>
> >> How should I pass an array of levels?
> >>
> >
> > ';' means 'and' while ',' means 'or' so in this case we have to use ',' :
> >
> > <endpoint_url>?_s=level==WARN,level==DEBUG
> >
> > Dates need to be passed differently, you can probably use the following
> > code:
> >
> > String format = "yyyy-MM-dd";
> > String date = new SimpleDateFormat(format).format(date);
> > and add it to the query.
>
> Thanks.
>
>
We also have a message property there so just wanted to ask you to add
wildcards on both sides of the typed phrase.
For example, "find all the log records with messages containing a 'problem'
word":
_s=message==*problem*
realistically we should let users do partial checks, and we can't expect
them to type the complete message for a match to succeed, so adding
wildcards allows for it. "*problem" and "problem*" are more specific
possibilities which won't apply in our case...
Hope I'm not stressing you too much :-) - I appreciate you are busy with
your own work,
Thanks, Sergey
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi Sergey,
On Wed, Mar 9, 2011 at 4:51 PM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi Tomasz
>
> On Wed, Mar 9, 2011 at 3:00 PM, Tomasz Oponowicz
> <to...@gmail.com> wrote:
>>
>> Hi Sergey,
>>
>> On Wed, Mar 9, 2011 at 2:28 PM, Sergey Beryozkin <sb...@gmail.com>
>> wrote:
>> > Hi Tomasz
>> >
>> > I'm finishing some work for AtomPullServer be able to connect to the
>> > existing file logs, I saw you committing some relevant updates to the
>> > LogBrowser code too, thanks. I'll will experiment with the Filter dialog
>> > shortly.
>>
>> I've just added logic to "Filter" dialog box.
>>
>> It doesn't work because we have to talk about how should I pass
>> attributes to AtomPullServer. At the moment I'm building URL:
>>
>>
>> <endpoint_url>?_s=date=ge=1299668400000;date=lt=1300273200000;level==DEBUG;level==WARN
>>
>> How should I pass an array of levels?
>>
>
> ';' means 'and' while ',' means 'or' so in this case we have to use ',' :
>
> <endpoint_url>?_s=level==WARN,level==DEBUG
>
> Dates need to be passed differently, you can probably use the following
> code:
>
> String format = "yyyy-MM-dd";
> String date = new SimpleDateFormat(format).format(date);
> and add it to the query.
Thanks.
> I'll need to write a test involving dates as well.
>
>> I've made assumption that feed's links (i.e. next, previous etc) will
>> contain filtering attributes.
>
> Yes, exactly, I've already written few tests confirming it.
>
>>
>> > What I'd really appreciate if you could help me with the following, hope
>> > it
>> > will take just a fraction of your time:
>> >
>> > - Remove the Authentication/SignIn/SignOut related code/resources. We
>> > just
>> > don't have time now to deal with the authentication/security issues and
>> > we'd
>> > need to have another discussion later on about it. I believe you agreed
>> > to
>> > it too.
>>
>> I'm working on it right now. I almost finished.
>>
>> > - Explain a bit which code/script deals with processing the dates. I
>> > know
>> > you spent a lot of time on trying to reproduce the application
>> > exception,
>> > including testing it on Ubuntu, but I'd like to give it a try and see if
>> > I
>> > can spot something that causes the issue on my particular OS - if I can
>> > trace it to some bad Java or Firefox installation then I'll get relaxed
>> > :-)...
>>
>> I will give more details at this afternoon.
In org.apache.cxf.management.web.browser.client.service.browser.Entry
class, line 213:
@Nullable
public Date getEventTimestamp() {
String value = getTagValue(EVENT_TIMESTAMP_TAG);
return value != null ? DATETIME_FORMATTER.parse(value) : null;
}
And the definition of the DATETIME_FORMATTER is:
private static final DateTimeFormat DATETIME_FORMATTER =
DateTimeFormat.getFormat("yyyy-MM-ddTHH:mm:ss'.'SSSZ");
During a merge you removed some test classes which allows you to
execute full feature debug mode.
I can provide a patch (with these classes) and some instructions if you want.
>
> thanks again :-)
>
>>
>> > Thanks, Sergey
>> >
>> >
>> > On Sun, Feb 27, 2011 at 6:20 PM, Sergey Beryozkin <sb...@gmail.com>
>> > wrote:
>> >>
>> >> String text = columnDefinition.getContent(entry);
>> >>
>> >> in SelectableTable.setData
>> >>
>> >> is causing it,
>> >>
>> >> and it's only happening when the Date column is filled in, the Level
>> >> and Message are fine.
>> >>
>> >> Looks like the code co0nverting the Atom/http-format date is brittle a
>> >> bit
>> >>
>> >> cheers, Sergey
>> >>
>> >> On Sun, Feb 27, 2011 at 5:10 PM, Sergey Beryozkin
>> >> <sb...@gmail.com>
>> >> wrote:
>> >> > Hi Tomasz
>> >> >
>> >> > I've spent some time and it is some sort of the asynchronous
>> >> > exception
>> >> > which when captured shows the Date of the log entry. It is thrown
>> >> > when
>> >> > the ViewerViewImpl.setEntries(List<Entry> entries) is called. We'll
>> >> > track it down soon :-)
>> >> >
>> >> > Sergey
>> >> >
>> >> > On Sat, Feb 26, 2011 at 1:23 PM, Tomasz Oponowicz
>> >> > <to...@gmail.com> wrote:
>> >> >> Hi Sergey,
>> >> >>
>> >> >> On Fri, Feb 25, 2011 at 1:19 PM, Sergey Beryozkin
>> >> >> <sb...@gmail.com> wrote:
>> >> >>> Hi Tomasz
>> >> >>>
>> >> >>> I've rerun the demo.
>> >> >>> First of all, the changes you've made recently have definitely made
>> >> >>> it
>> >> >>> look much better, thanks.
>> >> >>> Unfortunately, I'm still hitting this GWT ApplicationException when
>> >> >>> refreshing the endpoint :
>> >> >>> "
>> >> >>> Application Error
>> >> >>> Class$jcb
>> >> >>> 2011-02-25T11:46:55.078Z
>> >> >>> "
>> >> >>>
>> >> >>> Can this message, particularly Class$jcb, help somehow to identify
>> >> >>> the
>> >> >>> problem ? It does look like it's a platform/browser specific issue,
>> >> >>> I'm on Ubuntu 9, Firefox 3.6.13, but it would be good to get rid of
>> >> >>> it
>> >> >>> somehow.
>> >> >>> Just tried Chrome and it showed the same error but with the
>> >> >>> "Class$kbc" - it's probably the some gwt proxy...
>> >> >>>
>> >> >>> Can you please point to the code in the logbrowser project where
>> >> >>> the
>> >> >>> response from the remote endpoint is processed ? I will
>> >> >>> investigate...
>> >> >>
>> >> >> I will install similar environment as a virtual machine - I hope I
>> >> >> will reproduce this issue...
>> >> >>
>> >> >>> Few more comments. I agree, the way Tasks and Endpoints are
>> >> >>> currently
>> >> >>> shown is nice.
>> >> >>>
>> >> >>> - Can you please consider having both ManageEndpoints and Filter
>> >> >>> links
>> >> >>> located under the "Tasks" ? And have Endpoints shown first, with
>> >> >>> the
>> >> >>> "Tasks" in the bottom of the pane ? Ultimately, the user just wants
>> >> >>> to
>> >> >>> see the list of endpoints. Creating/deleting the endpoints and
>> >> >>> applying filters is critical but I'd just prefer the "Tasks" not to
>> >> >>> feature as the main activity of the LogBrowser users...
>> >> >>> - Filter dialog can not be closed at the moment, after it has been
>> >> >>> opened...
>> >> >>> - Probably makes sense to rename "Explore" to "Explorer" given that
>> >> >>> the "ManageEndpoints" panel has the "Back to Explorer" link...
>> >> >>>
>> >> >>> More comments inline...
>> >> >>>
>> >> >>>
>> >> >>>>> - Authentication: I've noticed there's AuthenticationRequired
>> >> >>>>> annotation attached to some of the BootstrapStorage methods - we
>> >> >>>>> really need to remove this annotation and for now just pop-up a
>> >> >>>>> login
>> >> >>>>> window on the browser start-up.
>> >> >>>>> Users will be configuring the LogBrowser application as part of
>> >> >>>>> the
>> >> >>>>> real deployments. So what would be good is to write the GWT
>> >> >>>>> client
>> >> >>>>> code such that it only pops up a window when the initial GET
>> >> >>>>> returns
>> >> >>>>> 401 - can you use CXF WebClient there and do 'Response r =
>> >> >>>>> webClient.get()' and if r.getStatus() == 401 then pop-up a login
>> >> >>>>> dialog ? We can deal with this issue later, when we have more
>> >> >>>>> time,
>> >> >>>>> and then we'll also decide whether to support https in cases when
>> >> >>>>> the
>> >> >>>>> authentication is needed or may be do the UT profile, we'll
>> >> >>>>> see...
>> >> >>>>
>> >> >>>> According to your list of tasks please consider also fallowing
>> >> >>>> tasks:
>> >> >>>
>> >> >>> Thanks for this analysis...
>> >> >>>
>> >> >>>>
>> >> >>>> - Removing "Sign in" feature;
>> >> >>>> - Pros:
>> >> >>>> - Simplify implementation;
>> >> >>>> - Easy configuration for end user;
>> >> >>>> - Every company has got their own internal user
>> >> >>>> authentication system (LDAP, OpenID, internal SSO etc.);
>> >> >>>> - Even if LogBrowser doesn't contain any user
>> >> >>>> authentication
>> >> >>>> system, it's still very easy to add integration with some
>> >> >>>> authentication system:
>> >> >>>> - Simply interceptor before request rich
>> >> >>>> controller;
>> >> >>>> - Apache directives (of course if user use Apache
>> >> >>>> before Tomcat);
>> >> >>>> - Cons:
>> >> >>>> - I understand that feeds should be secured, but I think
>> >> >>>> we
>> >> >>>> should rather concentrate on:
>> >> >>>> - HTTPS connection;
>> >> >>>> - password per feed (optional);
>> >> >>>>
>> >> >>>
>> >> >>> I think we are in agreement here. I'd like to propose:
>> >> >>> - remove the initial Sign-In dialog altogether
>> >> >>> - Remove SighIn and SignOut buttons
>> >> >>> - Remove AuthenticationRequired annotation from the coode
>> >> >>> - postpone dealing with the authentication issues - at the moment
>> >> >>> we
>> >> >>> just need to focus on making sure
>> >> >>> the browser is operational.
>> >> >>>
>> >> >>> After 2.4.0, we can enhance it for the authentication+HTTPS be
>> >> >>> supported.
>> >> >>>
>> >> >>> If you agree then please remove all the authentication-related
>> >> >>> code/settings...
>> >> >>>
>> >> >>>> - Removing storing user settings remotely on the servers;
>> >> >>>> - Pros:
>> >> >>>> - Simplify implementation;
>> >> >>>> - Easy configuration for end user;
>> >> >>>> - Very clear message - all settings are stored in browser
>> >> >>>> local storage. At the moment the logic it's to complicated. It
>> >> >>>> depend
>> >> >>>> on situation we keep settings in memory, browser local storage or
>> >> >>>> bootstrap settings;
>> >> >>>> - Cons:
>> >> >>>> - When end user clear cache all settings will be removed;
>> >> >>>> - Settings are stored per browser. When you add something
>> >> >>>> in
>> >> >>>> Firefox it won't be available in Chrome;
>> >> >>>>
>> >> >>>
>> >> >>> I think it makes sense to keep the list of endpoints and the filter
>> >> >>> properties the current user has created.
>> >> >>> *But*, these settings just need to shared across multiple restarts
>> >> >>> of
>> >> >>> the browser between all the users.
>> >> >>> This is because I don't really think it is realistic that one user
>> >> >>> will want to see EndpointA only, the other one, EndpointB, etc.
>> >> >>> So lets keep it - I'm not worried about many users using different
>> >> >>> browsers for checking the logs of the single server :-)
>> >> >>>
>> >> >>>> What do you think about these tasks? I'd like to keep LogBrowser
>> >> >>>> minimalistic.
>> >> >>>>
>> >> >>> makes sense :-)
>> >> >>>
>> >> >>> Cheers, Sergey
>> >> >>>
>> >> >>>> --
>> >> >>>> Best regards,
>> >> >>>> Tomasz Oponowicz
>> >> >>>>
>> >> >>>
>> >> >>
>> >> >> --
>> >> >> Best regards,
>> >> >> Tomasz Oponowicz
>> >> >>
>> >> >
>> >
>> >
>> >
>> > --
>> > Sergey Beryozkin
>> >
>> > Application Integration Division of Talend
>> > http://sberyozkin.blogspot.com
>> >
>>
>>
>>
>> --
>> Best regards,
>> Tomasz Oponowicz
>
>
>
> --
> Sergey Beryozkin
>
> Application Integration Division of Talend
> http://sberyozkin.blogspot.com
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz
On Wed, Mar 9, 2011 at 3:00 PM, Tomasz Oponowicz <tomasz.oponowicz@gmail.com
> wrote:
> Hi Sergey,
>
> On Wed, Mar 9, 2011 at 2:28 PM, Sergey Beryozkin <sb...@gmail.com>
> wrote:
> > Hi Tomasz
> >
> > I'm finishing some work for AtomPullServer be able to connect to the
> > existing file logs, I saw you committing some relevant updates to the
> > LogBrowser code too, thanks. I'll will experiment with the Filter dialog
> > shortly.
