status from the dojo update

[Werner Punz <we...@gmx.at>]

Hi good news,
I did a test on the dojo stuff.
It looks good,
so far