>
> I've just added logic to "Filter" dialog box.
>
> It doesn't work because we have to talk about how should I pass
> attributes to AtomPullServer. At the moment I'm building URL:
>
>
> <endpoint_url>?_s=date=ge=1299668400000;date=lt=1300273200000;level==DEBUG;level==WARN
>
> How should I pass an array of levels?
>
>
';' means 'and' while ',' means 'or' so in this case we have to use ',' :
<endpoint_url>?_s=level==WARN,level==DEBUG
Dates need to be passed differently, you can probably use the following
code:
String format = "yyyy-MM-dd";
String date = new SimpleDateFormat(format).format(date);
and add it to the query.
I'll need to write a test involving dates as well.
I've made assumption that feed's links (i.e. next, previous etc) will
> contain filtering attributes.
>
Yes, exactly, I've already written few tests confirming it.
>
> > What I'd really appreciate if you could help me with the following, hope
> it
> > will take just a fraction of your time:
> >
> > - Remove the Authentication/SignIn/SignOut related code/resources. We
> just
> > don't have time now to deal with the authentication/security issues and
> we'd
> > need to have another discussion later on about it. I believe you agreed
> to
> > it too.
>
> I'm working on it right now. I almost finished.
>
> > - Explain a bit which code/script deals with processing the dates. I know
> > you spent a lot of time on trying to reproduce the application exception,
> > including testing it on Ubuntu, but I'd like to give it a try and see if
> I
> > can spot something that causes the issue on my particular OS - if I can
> > trace it to some bad Java or Firefox installation then I'll get relaxed
> > :-)...
>
> I will give more details at this afternoon.
>
>
thanks again :-)
> > Thanks, Sergey
> >
> >
> > On Sun, Feb 27, 2011 at 6:20 PM, Sergey Beryozkin <sb...@gmail.com>
> > wrote:
> >>
> >> String text = columnDefinition.getContent(entry);
> >>
> >> in SelectableTable.setData
> >>
> >> is causing it,
> >>
> >> and it's only happening when the Date column is filled in, the Level
> >> and Message are fine.
> >>
> >> Looks like the code co0nverting the Atom/http-format date is brittle a
> bit
> >>
> >> cheers, Sergey
> >>
> >> On Sun, Feb 27, 2011 at 5:10 PM, Sergey Beryozkin <sberyozkin@gmail.com
> >
> >> wrote:
> >> > Hi Tomasz
> >> >
> >> > I've spent some time and it is some sort of the asynchronous exception
> >> > which when captured shows the Date of the log entry. It is thrown when
> >> > the ViewerViewImpl.setEntries(List<Entry> entries) is called. We'll
> >> > track it down soon :-)
> >> >
> >> > Sergey
> >> >
> >> > On Sat, Feb 26, 2011 at 1:23 PM, Tomasz Oponowicz
> >> > <to...@gmail.com> wrote:
> >> >> Hi Sergey,
> >> >>
> >> >> On Fri, Feb 25, 2011 at 1:19 PM, Sergey Beryozkin
> >> >> <sb...@gmail.com> wrote:
> >> >>> Hi Tomasz
> >> >>>
> >> >>> I've rerun the demo.
> >> >>> First of all, the changes you've made recently have definitely made
> it
> >> >>> look much better, thanks.
> >> >>> Unfortunately, I'm still hitting this GWT ApplicationException when
> >> >>> refreshing the endpoint :
> >> >>> "
> >> >>> Application Error
> >> >>> Class$jcb
> >> >>> 2011-02-25T11:46:55.078Z
> >> >>> "
> >> >>>
> >> >>> Can this message, particularly Class$jcb, help somehow to identify
> the
> >> >>> problem ? It does look like it's a platform/browser specific issue,
> >> >>> I'm on Ubuntu 9, Firefox 3.6.13, but it would be good to get rid of
> it
> >> >>> somehow.
> >> >>> Just tried Chrome and it showed the same error but with the
> >> >>> "Class$kbc" - it's probably the some gwt proxy...
> >> >>>
> >> >>> Can you please point to the code in the logbrowser project where the
> >> >>> response from the remote endpoint is processed ? I will
> investigate...
> >> >>
> >> >> I will install similar environment as a virtual machine - I hope I
> >> >> will reproduce this issue...
> >> >>
> >> >>> Few more comments. I agree, the way Tasks and Endpoints are
> currently
> >> >>> shown is nice.
> >> >>>
> >> >>> - Can you please consider having both ManageEndpoints and Filter
> links
> >> >>> located under the "Tasks" ? And have Endpoints shown first, with the
> >> >>> "Tasks" in the bottom of the pane ? Ultimately, the user just wants
> to
> >> >>> see the list of endpoints. Creating/deleting the endpoints and
> >> >>> applying filters is critical but I'd just prefer the "Tasks" not to
> >> >>> feature as the main activity of the LogBrowser users...
> >> >>> - Filter dialog can not be closed at the moment, after it has been
> >> >>> opened...
> >> >>> - Probably makes sense to rename "Explore" to "Explorer" given that
> >> >>> the "ManageEndpoints" panel has the "Back to Explorer" link...
> >> >>>
> >> >>> More comments inline...
> >> >>>
> >> >>>
> >> >>>>> - Authentication: I've noticed there's AuthenticationRequired
> >> >>>>> annotation attached to some of the BootstrapStorage methods - we
> >> >>>>> really need to remove this annotation and for now just pop-up a
> >> >>>>> login
> >> >>>>> window on the browser start-up.
> >> >>>>> Users will be configuring the LogBrowser application as part of
> the
> >> >>>>> real deployments. So what would be good is to write the GWT client
> >> >>>>> code such that it only pops up a window when the initial GET
> >> >>>>> returns
> >> >>>>> 401 - can you use CXF WebClient there and do 'Response r =
> >> >>>>> webClient.get()' and if r.getStatus() == 401 then pop-up a login
> >> >>>>> dialog ? We can deal with this issue later, when we have more
> time,
> >> >>>>> and then we'll also decide whether to support https in cases when
> >> >>>>> the
> >> >>>>> authentication is needed or may be do the UT profile, we'll see...
> >> >>>>
> >> >>>> According to your list of tasks please consider also fallowing
> tasks:
> >> >>>
> >> >>> Thanks for this analysis...
> >> >>>
> >> >>>>
> >> >>>> - Removing "Sign in" feature;
> >> >>>> - Pros:
> >> >>>> - Simplify implementation;
> >> >>>> - Easy configuration for end user;
> >> >>>> - Every company has got their own internal user
> >> >>>> authentication system (LDAP, OpenID, internal SSO etc.);
> >> >>>> - Even if LogBrowser doesn't contain any user
> authentication
> >> >>>> system, it's still very easy to add integration with some
> >> >>>> authentication system:
> >> >>>> - Simply interceptor before request rich
> controller;
> >> >>>> - Apache directives (of course if user use Apache
> >> >>>> before Tomcat);
> >> >>>> - Cons:
> >> >>>> - I understand that feeds should be secured, but I think we
> >> >>>> should rather concentrate on:
> >> >>>> - HTTPS connection;
> >> >>>> - password per feed (optional);
> >> >>>>
> >> >>>
> >> >>> I think we are in agreement here. I'd like to propose:
> >> >>> - remove the initial Sign-In dialog altogether
> >> >>> - Remove SighIn and SignOut buttons
> >> >>> - Remove AuthenticationRequired annotation from the coode
> >> >>> - postpone dealing with the authentication issues - at the moment we
> >> >>> just need to focus on making sure
> >> >>> the browser is operational.
> >> >>>
> >> >>> After 2.4.0, we can enhance it for the authentication+HTTPS be
> >> >>> supported.
> >> >>>
> >> >>> If you agree then please remove all the authentication-related
> >> >>> code/settings...
> >> >>>
> >> >>>> - Removing storing user settings remotely on the servers;
> >> >>>> - Pros:
> >> >>>> - Simplify implementation;
> >> >>>> - Easy configuration for end user;
> >> >>>> - Very clear message - all settings are stored in browser
> >> >>>> local storage. At the moment the logic it's to complicated. It
> depend
> >> >>>> on situation we keep settings in memory, browser local storage or
> >> >>>> bootstrap settings;
> >> >>>> - Cons:
> >> >>>> - When end user clear cache all settings will be removed;
> >> >>>> - Settings are stored per browser. When you add something
> in
> >> >>>> Firefox it won't be available in Chrome;
> >> >>>>
> >> >>>
> >> >>> I think it makes sense to keep the list of endpoints and the filter
> >> >>> properties the current user has created.
> >> >>> *But*, these settings just need to shared across multiple restarts
> of
> >> >>> the browser between all the users.
> >> >>> This is because I don't really think it is realistic that one user
> >> >>> will want to see EndpointA only, the other one, EndpointB, etc.
> >> >>> So lets keep it - I'm not worried about many users using different
> >> >>> browsers for checking the logs of the single server :-)
> >> >>>
> >> >>>> What do you think about these tasks? I'd like to keep LogBrowser
> >> >>>> minimalistic.
> >> >>>>
> >> >>> makes sense :-)
> >> >>>
> >> >>> Cheers, Sergey
> >> >>>
> >> >>>> --
> >> >>>> Best regards,
> >> >>>> Tomasz Oponowicz
> >> >>>>
> >> >>>
> >> >>
> >> >> --
> >> >> Best regards,
> >> >> Tomasz Oponowicz
> >> >>
> >> >
> >
> >
> >
> > --
> > Sergey Beryozkin
> >
> > Application Integration Division of Talend
> > http://sberyozkin.blogspot.com
> >
>
>
>
> --
> Best regards,
> Tomasz Oponowicz
>
--
Sergey Beryozkin
Application Integration Division of Talend <http://www.talend.com>
http://sberyozkin.blogspot.com
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi Sergey,
On Wed, Mar 9, 2011 at 2:28 PM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi Tomasz
>
> I'm finishing some work for AtomPullServer be able to connect to the
> existing file logs, I saw you committing some relevant updates to the
> LogBrowser code too, thanks. I'll will experiment with the Filter dialog
> shortly.
I've just added logic to "Filter" dialog box.
It doesn't work because we have to talk about how should I pass
attributes to AtomPullServer. At the moment I'm building URL:
<endpoint_url>?_s=date=ge=1299668400000;date=lt=1300273200000;level==DEBUG;level==WARN
How should I pass an array of levels?
I've made assumption that feed's links (i.e. next, previous etc) will
contain filtering attributes.
> What I'd really appreciate if you could help me with the following, hope it
> will take just a fraction of your time:
>
> - Remove the Authentication/SignIn/SignOut related code/resources. We just
> don't have time now to deal with the authentication/security issues and we'd
> need to have another discussion later on about it. I believe you agreed to
> it too.
I'm working on it right now. I almost finished.
> - Explain a bit which code/script deals with processing the dates. I know
> you spent a lot of time on trying to reproduce the application exception,
> including testing it on Ubuntu, but I'd like to give it a try and see if I
> can spot something that causes the issue on my particular OS - if I can
> trace it to some bad Java or Firefox installation then I'll get relaxed
> :-)...
I will give more details at this afternoon.
> Thanks, Sergey
>
>
> On Sun, Feb 27, 2011 at 6:20 PM, Sergey Beryozkin <sb...@gmail.com>
> wrote:
>>
>> String text = columnDefinition.getContent(entry);
>>
>> in SelectableTable.setData
>>
>> is causing it,
>>
>> and it's only happening when the Date column is filled in, the Level
>> and Message are fine.
>>
>> Looks like the code co0nverting the Atom/http-format date is brittle a bit
>>
>> cheers, Sergey
>>
>> On Sun, Feb 27, 2011 at 5:10 PM, Sergey Beryozkin <sb...@gmail.com>
>> wrote:
>> > Hi Tomasz
>> >
>> > I've spent some time and it is some sort of the asynchronous exception
>> > which when captured shows the Date of the log entry. It is thrown when
>> > the ViewerViewImpl.setEntries(List<Entry> entries) is called. We'll
>> > track it down soon :-)
>> >
>> > Sergey
>> >
>> > On Sat, Feb 26, 2011 at 1:23 PM, Tomasz Oponowicz
>> > <to...@gmail.com> wrote:
>> >> Hi Sergey,
>> >>
>> >> On Fri, Feb 25, 2011 at 1:19 PM, Sergey Beryozkin
>> >> <sb...@gmail.com> wrote:
>> >>> Hi Tomasz
>> >>>
>> >>> I've rerun the demo.
>> >>> First of all, the changes you've made recently have definitely made it
>> >>> look much better, thanks.
>> >>> Unfortunately, I'm still hitting this GWT ApplicationException when
>> >>> refreshing the endpoint :
>> >>> "
>> >>> Application Error
>> >>> Class$jcb
>> >>> 2011-02-25T11:46:55.078Z
>> >>> "
>> >>>
>> >>> Can this message, particularly Class$jcb, help somehow to identify the
>> >>> problem ? It does look like it's a platform/browser specific issue,
>> >>> I'm on Ubuntu 9, Firefox 3.6.13, but it would be good to get rid of it
>> >>> somehow.
>> >>> Just tried Chrome and it showed the same error but with the
>> >>> "Class$kbc" - it's probably the some gwt proxy...
>> >>>
>> >>> Can you please point to the code in the logbrowser project where the
>> >>> response from the remote endpoint is processed ? I will investigate...
>> >>
>> >> I will install similar environment as a virtual machine - I hope I
>> >> will reproduce this issue...
>> >>
>> >>> Few more comments. I agree, the way Tasks and Endpoints are currently
>> >>> shown is nice.
>> >>>
>> >>> - Can you please consider having both ManageEndpoints and Filter links
>> >>> located under the "Tasks" ? And have Endpoints shown first, with the
>> >>> "Tasks" in the bottom of the pane ? Ultimately, the user just wants to
>> >>> see the list of endpoints. Creating/deleting the endpoints and
>> >>> applying filters is critical but I'd just prefer the "Tasks" not to
>> >>> feature as the main activity of the LogBrowser users...
>> >>> - Filter dialog can not be closed at the moment, after it has been
>> >>> opened...
>> >>> - Probably makes sense to rename "Explore" to "Explorer" given that
>> >>> the "ManageEndpoints" panel has the "Back to Explorer" link...
>> >>>
>> >>> More comments inline...
>> >>>
>> >>>
>> >>>>> - Authentication: I've noticed there's AuthenticationRequired
>> >>>>> annotation attached to some of the BootstrapStorage methods - we
>> >>>>> really need to remove this annotation and for now just pop-up a
>> >>>>> login
>> >>>>> window on the browser start-up.
>> >>>>> Users will be configuring the LogBrowser application as part of the
>> >>>>> real deployments. So what would be good is to write the GWT client
>> >>>>> code such that it only pops up a window when the initial GET
>> >>>>> returns
>> >>>>> 401 - can you use CXF WebClient there and do 'Response r =
>> >>>>> webClient.get()' and if r.getStatus() == 401 then pop-up a login
>> >>>>> dialog ? We can deal with this issue later, when we have more time,
>> >>>>> and then we'll also decide whether to support https in cases when
>> >>>>> the
>> >>>>> authentication is needed or may be do the UT profile, we'll see...
>> >>>>
>> >>>> According to your list of tasks please consider also fallowing tasks:
>> >>>
>> >>> Thanks for this analysis...
>> >>>
>> >>>>
>> >>>> - Removing "Sign in" feature;
>> >>>> - Pros:
>> >>>> - Simplify implementation;
>> >>>> - Easy configuration for end user;
>> >>>> - Every company has got their own internal user
>> >>>> authentication system (LDAP, OpenID, internal SSO etc.);
>> >>>> - Even if LogBrowser doesn't contain any user authentication
>> >>>> system, it's still very easy to add integration with some
>> >>>> authentication system:
>> >>>> - Simply interceptor before request rich controller;
>> >>>> - Apache directives (of course if user use Apache
>> >>>> before Tomcat);
>> >>>> - Cons:
>> >>>> - I understand that feeds should be secured, but I think we
>> >>>> should rather concentrate on:
>> >>>> - HTTPS connection;
>> >>>> - password per feed (optional);
>> >>>>
>> >>>
>> >>> I think we are in agreement here. I'd like to propose:
>> >>> - remove the initial Sign-In dialog altogether
>> >>> - Remove SighIn and SignOut buttons
>> >>> - Remove AuthenticationRequired annotation from the coode
>> >>> - postpone dealing with the authentication issues - at the moment we
>> >>> just need to focus on making sure
>> >>> the browser is operational.
>> >>>
>> >>> After 2.4.0, we can enhance it for the authentication+HTTPS be
>> >>> supported.
>> >>>
>> >>> If you agree then please remove all the authentication-related
>> >>> code/settings...
>> >>>
>> >>>> - Removing storing user settings remotely on the servers;
>> >>>> - Pros:
>> >>>> - Simplify implementation;
>> >>>> - Easy configuration for end user;
>> >>>> - Very clear message - all settings are stored in browser
>> >>>> local storage. At the moment the logic it's to complicated. It depend
>> >>>> on situation we keep settings in memory, browser local storage or
>> >>>> bootstrap settings;
>> >>>> - Cons:
>> >>>> - When end user clear cache all settings will be removed;
>> >>>> - Settings are stored per browser. When you add something in
>> >>>> Firefox it won't be available in Chrome;
>> >>>>
>> >>>
>> >>> I think it makes sense to keep the list of endpoints and the filter
>> >>> properties the current user has created.
>> >>> *But*, these settings just need to shared across multiple restarts of
>> >>> the browser between all the users.
>> >>> This is because I don't really think it is realistic that one user
>> >>> will want to see EndpointA only, the other one, EndpointB, etc.
>> >>> So lets keep it - I'm not worried about many users using different
>> >>> browsers for checking the logs of the single server :-)
>> >>>
>> >>>> What do you think about these tasks? I'd like to keep LogBrowser
>> >>>> minimalistic.
>> >>>>
>> >>> makes sense :-)
>> >>>
>> >>> Cheers, Sergey
>> >>>
>> >>>> --
>> >>>> Best regards,
>> >>>> Tomasz Oponowicz
>> >>>>
>> >>>
>> >>
>> >> --
>> >> Best regards,
>> >> Tomasz Oponowicz
>> >>
>> >
>
>
>
> --
> Sergey Beryozkin
>
> Application Integration Division of Talend
> http://sberyozkin.blogspot.com
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz
I'm finishing some work for AtomPullServer be able to connect to the
existing file logs, I saw you committing some relevant updates to the
LogBrowser code too, thanks. I'll will experiment with the Filter dialog
shortly.
What I'd really appreciate if you could help me with the following, hope it
will take just a fraction of your time:
- Remove the Authentication/SignIn/SignOut related code/resources. We just
don't have time now to deal with the authentication/security issues and we'd
need to have another discussion later on about it. I believe you agreed to
it too.
- Explain a bit which code/script deals with processing the dates. I know
you spent a lot of time on trying to reproduce the application exception,
including testing it on Ubuntu, but I'd like to give it a try and see if I
can spot something that causes the issue on my particular OS - if I can
trace it to some bad Java or Firefox installation then I'll get relaxed
:-)...
Thanks, Sergey
On Sun, Feb 27, 2011 at 6:20 PM, Sergey Beryozkin <sb...@gmail.com>wrote:
> String text = columnDefinition.getContent(entry);
>
> in SelectableTable.setData
>
> is causing it,
>
> and it's only happening when the Date column is filled in, the Level
> and Message are fine.
>
> Looks like the code co0nverting the Atom/http-format date is brittle a bit
>
> cheers, Sergey
>
> On Sun, Feb 27, 2011 at 5:10 PM, Sergey Beryozkin <sb...@gmail.com>
> wrote:
> > Hi Tomasz
> >
> > I've spent some time and it is some sort of the asynchronous exception
> > which when captured shows the Date of the log entry. It is thrown when
> > the ViewerViewImpl.setEntries(List<Entry> entries) is called. We'll
> > track it down soon :-)
> >
> > Sergey
> >
> > On Sat, Feb 26, 2011 at 1:23 PM, Tomasz Oponowicz
> > <to...@gmail.com> wrote:
> >> Hi Sergey,
> >>
> >> On Fri, Feb 25, 2011 at 1:19 PM, Sergey Beryozkin <sb...@gmail.com>
> wrote:
> >>> Hi Tomasz
> >>>
> >>> I've rerun the demo.
> >>> First of all, the changes you've made recently have definitely made it
> >>> look much better, thanks.
> >>> Unfortunately, I'm still hitting this GWT ApplicationException when
> >>> refreshing the endpoint :
> >>> "
> >>> Application Error
> >>> Class$jcb
> >>> 2011-02-25T11:46:55.078Z
> >>> "
> >>>
> >>> Can this message, particularly Class$jcb, help somehow to identify the
> >>> problem ? It does look like it's a platform/browser specific issue,
> >>> I'm on Ubuntu 9, Firefox 3.6.13, but it would be good to get rid of it
> >>> somehow.
> >>> Just tried Chrome and it showed the same error but with the
> >>> "Class$kbc" - it's probably the some gwt proxy...
> >>>
> >>> Can you please point to the code in the logbrowser project where the
> >>> response from the remote endpoint is processed ? I will investigate...
> >>
> >> I will install similar environment as a virtual machine - I hope I
> >> will reproduce this issue...
> >>
> >>> Few more comments. I agree, the way Tasks and Endpoints are currently
> >>> shown is nice.
> >>>
> >>> - Can you please consider having both ManageEndpoints and Filter links
> >>> located under the "Tasks" ? And have Endpoints shown first, with the
> >>> "Tasks" in the bottom of the pane ? Ultimately, the user just wants to
> >>> see the list of endpoints. Creating/deleting the endpoints and
> >>> applying filters is critical but I'd just prefer the "Tasks" not to
> >>> feature as the main activity of the LogBrowser users...
> >>> - Filter dialog can not be closed at the moment, after it has been
> opened...
> >>> - Probably makes sense to rename "Explore" to "Explorer" given that
> >>> the "ManageEndpoints" panel has the "Back to Explorer" link...
> >>>
> >>> More comments inline...
> >>>
> >>>
> >>>>> - Authentication: I've noticed there's AuthenticationRequired
> >>>>> annotation attached to some of the BootstrapStorage methods - we
> >>>>> really need to remove this annotation and for now just pop-up a login
> >>>>> window on the browser start-up.
> >>>>> Users will be configuring the LogBrowser application as part of the
> >>>>> real deployments. So what would be good is to write the GWT client
> >>>>> code such that it only pops up a window when the initial GET returns
> >>>>> 401 - can you use CXF WebClient there and do 'Response r =
> >>>>> webClient.get()' and if r.getStatus() == 401 then pop-up a login
> >>>>> dialog ? We can deal with this issue later, when we have more time,
> >>>>> and then we'll also decide whether to support https in cases when the
> >>>>> authentication is needed or may be do the UT profile, we'll see...
> >>>>
> >>>> According to your list of tasks please consider also fallowing tasks:
> >>>
> >>> Thanks for this analysis...
> >>>
> >>>>
> >>>> - Removing "Sign in" feature;
> >>>> - Pros:
> >>>> - Simplify implementation;
> >>>> - Easy configuration for end user;
> >>>> - Every company has got their own internal user
> >>>> authentication system (LDAP, OpenID, internal SSO etc.);
> >>>> - Even if LogBrowser doesn't contain any user authentication
> >>>> system, it's still very easy to add integration with some
> >>>> authentication system:
> >>>> - Simply interceptor before request rich controller;
> >>>> - Apache directives (of course if user use Apache
> >>>> before Tomcat);
> >>>> - Cons:
> >>>> - I understand that feeds should be secured, but I think we
> >>>> should rather concentrate on:
> >>>> - HTTPS connection;
> >>>> - password per feed (optional);
> >>>>
> >>>
> >>> I think we are in agreement here. I'd like to propose:
> >>> - remove the initial Sign-In dialog altogether
> >>> - Remove SighIn and SignOut buttons
> >>> - Remove AuthenticationRequired annotation from the coode
> >>> - postpone dealing with the authentication issues - at the moment we
> >>> just need to focus on making sure
> >>> the browser is operational.
> >>>
> >>> After 2.4.0, we can enhance it for the authentication+HTTPS be
> supported.
> >>>
> >>> If you agree then please remove all the authentication-related
> code/settings...
> >>>
> >>>> - Removing storing user settings remotely on the servers;
> >>>> - Pros:
> >>>> - Simplify implementation;
> >>>> - Easy configuration for end user;
> >>>> - Very clear message - all settings are stored in browser
> >>>> local storage. At the moment the logic it's to complicated. It depend
> >>>> on situation we keep settings in memory, browser local storage or
> >>>> bootstrap settings;
> >>>> - Cons:
> >>>> - When end user clear cache all settings will be removed;
> >>>> - Settings are stored per browser. When you add something in
> >>>> Firefox it won't be available in Chrome;
> >>>>
> >>>
> >>> I think it makes sense to keep the list of endpoints and the filter
> >>> properties the current user has created.
> >>> *But*, these settings just need to shared across multiple restarts of
> >>> the browser between all the users.
> >>> This is because I don't really think it is realistic that one user
> >>> will want to see EndpointA only, the other one, EndpointB, etc.
> >>> So lets keep it - I'm not worried about many users using different
> >>> browsers for checking the logs of the single server :-)
> >>>
> >>>> What do you think about these tasks? I'd like to keep LogBrowser
> minimalistic.
> >>>>
> >>> makes sense :-)
> >>>
> >>> Cheers, Sergey
> >>>
> >>>> --
> >>>> Best regards,
> >>>> Tomasz Oponowicz
> >>>>
> >>>
> >>
> >> --
> >> Best regards,
> >> Tomasz Oponowicz
> >>
> >
>
--
Sergey Beryozkin
Application Integration Division of Talend <http://www.talend.com>
http://sberyozkin.blogspot.com
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
String text = columnDefinition.getContent(entry);
in SelectableTable.setData
is causing it,
and it's only happening when the Date column is filled in, the Level
and Message are fine.
Looks like the code co0nverting the Atom/http-format date is brittle a bit
cheers, Sergey
On Sun, Feb 27, 2011 at 5:10 PM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi Tomasz
>
> I've spent some time and it is some sort of the asynchronous exception
> which when captured shows the Date of the log entry. It is thrown when
> the ViewerViewImpl.setEntries(List<Entry> entries) is called. We'll
> track it down soon :-)
>
> Sergey
>
> On Sat, Feb 26, 2011 at 1:23 PM, Tomasz Oponowicz
> <to...@gmail.com> wrote:
>> Hi Sergey,
>>
>> On Fri, Feb 25, 2011 at 1:19 PM, Sergey Beryozkin <sb...@gmail.com> wrote:
>>> Hi Tomasz
>>>
>>> I've rerun the demo.
>>> First of all, the changes you've made recently have definitely made it
>>> look much better, thanks.
>>> Unfortunately, I'm still hitting this GWT ApplicationException when
>>> refreshing the endpoint :
>>> "
>>> Application Error
>>> Class$jcb
>>> 2011-02-25T11:46:55.078Z
>>> "
>>>
>>> Can this message, particularly Class$jcb, help somehow to identify the
>>> problem ? It does look like it's a platform/browser specific issue,
>>> I'm on Ubuntu 9, Firefox 3.6.13, but it would be good to get rid of it
>>> somehow.
>>> Just tried Chrome and it showed the same error but with the
>>> "Class$kbc" - it's probably the some gwt proxy...
>>>
>>> Can you please point to the code in the logbrowser project where the
>>> response from the remote endpoint is processed ? I will investigate...
>>
>> I will install similar environment as a virtual machine - I hope I
>> will reproduce this issue...
>>
>>> Few more comments. I agree, the way Tasks and Endpoints are currently
>>> shown is nice.
>>>
>>> - Can you please consider having both ManageEndpoints and Filter links
>>> located under the "Tasks" ? And have Endpoints shown first, with the
>>> "Tasks" in the bottom of the pane ? Ultimately, the user just wants to
>>> see the list of endpoints. Creating/deleting the endpoints and
>>> applying filters is critical but I'd just prefer the "Tasks" not to
>>> feature as the main activity of the LogBrowser users...
>>> - Filter dialog can not be closed at the moment, after it has been opened...
>>> - Probably makes sense to rename "Explore" to "Explorer" given that
>>> the "ManageEndpoints" panel has the "Back to Explorer" link...
>>>
>>> More comments inline...
>>>
>>>
>>>>> - Authentication: I've noticed there's AuthenticationRequired
>>>>> annotation attached to some of the BootstrapStorage methods - we
>>>>> really need to remove this annotation and for now just pop-up a login
>>>>> window on the browser start-up.
>>>>> Users will be configuring the LogBrowser application as part of the
>>>>> real deployments. So what would be good is to write the GWT client
>>>>> code such that it only pops up a window when the initial GET returns
>>>>> 401 - can you use CXF WebClient there and do 'Response r =
>>>>> webClient.get()' and if r.getStatus() == 401 then pop-up a login
>>>>> dialog ? We can deal with this issue later, when we have more time,
>>>>> and then we'll also decide whether to support https in cases when the
>>>>> authentication is needed or may be do the UT profile, we'll see...
>>>>
>>>> According to your list of tasks please consider also fallowing tasks:
>>>
>>> Thanks for this analysis...
>>>
>>>>
>>>> - Removing "Sign in" feature;
>>>> - Pros:
>>>> - Simplify implementation;
>>>> - Easy configuration for end user;
>>>> - Every company has got their own internal user
>>>> authentication system (LDAP, OpenID, internal SSO etc.);
>>>> - Even if LogBrowser doesn't contain any user authentication
>>>> system, it's still very easy to add integration with some
>>>> authentication system:
>>>> - Simply interceptor before request rich controller;
>>>> - Apache directives (of course if user use Apache
>>>> before Tomcat);
>>>> - Cons:
>>>> - I understand that feeds should be secured, but I think we
>>>> should rather concentrate on:
>>>> - HTTPS connection;
>>>> - password per feed (optional);
>>>>
>>>
>>> I think we are in agreement here. I'd like to propose:
>>> - remove the initial Sign-In dialog altogether
>>> - Remove SighIn and SignOut buttons
>>> - Remove AuthenticationRequired annotation from the coode
>>> - postpone dealing with the authentication issues - at the moment we
>>> just need to focus on making sure
>>> the browser is operational.
>>>
>>> After 2.4.0, we can enhance it for the authentication+HTTPS be supported.
>>>
>>> If you agree then please remove all the authentication-related code/settings...
>>>
>>>> - Removing storing user settings remotely on the servers;
>>>> - Pros:
>>>> - Simplify implementation;
>>>> - Easy configuration for end user;
>>>> - Very clear message - all settings are stored in browser
>>>> local storage. At the moment the logic it's to complicated. It depend
>>>> on situation we keep settings in memory, browser local storage or
>>>> bootstrap settings;
>>>> - Cons:
>>>> - When end user clear cache all settings will be removed;
>>>> - Settings are stored per browser. When you add something in
>>>> Firefox it won't be available in Chrome;
>>>>
>>>
>>> I think it makes sense to keep the list of endpoints and the filter
>>> properties the current user has created.
>>> *But*, these settings just need to shared across multiple restarts of
>>> the browser between all the users.
>>> This is because I don't really think it is realistic that one user
>>> will want to see EndpointA only, the other one, EndpointB, etc.
>>> So lets keep it - I'm not worried about many users using different
>>> browsers for checking the logs of the single server :-)
>>>
>>>> What do you think about these tasks? I'd like to keep LogBrowser minimalistic.
>>>>
>>> makes sense :-)
>>>
>>> Cheers, Sergey
>>>
>>>> --
>>>> Best regards,
>>>> Tomasz Oponowicz
>>>>
>>>
>>
>> --
>> Best regards,
>> Tomasz Oponowicz
>>
>
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz
I've spent some time and it is some sort of the asynchronous exception
which when captured shows the Date of the log entry. It is thrown when
the ViewerViewImpl.setEntries(List<Entry> entries) is called. We'll
track it down soon :-)
Sergey
On Sat, Feb 26, 2011 at 1:23 PM, Tomasz Oponowicz
<to...@gmail.com> wrote:
> Hi Sergey,
>
> On Fri, Feb 25, 2011 at 1:19 PM, Sergey Beryozkin <sb...@gmail.com> wrote:
>> Hi Tomasz
>>
>> I've rerun the demo.
>> First of all, the changes you've made recently have definitely made it
>> look much better, thanks.
>> Unfortunately, I'm still hitting this GWT ApplicationException when
>> refreshing the endpoint :
>> "
>> Application Error
>> Class$jcb
>> 2011-02-25T11:46:55.078Z
>> "
>>
>> Can this message, particularly Class$jcb, help somehow to identify the
>> problem ? It does look like it's a platform/browser specific issue,
>> I'm on Ubuntu 9, Firefox 3.6.13, but it would be good to get rid of it
>> somehow.
>> Just tried Chrome and it showed the same error but with the
>> "Class$kbc" - it's probably the some gwt proxy...
>>
>> Can you please point to the code in the logbrowser project where the
>> response from the remote endpoint is processed ? I will investigate...
>
> I will install similar environment as a virtual machine - I hope I
> will reproduce this issue...
>
>> Few more comments. I agree, the way Tasks and Endpoints are currently
>> shown is nice.
>>
>> - Can you please consider having both ManageEndpoints and Filter links
>> located under the "Tasks" ? And have Endpoints shown first, with the
>> "Tasks" in the bottom of the pane ? Ultimately, the user just wants to
>> see the list of endpoints. Creating/deleting the endpoints and
>> applying filters is critical but I'd just prefer the "Tasks" not to
>> feature as the main activity of the LogBrowser users...
>> - Filter dialog can not be closed at the moment, after it has been opened...
>> - Probably makes sense to rename "Explore" to "Explorer" given that
>> the "ManageEndpoints" panel has the "Back to Explorer" link...
>>
>> More comments inline...
>>
>>
>>>> - Authentication: I've noticed there's AuthenticationRequired
>>>> annotation attached to some of the BootstrapStorage methods - we
>>>> really need to remove this annotation and for now just pop-up a login
>>>> window on the browser start-up.
>>>> Users will be configuring the LogBrowser application as part of the
>>>> real deployments. So what would be good is to write the GWT client
>>>> code such that it only pops up a window when the initial GET returns
>>>> 401 - can you use CXF WebClient there and do 'Response r =
>>>> webClient.get()' and if r.getStatus() == 401 then pop-up a login
>>>> dialog ? We can deal with this issue later, when we have more time,
>>>> and then we'll also decide whether to support https in cases when the
>>>> authentication is needed or may be do the UT profile, we'll see...
>>>
>>> According to your list of tasks please consider also fallowing tasks:
>>
>> Thanks for this analysis...
>>
>>>
>>> - Removing "Sign in" feature;
>>> - Pros:
>>> - Simplify implementation;
>>> - Easy configuration for end user;
>>> - Every company has got their own internal user
>>> authentication system (LDAP, OpenID, internal SSO etc.);
>>> - Even if LogBrowser doesn't contain any user authentication
>>> system, it's still very easy to add integration with some
>>> authentication system:
>>> - Simply interceptor before request rich controller;
>>> - Apache directives (of course if user use Apache
>>> before Tomcat);
>>> - Cons:
>>> - I understand that feeds should be secured, but I think we
>>> should rather concentrate on:
>>> - HTTPS connection;
>>> - password per feed (optional);
>>>
>>
>> I think we are in agreement here. I'd like to propose:
>> - remove the initial Sign-In dialog altogether
>> - Remove SighIn and SignOut buttons
>> - Remove AuthenticationRequired annotation from the coode
>> - postpone dealing with the authentication issues - at the moment we
>> just need to focus on making sure
>> the browser is operational.
>>
>> After 2.4.0, we can enhance it for the authentication+HTTPS be supported.
>>
>> If you agree then please remove all the authentication-related code/settings...
>>
>>> - Removing storing user settings remotely on the servers;
>>> - Pros:
>>> - Simplify implementation;
>>> - Easy configuration for end user;
>>> - Very clear message - all settings are stored in browser
>>> local storage. At the moment the logic it's to complicated. It depend
>>> on situation we keep settings in memory, browser local storage or
>>> bootstrap settings;
>>> - Cons:
>>> - When end user clear cache all settings will be removed;
>>> - Settings are stored per browser. When you add something in
>>> Firefox it won't be available in Chrome;
>>>
>>
>> I think it makes sense to keep the list of endpoints and the filter
>> properties the current user has created.
>> *But*, these settings just need to shared across multiple restarts of
>> the browser between all the users.
>> This is because I don't really think it is realistic that one user
>> will want to see EndpointA only, the other one, EndpointB, etc.
>> So lets keep it - I'm not worried about many users using different
>> browsers for checking the logs of the single server :-)
>>
>>> What do you think about these tasks? I'd like to keep LogBrowser minimalistic.
>>>
>> makes sense :-)
>>
>> Cheers, Sergey
>>
>>> --
>>> Best regards,
>>> Tomasz Oponowicz
>>>
>>
>
> --
> Best regards,
> Tomasz Oponowicz
>
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi Sergey,
On Fri, Feb 25, 2011 at 1:19 PM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi Tomasz
>
> I've rerun the demo.
> First of all, the changes you've made recently have definitely made it
> look much better, thanks.
> Unfortunately, I'm still hitting this GWT ApplicationException when
> refreshing the endpoint :
> "
> Application Error
> Class$jcb
> 2011-02-25T11:46:55.078Z
> "
>
> Can this message, particularly Class$jcb, help somehow to identify the
> problem ? It does look like it's a platform/browser specific issue,
> I'm on Ubuntu 9, Firefox 3.6.13, but it would be good to get rid of it
> somehow.
> Just tried Chrome and it showed the same error but with the
> "Class$kbc" - it's probably the some gwt proxy...
>
> Can you please point to the code in the logbrowser project where the
> response from the remote endpoint is processed ? I will investigate...
I will install similar environment as a virtual machine - I hope I
will reproduce this issue...
> Few more comments. I agree, the way Tasks and Endpoints are currently
> shown is nice.
>
> - Can you please consider having both ManageEndpoints and Filter links
> located under the "Tasks" ? And have Endpoints shown first, with the
> "Tasks" in the bottom of the pane ? Ultimately, the user just wants to
> see the list of endpoints. Creating/deleting the endpoints and
> applying filters is critical but I'd just prefer the "Tasks" not to
> feature as the main activity of the LogBrowser users...
> - Filter dialog can not be closed at the moment, after it has been opened...
> - Probably makes sense to rename "Explore" to "Explorer" given that
> the "ManageEndpoints" panel has the "Back to Explorer" link...
>
> More comments inline...
>
>
>>> - Authentication: I've noticed there's AuthenticationRequired
>>> annotation attached to some of the BootstrapStorage methods - we
>>> really need to remove this annotation and for now just pop-up a login
>>> window on the browser start-up.
>>> Users will be configuring the LogBrowser application as part of the
>>> real deployments. So what would be good is to write the GWT client
>>> code such that it only pops up a window when the initial GET returns
>>> 401 - can you use CXF WebClient there and do 'Response r =
>>> webClient.get()' and if r.getStatus() == 401 then pop-up a login
>>> dialog ? We can deal with this issue later, when we have more time,
>>> and then we'll also decide whether to support https in cases when the
>>> authentication is needed or may be do the UT profile, we'll see...
>>
>> According to your list of tasks please consider also fallowing tasks:
>
> Thanks for this analysis...
>
>>
>> - Removing "Sign in" feature;
>> - Pros:
>> - Simplify implementation;
>> - Easy configuration for end user;
>> - Every company has got their own internal user
>> authentication system (LDAP, OpenID, internal SSO etc.);
>> - Even if LogBrowser doesn't contain any user authentication
>> system, it's still very easy to add integration with some
>> authentication system:
>> - Simply interceptor before request rich controller;
>> - Apache directives (of course if user use Apache
>> before Tomcat);
>> - Cons:
>> - I understand that feeds should be secured, but I think we
>> should rather concentrate on:
>> - HTTPS connection;
>> - password per feed (optional);
>>
>
> I think we are in agreement here. I'd like to propose:
> - remove the initial Sign-In dialog altogether
> - Remove SighIn and SignOut buttons
> - Remove AuthenticationRequired annotation from the coode
> - postpone dealing with the authentication issues - at the moment we
> just need to focus on making sure
> the browser is operational.
>
> After 2.4.0, we can enhance it for the authentication+HTTPS be supported.
>
> If you agree then please remove all the authentication-related code/settings...
>
>> - Removing storing user settings remotely on the servers;
>> - Pros:
>> - Simplify implementation;
>> - Easy configuration for end user;
>> - Very clear message - all settings are stored in browser
>> local storage. At the moment the logic it's to complicated. It depend
>> on situation we keep settings in memory, browser local storage or
>> bootstrap settings;
>> - Cons:
>> - When end user clear cache all settings will be removed;
>> - Settings are stored per browser. When you add something in
>> Firefox it won't be available in Chrome;
>>
>
> I think it makes sense to keep the list of endpoints and the filter
> properties the current user has created.
> *But*, these settings just need to shared across multiple restarts of
> the browser between all the users.
> This is because I don't really think it is realistic that one user
> will want to see EndpointA only, the other one, EndpointB, etc.
> So lets keep it - I'm not worried about many users using different
> browsers for checking the logs of the single server :-)
>
>> What do you think about these tasks? I'd like to keep LogBrowser minimalistic.
>>
> makes sense :-)
>
> Cheers, Sergey
>
>> --
>> Best regards,
>> Tomasz Oponowicz
>>
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi
>
> Few more comments. I agree, the way Tasks and Endpoints are currently
> shown is nice.
>
> - Can you please consider having both ManageEndpoints and Filter links
> located under the "Tasks" ? And have Endpoints shown first, with the
> "Tasks" in the bottom of the pane ? Ultimately, the user just wants to
> see the list of endpoints. Creating/deleting the endpoints and
> applying filters is critical but I'd just prefer the "Tasks" not to
> feature as the main activity of the LogBrowser users...
I'm looking at it again and I kind of like it, the way the pane looks
now, so I'll let it up to you to decide how the final layout should
look like. May be try to move the Tasks to the bottom of the pane and
if you like it then update the code, if not then lets leave the things
the way they are now at the moment :-). I'm sure we'll get some
feedback later on too...
thanks, Sergey
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz
I've rerun the demo.
First of all, the changes you've made recently have definitely made it
look much better, thanks.
Unfortunately, I'm still hitting this GWT ApplicationException when
refreshing the endpoint :
"
Application Error
Class$jcb
2011-02-25T11:46:55.078Z
"
Can this message, particularly Class$jcb, help somehow to identify the
problem ? It does look like it's a platform/browser specific issue,
I'm on Ubuntu 9, Firefox 3.6.13, but it would be good to get rid of it
somehow.
Just tried Chrome and it showed the same error but with the
"Class$kbc" - it's probably the some gwt proxy...
Can you please point to the code in the logbrowser project where the
response from the remote endpoint is processed ? I will investigate...
Few more comments. I agree, the way Tasks and Endpoints are currently
shown is nice.
- Can you please consider having both ManageEndpoints and Filter links
located under the "Tasks" ? And have Endpoints shown first, with the
"Tasks" in the bottom of the pane ? Ultimately, the user just wants to
see the list of endpoints. Creating/deleting the endpoints and
applying filters is critical but I'd just prefer the "Tasks" not to
feature as the main activity of the LogBrowser users...
- Filter dialog can not be closed at the moment, after it has been opened...
- Probably makes sense to rename "Explore" to "Explorer" given that
the "ManageEndpoints" panel has the "Back to Explorer" link...
More comments inline...
>> - Authentication: I've noticed there's AuthenticationRequired
>> annotation attached to some of the BootstrapStorage methods - we
>> really need to remove this annotation and for now just pop-up a login
>> window on the browser start-up.
>> Users will be configuring the LogBrowser application as part of the
>> real deployments. So what would be good is to write the GWT client
>> code such that it only pops up a window when the initial GET returns
>> 401 - can you use CXF WebClient there and do 'Response r =
>> webClient.get()' and if r.getStatus() == 401 then pop-up a login
>> dialog ? We can deal with this issue later, when we have more time,
>> and then we'll also decide whether to support https in cases when the
>> authentication is needed or may be do the UT profile, we'll see...
>
> According to your list of tasks please consider also fallowing tasks:
Thanks for this analysis...
>
> - Removing "Sign in" feature;
> - Pros:
> - Simplify implementation;
> - Easy configuration for end user;
> - Every company has got their own internal user
> authentication system (LDAP, OpenID, internal SSO etc.);
> - Even if LogBrowser doesn't contain any user authentication
> system, it's still very easy to add integration with some
> authentication system:
> - Simply interceptor before request rich controller;
> - Apache directives (of course if user use Apache
> before Tomcat);
> - Cons:
> - I understand that feeds should be secured, but I think we
> should rather concentrate on:
> - HTTPS connection;
> - password per feed (optional);
>
I think we are in agreement here. I'd like to propose:
- remove the initial Sign-In dialog altogether
- Remove SighIn and SignOut buttons
- Remove AuthenticationRequired annotation from the coode
- postpone dealing with the authentication issues - at the moment we
just need to focus on making sure
the browser is operational.
After 2.4.0, we can enhance it for the authentication+HTTPS be supported.
If you agree then please remove all the authentication-related code/settings...
> - Removing storing user settings remotely on the servers;
> - Pros:
> - Simplify implementation;
> - Easy configuration for end user;
> - Very clear message - all settings are stored in browser
> local storage. At the moment the logic it's to complicated. It depend
> on situation we keep settings in memory, browser local storage or
> bootstrap settings;
> - Cons:
> - When end user clear cache all settings will be removed;
> - Settings are stored per browser. When you add something in
> Firefox it won't be available in Chrome;
>
I think it makes sense to keep the list of endpoints and the filter
properties the current user has created.
*But*, these settings just need to shared across multiple restarts of
the browser between all the users.
This is because I don't really think it is realistic that one user
will want to see EndpointA only, the other one, EndpointB, etc.
So lets keep it - I'm not worried about many users using different
browsers for checking the logs of the single server :-)
> What do you think about these tasks? I'd like to keep LogBrowser minimalistic.
>
makes sense :-)
Cheers, Sergey
> --
> Best regards,
> Tomasz Oponowicz
>
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi Sergey,
On Thu, Feb 17, 2011 at 12:34 PM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi Tomasz
>>>>>
>>>>> 1. LogBrowser has a showstopper bug which I'm sure is due to the fact we
>>>>> just did not look at it for a while :-). Basically, when I go and generate
>>>>> the logs, and then click 'refresh' on the endpoint, the browser reports an
>>>>> exception.
>>
>> I can't reproduce this (however it doesn't mean that there isn't bug ;) ).
>>
>> To solidify my understanding, use case:
>>
>> 1) Run 'logbrowser' through sample;
>> 2) Go to 'http://localhost:9002/log/browser/LogBrowser.html';
>> 3) Add new endpoint with URL 'http://localhost:9002/log/logs';
>> 4) Select newly created endpoint;
>> 5) In new tab open 'http://localhost:9002/customer-service.html' and
>> add few 'customers';
>> 6) Go back to 'logbrowser';
>> 6) Click 'refresh' link;
>>
>> Expected:
>> List is refreshed and new items are shown.
>>
>> Environment:
>> Firefox 3.6.x
>>
>> Is this use case correct? Thanks for your help.
>>
>
> It is, I'll try again asap - definitely looks like a platform/browser
> specific issue, it's Ubuntu 9 + FireFox 3.6.13 which I upgraded to
> recently using the apt-get facility...So I'm not worried really then,
> but I'll try again, and on Windows too...
>
>> Maybe try with force refresh (i.e. "Ctrl + R"). However I have to add
>> revision number to static resources URL to avoid cache problems.
>>
>>> this is is the only main issue at the moment
>>>>>
>>>>> 2. Please move ManageEndpoints button either immediately above or below the
>>>>> Filter button
>>>>
>>>> You mean something similar to "original" layout - "Manage endpoint" is
>>>> "attached" to the bottom of the page?
>>>>
>>>
>>> I'm thinking that given we have a Filter button in the bottom of the
>>> left-side pane, it would be ok to have both 'Manage Endpoints' and
>>> 'Filters' co-located...
>>>
>>>>> 3. Remove the Tasks and Endpoints buttons/entries, lets have it the way you
>>>>> did it originally. We only need to see the list of endpoints which will be
>>>>> added via "ManageEndpoints", the explorer style is just too complex.
According to 2, 3:
I left old layout, but I fixed CSS. It looks really nice now.
We need layout which is easy to extend. Section "tasks" give us
ability to add new actions later. I think it makes sense to keep this
layout.
>>>>> 4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel and
>>>>> overlaps with the "Settings" entry,
>>
>> Fixed.
>
> Super
>
>>
>>>> Looks like 2, 3, 4 are layout problems... I have to clean up this. At
>>>> the moment we are using mix of CSS and "table layout". I'm not CSS
>>>> expert so I will fix it by moving problematic parts to "table layout".
In general I committed many layout fixes.
>> I noticed that when you use latest Chrome or Safari, list of logs and
>> detailed view isn't shown. I consider this as a blocker. I will
>> prepare fix for this ASAP.
Fixed.
> May be it's somehow related to what I see with the application
> exception above...
>
> What I will need to do is to provide an abstract utility
> ReadableStorage implementation which can be easily overridden to have
> the file-based logs viewable. I don't think we can make LogBrowser
> perfect and feature-complete by the 2.4.0 is released, but hope users
> will give it a try anyway...
>
> Few other "would be nice to fix" issues - definitely not show stoppers :
> - LogBrowser has the 'embedded' providers, one for servicing the
> gzipped file and another one for unqualified JSON, this would be nice
> to move to the demo's Application (JSONProvider can be configured
> directly to drop the namespaces) - I can look into it
> - Authentication: I've noticed there's AuthenticationRequired
> annotation attached to some of the BootstrapStorage methods - we
> really need to remove this annotation and for now just pop-up a login
> window on the browser start-up.
> Users will be configuring the LogBrowser application as part of the
> real deployments. So what would be good is to write the GWT client
> code such that it only pops up a window when the initial GET returns
> 401 - can you use CXF WebClient there and do 'Response r =
> webClient.get()' and if r.getStatus() == 401 then pop-up a login
> dialog ? We can deal with this issue later, when we have more time,
> and then we'll also decide whether to support https in cases when the
> authentication is needed or may be do the UT profile, we'll see...
According to your list of tasks please consider also fallowing tasks:
- Removing "Sign in" feature;
- Pros:
- Simplify implementation;
- Easy configuration for end user;
- Every company has got their own internal user
authentication system (LDAP, OpenID, internal SSO etc.);
- Even if LogBrowser doesn't contain any user authentication
system, it's still very easy to add integration with some
authentication system:
- Simply interceptor before request rich controller;
- Apache directives (of course if user use Apache
before Tomcat);
- Cons:
- I understand that feeds should be secured, but I think we
should rather concentrate on:
- HTTPS connection;
- password per feed (optional);
- Removing storing user settings remotely on the servers;
- Pros:
- Simplify implementation;
- Easy configuration for end user;
- Very clear message - all settings are stored in browser
local storage. At the moment the logic it's to complicated. It depend
on situation we keep settings in memory, browser local storage or
bootstrap settings;
- Cons:
- When end user clear cache all settings will be removed;
- Settings are stored per browser. When you add something in
Firefox it won't be available in Chrome;
What do you think about these tasks? I'd like to keep LogBrowser minimalistic.
> Thanks, Sergey.
>
>>
>> --
>> Best regards,
>> Tomasz Oponowicz
>>
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz -
I've seen you doing a bunch of fixes - thanks a million.
I'm going to start focusing on the browser too very soon - just
finalizing the last few improvements for the transformation feature.
Hope we'll get the shiny log browser ready in time for 2.4.0 :-)
talk to you later
Sergey
On Thu, Feb 17, 2011 at 11:34 AM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi Tomasz
>>>>>
>>>>> 1. LogBrowser has a showstopper bug which I'm sure is due to the fact we
>>>>> just did not look at it for a while :-). Basically, when I go and generate
>>>>> the logs, and then click 'refresh' on the endpoint, the browser reports an
>>>>> exception.
>>
>> I can't reproduce this (however it doesn't mean that there isn't bug ;) ).
>>
>> To solidify my understanding, use case:
>>
>> 1) Run 'logbrowser' through sample;
>> 2) Go to 'http://localhost:9002/log/browser/LogBrowser.html';
>> 3) Add new endpoint with URL 'http://localhost:9002/log/logs';
>> 4) Select newly created endpoint;
>> 5) In new tab open 'http://localhost:9002/customer-service.html' and
>> add few 'customers';
>> 6) Go back to 'logbrowser';
>> 6) Click 'refresh' link;
>>
>> Expected:
>> List is refreshed and new items are shown.
>>
>> Environment:
>> Firefox 3.6.x
>>
>> Is this use case correct? Thanks for your help.
>>
>
> It is, I'll try again asap - definitely looks like a platform/browser
> specific issue, it's Ubuntu 9 + FireFox 3.6.13 which I upgraded to
> recently using the apt-get facility...So I'm not worried really then,
> but I'll try again, and on Windows too...
>
>> Maybe try with force refresh (i.e. "Ctrl + R"). However I have to add
>> revision number to static resources URL to avoid cache problems.
>>
>>> this is is the only main issue at the moment
>>>>>
>>>>> 2. Please move ManageEndpoints button either immediately above or below the
>>>>> Filter button
>>>>
>>>> You mean something similar to "original" layout - "Manage endpoint" is
>>>> "attached" to the bottom of the page?
>>>>
>>>
>>> I'm thinking that given we have a Filter button in the bottom of the
>>> left-side pane, it would be ok to have both 'Manage Endpoints' and
>>> 'Filters' co-located...
>>>
>>>>> 3. Remove the Tasks and Endpoints buttons/entries, lets have it the way you
>>>>> did it originally. We only need to see the list of endpoints which will be
>>>>> added via "ManageEndpoints", the explorer style is just too complex.
>>>>>
>>>>> 4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel and
>>>>> overlaps with the "Settings" entry,
>>
>> Fixed.
>
> Super
>
>>
>>>> Looks like 2, 3, 4 are layout problems... I have to clean up this. At
>>>> the moment we are using mix of CSS and "table layout". I'm not CSS
>>>> expert so I will fix it by moving problematic parts to "table layout".
>>
>> I noticed that when you use latest Chrome or Safari, list of logs and
>> detailed view isn't shown. I consider this as a blocker. I will
>> prepare fix for this ASAP.
>
> May be it's somehow related to what I see with the application
> exception above...
>
> What I will need to do is to provide an abstract utility
> ReadableStorage implementation which can be easily overridden to have
> the file-based logs viewable. I don't think we can make LogBrowser
> perfect and feature-complete by the 2.4.0 is released, but hope users
> will give it a try anyway...
>
> Few other "would be nice to fix" issues - definitely not show stoppers :
> - LogBrowser has the 'embedded' providers, one for servicing the
> gzipped file and another one for unqualified JSON, this would be nice
> to move to the demo's Application (JSONProvider can be configured
> directly to drop the namespaces) - I can look into it
> - Authentication: I've noticed there's AuthenticationRequired
> annotation attached to some of the BootstrapStorage methods - we
> really need to remove this annotation and for now just pop-up a login
> window on the browser start-up.
> Users will be configuring the LogBrowser application as part of the
> real deployments. So what would be good is to write the GWT client
> code such that it only pops up a window when the initial GET returns
> 401 - can you use CXF WebClient there and do 'Response r =
> webClient.get()' and if r.getStatus() == 401 then pop-up a login
> dialog ? We can deal with this issue later, when we have more time,
> and then we'll also decide whether to support https in cases when the
> authentication is needed or may be do the UT profile, we'll see...
>
> Thanks, Sergey.
>
>>
>> --
>> Best regards,
>> Tomasz Oponowicz
>>
>
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz
>>>>
>>>> 1. LogBrowser has a showstopper bug which I'm sure is due to the fact we
>>>> just did not look at it for a while :-). Basically, when I go and generate
>>>> the logs, and then click 'refresh' on the endpoint, the browser reports an
>>>> exception.
>
> I can't reproduce this (however it doesn't mean that there isn't bug ;) ).
>
> To solidify my understanding, use case:
>
> 1) Run 'logbrowser' through sample;
> 2) Go to 'http://localhost:9002/log/browser/LogBrowser.html';
> 3) Add new endpoint with URL 'http://localhost:9002/log/logs';
> 4) Select newly created endpoint;
> 5) In new tab open 'http://localhost:9002/customer-service.html' and
> add few 'customers';
> 6) Go back to 'logbrowser';
> 6) Click 'refresh' link;
>
> Expected:
> List is refreshed and new items are shown.
>
> Environment:
> Firefox 3.6.x
>
> Is this use case correct? Thanks for your help.
>
It is, I'll try again asap - definitely looks like a platform/browser
specific issue, it's Ubuntu 9 + FireFox 3.6.13 which I upgraded to
recently using the apt-get facility...So I'm not worried really then,
but I'll try again, and on Windows too...
> Maybe try with force refresh (i.e. "Ctrl + R"). However I have to add
> revision number to static resources URL to avoid cache problems.
>
>> this is is the only main issue at the moment
>>>>
>>>> 2. Please move ManageEndpoints button either immediately above or below the
>>>> Filter button
>>>
>>> You mean something similar to "original" layout - "Manage endpoint" is
>>> "attached" to the bottom of the page?
>>>
>>
>> I'm thinking that given we have a Filter button in the bottom of the
>> left-side pane, it would be ok to have both 'Manage Endpoints' and
>> 'Filters' co-located...
>>
>>>> 3. Remove the Tasks and Endpoints buttons/entries, lets have it the way you
>>>> did it originally. We only need to see the list of endpoints which will be
>>>> added via "ManageEndpoints", the explorer style is just too complex.
>>>>
>>>> 4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel and
>>>> overlaps with the "Settings" entry,
>
> Fixed.
Super
>
>>> Looks like 2, 3, 4 are layout problems... I have to clean up this. At
>>> the moment we are using mix of CSS and "table layout". I'm not CSS
>>> expert so I will fix it by moving problematic parts to "table layout".
>
> I noticed that when you use latest Chrome or Safari, list of logs and
> detailed view isn't shown. I consider this as a blocker. I will
> prepare fix for this ASAP.
May be it's somehow related to what I see with the application
exception above...
What I will need to do is to provide an abstract utility
ReadableStorage implementation which can be easily overridden to have
the file-based logs viewable. I don't think we can make LogBrowser
perfect and feature-complete by the 2.4.0 is released, but hope users
will give it a try anyway...
Few other "would be nice to fix" issues - definitely not show stoppers :
- LogBrowser has the 'embedded' providers, one for servicing the
gzipped file and another one for unqualified JSON, this would be nice
to move to the demo's Application (JSONProvider can be configured
directly to drop the namespaces) - I can look into it
- Authentication: I've noticed there's AuthenticationRequired
annotation attached to some of the BootstrapStorage methods - we
really need to remove this annotation and for now just pop-up a login
window on the browser start-up.
Users will be configuring the LogBrowser application as part of the
real deployments. So what would be good is to write the GWT client
code such that it only pops up a window when the initial GET returns
401 - can you use CXF WebClient there and do 'Response r =
webClient.get()' and if r.getStatus() == 401 then pop-up a login
dialog ? We can deal with this issue later, when we have more time,
and then we'll also decide whether to support https in cases when the
authentication is needed or may be do the UT profile, we'll see...
Thanks, Sergey.
>
> --
> Best regards,
> Tomasz Oponowicz
>
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi Sergey,
On Tue, Feb 15, 2011 at 10:50 AM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi Tomasz
>
> It's really nice to hear from you again :-)
>
>
>>> Here is the initial list of issues which I need some help with :
>>>
>>> 1. LogBrowser has a showstopper bug which I'm sure is due to the fact we
>>> just did not look at it for a while :-). Basically, when I go and generate
>>> the logs, and then click 'refresh' on the endpoint, the browser reports an
>>> exception.
I can't reproduce this (however it doesn't mean that there isn't bug ;) ).
To solidify my understanding, use case:
1) Run 'logbrowser' through sample;
2) Go to 'http://localhost:9002/log/browser/LogBrowser.html';
3) Add new endpoint with URL 'http://localhost:9002/log/logs';
4) Select newly created endpoint;
5) In new tab open 'http://localhost:9002/customer-service.html' and
add few 'customers';
6) Go back to 'logbrowser';
6) Click 'refresh' link;
Expected:
List is refreshed and new items are shown.
Environment:
Firefox 3.6.x
Is this use case correct? Thanks for your help.
Maybe try with force refresh (i.e. "Ctrl + R"). However I have to add
revision number to static resources URL to avoid cache problems.
> this is is the only main issue at the moment
>>>
>>> 2. Please move ManageEndpoints button either immediately above or below the
>>> Filter button
>>
>> You mean something similar to "original" layout - "Manage endpoint" is
>> "attached" to the bottom of the page?
>>
>
> I'm thinking that given we have a Filter button in the bottom of the
> left-side pane, it would be ok to have both 'Manage Endpoints' and
> 'Filters' co-located...
>
>>> 3. Remove the Tasks and Endpoints buttons/entries, lets have it the way you
>>> did it originally. We only need to see the list of endpoints which will be
>>> added via "ManageEndpoints", the explorer style is just too complex.
>>>
>>> 4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel and
>>> overlaps with the "Settings" entry,
Fixed.
>> Looks like 2, 3, 4 are layout problems... I have to clean up this. At
>> the moment we are using mix of CSS and "table layout". I'm not CSS
>> expert so I will fix it by moving problematic parts to "table layout".
I noticed that when you use latest Chrome or Safari, list of logs and
detailed view isn't shown. I consider this as a blocker. I will
prepare fix for this ASAP.
>
> Please give it a try. The left pane has the "Explorer" title which is
> fine, but then the actual space is a bit 'busy', it would be nice to
> have a list of endpoints there only, with the 'Manage Endpoints' and
> 'Filters' being collapsed in the bottom of the pane...
>
>>> 5. In ManageEndpoints please change "back to browse" to "Back to Explorer"
>>
>> Fixed.
>
> I've seen the merge - and I know now how to do it too :-)
>>
>>> 6. We will need to think what to do with the jscript library added to the
>>> demo - we will have to link to its via remore URI or simple do a simpler
>>> plain HTML form - not a priority right now.
>
> If linking to the remote jscript.js proves to be problematic then I
> can just introduce a CXF JAX-WS & JAX-RS client invoking in the
> endpoint and generating the logs. Or we can have a plain form without
> trying to analyze the response...
>
> thanks, Sergey
>
>>>
>>> Hope it's not a big list :-). We will catch up on the GWT dev a bit and I
>>> also hope other users with the GWT experience contribute in time. What I
>>> will do before 2.4 is look at the browser code and see what I can refactor
>>> but also add some helpers to the AtomPullServer so that users can use it for
>>> reading from the existing file logs...
>>>
>>> Thanks, Sergey
>>>
>>>
>>>
>>
>> --
>> Best regards,
>> Tomasz Oponowicz
>>
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Hi Tomasz
It's really nice to hear from you again :-)
>> Here is the initial list of issues which I need some help with :
>>
>> 1. LogBrowser has a showstopper bug which I'm sure is due to the fact we
>> just did not look at it for a while :-). Basically, when I go and generate
>> the logs, and then click 'refresh' on the endpoint, the browser reports an
>> exception.
this is is the only main issue at the moment
>>
>> 2. Please move ManageEndpoints button either immediately above or below the
>> Filter button
>
> You mean something similar to "original" layout - "Manage endpoint" is
> "attached" to the bottom of the page?
>
I'm thinking that given we have a Filter button in the bottom of the
left-side pane, it would be ok to have both 'Manage Endpoints' and
'Filters' co-located...
>> 3. Remove the Tasks and Endpoints buttons/entries, lets have it the way you
>> did it originally. We only need to see the list of endpoints which will be
>> added via "ManageEndpoints", the explorer style is just too complex.
>>
>> 4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel and
>> overlaps with the "Settings" entry,
>
> Looks like 2, 3, 4 are layout problems... I have to clean up this. At
> the moment we are using mix of CSS and "table layout". I'm not CSS
> expert so I will fix it by moving problematic parts to "table layout".
>
Please give it a try. The left pane has the "Explorer" title which is
fine, but then the actual space is a bit 'busy', it would be nice to
have a list of endpoints there only, with the 'Manage Endpoints' and
'Filters' being collapsed in the bottom of the pane...
>> 5. In ManageEndpoints please change "back to browse" to "Back to Explorer"
>
> Fixed.
I've seen the merge - and I know now how to do it too :-)
>
>> 6. We will need to think what to do with the jscript library added to the
>> demo - we will have to link to its via remore URI or simple do a simpler
>> plain HTML form - not a priority right now.
If linking to the remote jscript.js proves to be problematic then I
can just introduce a CXF JAX-WS & JAX-RS client invoking in the
endpoint and generating the logs. Or we can have a plain form without
trying to analyze the response...
thanks, Sergey
>>
>> Hope it's not a big list :-). We will catch up on the GWT dev a bit and I
>> also hope other users with the GWT experience contribute in time. What I
>> will do before 2.4 is look at the browser code and see what I can refactor
>> but also add some helpers to the AtomPullServer so that users can use it for
>> reading from the existing file logs...
>>
>> Thanks, Sergey
>>
>>
>>
>
> --
> Best regards,
> Tomasz Oponowicz
>
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi Sergey,
On Fri, Feb 4, 2011 at 6:55 PM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Hi all,
>
> I have just moved the LogBrowser project Tomasz was working upon as part of
> GSOC 2010 to the trunk. I've done few minor changes before moving it in.
>
> - rt/management-web has a new package,
> org.apache.cxf.management.web.browser* which contains the code for the
> browser. It was org.apache.cxf.management.web.logging.logbrowser* in the
> sandbox but I thought we might enhance it in the future to work
> PersistingIn/Out interceptors in the rt/management so decided to simplify
> the package name.
>
> - samples/logbrowser is the demo showing the LogBrowser in action.
>
> It may make sense even move the whole rt/management-web into a CXF
> subproject, we'll see. For now, I'd like to have the browser polished and
> then we can decide.
>
> The GWT compilation will happen as part of building the rt/management-web,
> so the final jar will actually have the generated GWT code and resources
> included. We may consider doing it during the demo run, but again, lets try
> to polish the browser first.
>
> Tomasz, first of all congrats, I think it is good that your project is
> moving to the trunk.
> I'd appreciate if you could help me a bit with the initial clean-up, hope
> you'll find some time. You are a CXF committer now so you can do changes on
> the trunk...
>
> Here is the initial list of issues which I need some help with :
>
> 1. LogBrowser has a showstopper bug which I'm sure is due to the fact we
> just did not look at it for a while :-). Basically, when I go and generate
> the logs, and then click 'refresh' on the endpoint, the browser reports an
> exception.
> The AtomPullServer is functioning ok, Firefox subscribes to the Atom Feed
> with no probs.
>
> 2. Please move ManageEndpoints button either immediately above or below the
> Filter button
You mean something similar to "original" layout - "Manage endpoint" is
"attached" to the bottom of the page?
> 3. Remove the Tasks and Endpoints buttons/entries, lets have it the way you
> did it originally. We only need to see the list of endpoints which will be
> added via "ManageEndpoints", the explorer style is just too complex.
>
> 4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel and
> overlaps with the "Settings" entry,
Looks like 2, 3, 4 are layout problems... I have to clean up this. At
the moment we are using mix of CSS and "table layout". I'm not CSS
expert so I will fix it by moving problematic parts to "table layout".
> 5. In ManageEndpoints please change "back to browse" to "Back to Explorer"
Fixed.
> 6. We will need to think what to do with the jscript library added to the
> demo - we will have to link to its via remore URI or simple do a simpler
> plain HTML form - not a priority right now.
>
> Hope it's not a big list :-). We will catch up on the GWT dev a bit and I
> also hope other users with the GWT experience contribute in time. What I
> will do before 2.4 is look at the browser code and see what I can refactor
> but also add some helpers to the AtomPullServer so that users can use it for
> reading from the existing file logs...
>
> Thanks, Sergey
>
>
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Hi,
Thanks Daniel for a fix. Looks fine - error dialog works without issue.
On Fri, Feb 4, 2011 at 10:54 PM, Daniel Kulp <dk...@apache.org> wrote:
> On Friday 04 February 2011 4:29:12 pm Daniel Kulp wrote:
>> On Friday 04 February 2011 4:24:39 pm Sergey Beryozkin wrote:
>> > Tomasz, I had to comment out few lines in the rt/management-web/pom.xml
>> > (excluded the browser package from the compilation).
>> >
>> > I won't have time in the next couple of weeks to push the browser
>> > subpackage to a subproject, so for a while I suggest
>> > we work on it in our own snapshots, and then, before committing the
>> > changes we exclude the browser subpackage in the pom.xml.
>>
>> The only problematic class is the ErrorDialog. Is there a way to get that
>> reworked to not use the extgui stuff? That would fix the issue just
>> fine.
>
> I've gone ahead and made ErrorDialog just extend PopUp for now. I'm really
> not sure what that will do, but it at least now compiles and such.
>
> Dan
>
>
>
>> Dan
>>
>> > thanks, Sergey
>> >
>> > On Fri, Feb 4, 2011 at 5:55 PM, Sergey Beryozkin
>>
>> <sb...@gmail.com>wrote:
>> > > Hi all,
>> > >
>> > > I have just moved the LogBrowser project Tomasz was working upon as
>> > > part of GSOC 2010 to the trunk. I've done few minor changes before
>> > > moving it in.
>> > >
>> > > - rt/management-web has a new package,
>> > > org.apache.cxf.management.web.browser* which contains the code for the
>> > > browser. It was org.apache.cxf.management.web.logging.logbrowser* in
>> > > the sandbox but I thought we might enhance it in the future to work
>> > > PersistingIn/Out interceptors in the rt/management so decided to
>> > > simplify the package name.
>> > >
>> > > - samples/logbrowser is the demo showing the LogBrowser in action.
>> > >
>> > > It may make sense even move the whole rt/management-web into a CXF
>> > > subproject, we'll see. For now, I'd like to have the browser polished
>> > > and then we can decide.
>> > >
>> > > The GWT compilation will happen as part of building the
>> > > rt/management-web, so the final jar will actually have the generated
>> > > GWT code and resources included. We may consider doing it during the
>> > > demo run, but again, lets try to polish the browser first.
>> > >
>> > > Tomasz, first of all congrats, I think it is good that your project is
>> > > moving to the trunk.
>> > > I'd appreciate if you could help me a bit with the initial clean-up,
>> > > hope you'll find some time. You are a CXF committer now so you can do
>> > > changes on the trunk...
>> > >
>> > > Here is the initial list of issues which I need some help with :
>> > >
>> > > 1. LogBrowser has a showstopper bug which I'm sure is due to the fact
>> > > we just did not look at it for a while :-). Basically, when I go and
>> > > generate the logs, and then click 'refresh' on the endpoint, the
>> > > browser reports an exception.
>> > > The AtomPullServer is functioning ok, Firefox subscribes to the Atom
>> > > Feed with no probs.
>> > >
>> > > 2. Please move ManageEndpoints button either immediately above or below
>> > > the Filter button
>> > >
>> > > 3. Remove the Tasks and Endpoints buttons/entries, lets have it the way
>> > > you did it originally. We only need to see the list of endpoints which
>> > > will be added via "ManageEndpoints", the explorer style is just too
>> > > complex.
>> > >
>> > > 4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel
>> > > and overlaps with the "Settings" entry,
>> > >
>> > > 5. In ManageEndpoints please change "back to browse" to "Back to
>> > > Explorer"
>> > >
>> > >
>> > > 6. We will need to think what to do with the jscript library added to
>> > > the demo - we will have to link to its via remore URI or simple do a
>> > > simpler plain HTML form - not a priority right now.
>> > >
>> > > Hope it's not a big list :-). We will catch up on the GWT dev a bit and
>> > > I also hope other users with the GWT experience contribute in time.
>> > > What I will do before 2.4 is look at the browser code and see what I
>> > > can refactor but also add some helpers to the AtomPullServer so that
>> > > users can use it for reading from the existing file logs...
>> > >
>> > > Thanks, Sergey
>
> --
> Daniel Kulp
> dkulp@apache.org
> http://dankulp.com/blog
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Tomasz Oponowicz <to...@gmail.com>.
Sorry for long silence... I'll look at all issues/problems
On Sun, Feb 6, 2011 at 6:34 PM, Sergey Beryozkin <sb...@gmail.com> wrote:
> Thanks Dan, this also let me remove the repository tag...
> Cheers, Sergey
>
> On Fri, Feb 4, 2011 at 9:54 PM, Daniel Kulp <dk...@apache.org> wrote:
>>
>> On Friday 04 February 2011 4:29:12 pm Daniel Kulp wrote:
>> > On Friday 04 February 2011 4:24:39 pm Sergey Beryozkin wrote:
>> > > Tomasz, I had to comment out few lines in the
>> > > rt/management-web/pom.xml
>> > > (excluded the browser package from the compilation).
>> > >
>> > > I won't have time in the next couple of weeks to push the browser
>> > > subpackage to a subproject, so for a while I suggest
>> > > we work on it in our own snapshots, and then, before committing the
>> > > changes we exclude the browser subpackage in the pom.xml.
>> >
>> > The only problematic class is the ErrorDialog. Is there a way to get
>> > that
>> > reworked to not use the extgui stuff? That would fix the issue just
>> > fine.
>>
>> I've gone ahead and made ErrorDialog just extend PopUp for now. I'm
>> really
>> not sure what that will do, but it at least now compiles and such.
>>
>> Dan
>>
>>
>>
>> > Dan
>> >
>> > > thanks, Sergey
>> > >
>> > > On Fri, Feb 4, 2011 at 5:55 PM, Sergey Beryozkin
>> >
>> > <sb...@gmail.com>wrote:
>> > > > Hi all,
>> > > >
>> > > > I have just moved the LogBrowser project Tomasz was working upon as
>> > > > part of GSOC 2010 to the trunk. I've done few minor changes before
>> > > > moving it in.
>> > > >
>> > > > - rt/management-web has a new package,
>> > > > org.apache.cxf.management.web.browser* which contains the code for
>> > > > the
>> > > > browser. It was org.apache.cxf.management.web.logging.logbrowser* in
>> > > > the sandbox but I thought we might enhance it in the future to work
>> > > > PersistingIn/Out interceptors in the rt/management so decided to
>> > > > simplify the package name.
>> > > >
>> > > > - samples/logbrowser is the demo showing the LogBrowser in action.
>> > > >
>> > > > It may make sense even move the whole rt/management-web into a CXF
>> > > > subproject, we'll see. For now, I'd like to have the browser
>> > > > polished
>> > > > and then we can decide.
>> > > >
>> > > > The GWT compilation will happen as part of building the
>> > > > rt/management-web, so the final jar will actually have the generated
>> > > > GWT code and resources included. We may consider doing it during the
>> > > > demo run, but again, lets try to polish the browser first.
>> > > >
>> > > > Tomasz, first of all congrats, I think it is good that your project
>> > > > is
>> > > > moving to the trunk.
>> > > > I'd appreciate if you could help me a bit with the initial clean-up,
>> > > > hope you'll find some time. You are a CXF committer now so you can
>> > > > do
>> > > > changes on the trunk...
>> > > >
>> > > > Here is the initial list of issues which I need some help with :
>> > > >
>> > > > 1. LogBrowser has a showstopper bug which I'm sure is due to the
>> > > > fact
>> > > > we just did not look at it for a while :-). Basically, when I go and
>> > > > generate the logs, and then click 'refresh' on the endpoint, the
>> > > > browser reports an exception.
>> > > > The AtomPullServer is functioning ok, Firefox subscribes to the Atom
>> > > > Feed with no probs.
>> > > >
>> > > > 2. Please move ManageEndpoints button either immediately above or
>> > > > below
>> > > > the Filter button
>> > > >
>> > > > 3. Remove the Tasks and Endpoints buttons/entries, lets have it the
>> > > > way
>> > > > you did it originally. We only need to see the list of endpoints
>> > > > which
>> > > > will be added via "ManageEndpoints", the explorer style is just too
>> > > > complex.
>> > > >
>> > > > 4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel
>> > > > and overlaps with the "Settings" entry,
>> > > >
>> > > > 5. In ManageEndpoints please change "back to browse" to "Back to
>> > > > Explorer"
>> > > >
>> > > >
>> > > > 6. We will need to think what to do with the jscript library added
>> > > > to
>> > > > the demo - we will have to link to its via remore URI or simple do a
>> > > > simpler plain HTML form - not a priority right now.
>> > > >
>> > > > Hope it's not a big list :-). We will catch up on the GWT dev a bit
>> > > > and
>> > > > I also hope other users with the GWT experience contribute in time.
>> > > > What I will do before 2.4 is look at the browser code and see what I
>> > > > can refactor but also add some helpers to the AtomPullServer so that
>> > > > users can use it for reading from the existing file logs...
>> > > >
>> > > > Thanks, Sergey
>>
>> --
>> Daniel Kulp
>> dkulp@apache.org
>> http://dankulp.com/blog
>
>
--
Best regards,
Tomasz Oponowicz
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Thanks Dan, this also let me remove the repository tag...
Cheers, Sergey
On Fri, Feb 4, 2011 at 9:54 PM, Daniel Kulp <dk...@apache.org> wrote:
> On Friday 04 February 2011 4:29:12 pm Daniel Kulp wrote:
> > On Friday 04 February 2011 4:24:39 pm Sergey Beryozkin wrote:
> > > Tomasz, I had to comment out few lines in the rt/management-web/pom.xml
> > > (excluded the browser package from the compilation).
> > >
> > > I won't have time in the next couple of weeks to push the browser
> > > subpackage to a subproject, so for a while I suggest
> > > we work on it in our own snapshots, and then, before committing the
> > > changes we exclude the browser subpackage in the pom.xml.
> >
> > The only problematic class is the ErrorDialog. Is there a way to get
> that
> > reworked to not use the extgui stuff? That would fix the issue just
> > fine.
>
> I've gone ahead and made ErrorDialog just extend PopUp for now. I'm really
> not sure what that will do, but it at least now compiles and such.
>
> Dan
>
>
>
> > Dan
> >
> > > thanks, Sergey
> > >
> > > On Fri, Feb 4, 2011 at 5:55 PM, Sergey Beryozkin
> >
> > <sb...@gmail.com>wrote:
> > > > Hi all,
> > > >
> > > > I have just moved the LogBrowser project Tomasz was working upon as
> > > > part of GSOC 2010 to the trunk. I've done few minor changes before
> > > > moving it in.
> > > >
> > > > - rt/management-web has a new package,
> > > > org.apache.cxf.management.web.browser* which contains the code for
> the
> > > > browser. It was org.apache.cxf.management.web.logging.logbrowser* in
> > > > the sandbox but I thought we might enhance it in the future to work
> > > > PersistingIn/Out interceptors in the rt/management so decided to
> > > > simplify the package name.
> > > >
> > > > - samples/logbrowser is the demo showing the LogBrowser in action.
> > > >
> > > > It may make sense even move the whole rt/management-web into a CXF
> > > > subproject, we'll see. For now, I'd like to have the browser polished
> > > > and then we can decide.
> > > >
> > > > The GWT compilation will happen as part of building the
> > > > rt/management-web, so the final jar will actually have the generated
> > > > GWT code and resources included. We may consider doing it during the
> > > > demo run, but again, lets try to polish the browser first.
> > > >
> > > > Tomasz, first of all congrats, I think it is good that your project
> is
> > > > moving to the trunk.
> > > > I'd appreciate if you could help me a bit with the initial clean-up,
> > > > hope you'll find some time. You are a CXF committer now so you can
> do
> > > > changes on the trunk...
> > > >
> > > > Here is the initial list of issues which I need some help with :
> > > >
> > > > 1. LogBrowser has a showstopper bug which I'm sure is due to the fact
> > > > we just did not look at it for a while :-). Basically, when I go and
> > > > generate the logs, and then click 'refresh' on the endpoint, the
> > > > browser reports an exception.
> > > > The AtomPullServer is functioning ok, Firefox subscribes to the Atom
> > > > Feed with no probs.
> > > >
> > > > 2. Please move ManageEndpoints button either immediately above or
> below
> > > > the Filter button
> > > >
> > > > 3. Remove the Tasks and Endpoints buttons/entries, lets have it the
> way
> > > > you did it originally. We only need to see the list of endpoints
> which
> > > > will be added via "ManageEndpoints", the explorer style is just too
> > > > complex.
> > > >
> > > > 4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel
> > > > and overlaps with the "Settings" entry,
> > > >
> > > > 5. In ManageEndpoints please change "back to browse" to "Back to
> > > > Explorer"
> > > >
> > > >
> > > > 6. We will need to think what to do with the jscript library added to
> > > > the demo - we will have to link to its via remore URI or simple do a
> > > > simpler plain HTML form - not a priority right now.
> > > >
> > > > Hope it's not a big list :-). We will catch up on the GWT dev a bit
> and
> > > > I also hope other users with the GWT experience contribute in time.
> > > > What I will do before 2.4 is look at the browser code and see what I
> > > > can refactor but also add some helpers to the AtomPullServer so that
> > > > users can use it for reading from the existing file logs...
> > > >
> > > > Thanks, Sergey
>
> --
> Daniel Kulp
> dkulp@apache.org
> http://dankulp.com/blog
>
Re: LogBrowser project is on the trunk
Posted by Daniel Kulp <dk...@apache.org>.
On Friday 04 February 2011 4:29:12 pm Daniel Kulp wrote:
> On Friday 04 February 2011 4:24:39 pm Sergey Beryozkin wrote:
> > Tomasz, I had to comment out few lines in the rt/management-web/pom.xml
> > (excluded the browser package from the compilation).
> >
> > I won't have time in the next couple of weeks to push the browser
> > subpackage to a subproject, so for a while I suggest
> > we work on it in our own snapshots, and then, before committing the
> > changes we exclude the browser subpackage in the pom.xml.
>
> The only problematic class is the ErrorDialog. Is there a way to get that
> reworked to not use the extgui stuff? That would fix the issue just
> fine.
I've gone ahead and made ErrorDialog just extend PopUp for now. I'm really
not sure what that will do, but it at least now compiles and such.
Dan
> Dan
>
> > thanks, Sergey
> >
> > On Fri, Feb 4, 2011 at 5:55 PM, Sergey Beryozkin
>
> <sb...@gmail.com>wrote:
> > > Hi all,
> > >
> > > I have just moved the LogBrowser project Tomasz was working upon as
> > > part of GSOC 2010 to the trunk. I've done few minor changes before
> > > moving it in.
> > >
> > > - rt/management-web has a new package,
> > > org.apache.cxf.management.web.browser* which contains the code for the
> > > browser. It was org.apache.cxf.management.web.logging.logbrowser* in
> > > the sandbox but I thought we might enhance it in the future to work
> > > PersistingIn/Out interceptors in the rt/management so decided to
> > > simplify the package name.
> > >
> > > - samples/logbrowser is the demo showing the LogBrowser in action.
> > >
> > > It may make sense even move the whole rt/management-web into a CXF
> > > subproject, we'll see. For now, I'd like to have the browser polished
> > > and then we can decide.
> > >
> > > The GWT compilation will happen as part of building the
> > > rt/management-web, so the final jar will actually have the generated
> > > GWT code and resources included. We may consider doing it during the
> > > demo run, but again, lets try to polish the browser first.
> > >
> > > Tomasz, first of all congrats, I think it is good that your project is
> > > moving to the trunk.
> > > I'd appreciate if you could help me a bit with the initial clean-up,
> > > hope you'll find some time. You are a CXF committer now so you can do
> > > changes on the trunk...
> > >
> > > Here is the initial list of issues which I need some help with :
> > >
> > > 1. LogBrowser has a showstopper bug which I'm sure is due to the fact
> > > we just did not look at it for a while :-). Basically, when I go and
> > > generate the logs, and then click 'refresh' on the endpoint, the
> > > browser reports an exception.
> > > The AtomPullServer is functioning ok, Firefox subscribes to the Atom
> > > Feed with no probs.
> > >
> > > 2. Please move ManageEndpoints button either immediately above or below
> > > the Filter button
> > >
> > > 3. Remove the Tasks and Endpoints buttons/entries, lets have it the way
> > > you did it originally. We only need to see the list of endpoints which
> > > will be added via "ManageEndpoints", the explorer style is just too
> > > complex.
> > >
> > > 4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel
> > > and overlaps with the "Settings" entry,
> > >
> > > 5. In ManageEndpoints please change "back to browse" to "Back to
> > > Explorer"
> > >
> > >
> > > 6. We will need to think what to do with the jscript library added to
> > > the demo - we will have to link to its via remore URI or simple do a
> > > simpler plain HTML form - not a priority right now.
> > >
> > > Hope it's not a big list :-). We will catch up on the GWT dev a bit and
> > > I also hope other users with the GWT experience contribute in time.
> > > What I will do before 2.4 is look at the browser code and see what I
> > > can refactor but also add some helpers to the AtomPullServer so that
> > > users can use it for reading from the existing file logs...
> > >
> > > Thanks, Sergey
--
Daniel Kulp
dkulp@apache.org
http://dankulp.com/blog
Re: LogBrowser project is on the trunk
Posted by Daniel Kulp <dk...@apache.org>.
On Friday 04 February 2011 4:24:39 pm Sergey Beryozkin wrote:
> Tomasz, I had to comment out few lines in the rt/management-web/pom.xml
> (excluded the browser package from the compilation).
>
> I won't have time in the next couple of weeks to push the browser
> subpackage to a subproject, so for a while I suggest
> we work on it in our own snapshots, and then, before committing the changes
> we exclude the browser subpackage in the pom.xml.
The only problematic class is the ErrorDialog. Is there a way to get that
reworked to not use the extgui stuff? That would fix the issue just fine.
Dan
>
> thanks, Sergey
>
> On Fri, Feb 4, 2011 at 5:55 PM, Sergey Beryozkin
<sb...@gmail.com>wrote:
> > Hi all,
> >
> > I have just moved the LogBrowser project Tomasz was working upon as part
> > of GSOC 2010 to the trunk. I've done few minor changes before moving it
> > in.
> >
> > - rt/management-web has a new package,
> > org.apache.cxf.management.web.browser* which contains the code for the
> > browser. It was org.apache.cxf.management.web.logging.logbrowser* in the
> > sandbox but I thought we might enhance it in the future to work
> > PersistingIn/Out interceptors in the rt/management so decided to simplify
> > the package name.
> >
> > - samples/logbrowser is the demo showing the LogBrowser in action.
> >
> > It may make sense even move the whole rt/management-web into a CXF
> > subproject, we'll see. For now, I'd like to have the browser polished and
> > then we can decide.
> >
> > The GWT compilation will happen as part of building the
> > rt/management-web, so the final jar will actually have the generated GWT
> > code and resources included. We may consider doing it during the demo
> > run, but again, lets try to polish the browser first.
> >
> > Tomasz, first of all congrats, I think it is good that your project is
> > moving to the trunk.
> > I'd appreciate if you could help me a bit with the initial clean-up, hope
> > you'll find some time. You are a CXF committer now so you can do changes
> > on the trunk...
> >
> > Here is the initial list of issues which I need some help with :
> >
> > 1. LogBrowser has a showstopper bug which I'm sure is due to the fact we
> > just did not look at it for a while :-). Basically, when I go and
> > generate the logs, and then click 'refresh' on the endpoint, the browser
> > reports an exception.
> > The AtomPullServer is functioning ok, Firefox subscribes to the Atom Feed
> > with no probs.
> >
> > 2. Please move ManageEndpoints button either immediately above or below
> > the Filter button
> >
> > 3. Remove the Tasks and Endpoints buttons/entries, lets have it the way
> > you did it originally. We only need to see the list of endpoints which
> > will be added via "ManageEndpoints", the explorer style is just too
> > complex.
> >
> > 4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel and
> > overlaps with the "Settings" entry,
> >
> > 5. In ManageEndpoints please change "back to browse" to "Back to
> > Explorer"
> >
> >
> > 6. We will need to think what to do with the jscript library added to the
> > demo - we will have to link to its via remore URI or simple do a simpler
> > plain HTML form - not a priority right now.
> >
> > Hope it's not a big list :-). We will catch up on the GWT dev a bit and I
> > also hope other users with the GWT experience contribute in time. What I
> > will do before 2.4 is look at the browser code and see what I can
> > refactor but also add some helpers to the AtomPullServer so that users
> > can use it for reading from the existing file logs...
> >
> > Thanks, Sergey
--
Daniel Kulp
dkulp@apache.org
http://dankulp.com/blog
Re: LogBrowser project is on the trunk
Posted by Sergey Beryozkin <sb...@gmail.com>.
Tomasz, I had to comment out few lines in the rt/management-web/pom.xml
(excluded the browser package from the compilation).
I won't have time in the next couple of weeks to push the browser subpackage
to a subproject, so for a while I suggest
we work on it in our own snapshots, and then, before committing the changes
we exclude the browser subpackage in the pom.xml.
thanks, Sergey
On Fri, Feb 4, 2011 at 5:55 PM, Sergey Beryozkin <sb...@gmail.com>wrote:
> Hi all,
>
> I have just moved the LogBrowser project Tomasz was working upon as part of
> GSOC 2010 to the trunk. I've done few minor changes before moving it in.
>
> - rt/management-web has a new package,
> org.apache.cxf.management.web.browser* which contains the code for the
> browser. It was org.apache.cxf.management.web.logging.logbrowser* in the
> sandbox but I thought we might enhance it in the future to work
> PersistingIn/Out interceptors in the rt/management so decided to simplify
> the package name.
>
> - samples/logbrowser is the demo showing the LogBrowser in action.
>
> It may make sense even move the whole rt/management-web into a CXF
> subproject, we'll see. For now, I'd like to have the browser polished and
> then we can decide.
>
> The GWT compilation will happen as part of building the rt/management-web,
> so the final jar will actually have the generated GWT code and resources
> included. We may consider doing it during the demo run, but again, lets try
> to polish the browser first.
>
> Tomasz, first of all congrats, I think it is good that your project is
> moving to the trunk.
> I'd appreciate if you could help me a bit with the initial clean-up, hope
> you'll find some time. You are a CXF committer now so you can do changes on
> the trunk...
>
> Here is the initial list of issues which I need some help with :
>
> 1. LogBrowser has a showstopper bug which I'm sure is due to the fact we
> just did not look at it for a while :-). Basically, when I go and generate
> the logs, and then click 'refresh' on the endpoint, the browser reports an
> exception.
> The AtomPullServer is functioning ok, Firefox subscribes to the Atom Feed
> with no probs.
>
> 2. Please move ManageEndpoints button either immediately above or below the
> Filter button
>
> 3. Remove the Tasks and Endpoints buttons/entries, lets have it the way you
> did it originally. We only need to see the list of endpoints which will be
> added via "ManageEndpoints", the explorer style is just too complex.
>
> 4. When I go to ManageEndpoints, "Sign Out" leaks into the new panel and
> overlaps with the "Settings" entry,
>
> 5. In ManageEndpoints please change "back to browse" to "Back to Explorer"
>
>
> 6. We will need to think what to do with the jscript library added to the
> demo - we will have to link to its via remore URI or simple do a simpler
> plain HTML form - not a priority right now.
>
> Hope it's not a big list :-). We will catch up on the GWT dev a bit and I
> also hope other users with the GWT experience contribute in time. What I
> will do before 2.4 is look at the browser code and see what I can refactor
> but also add some helpers to the AtomPullServer so that users can use it for
> reading from the existing file logs...
>
> Thanks, Sergey
>
>
>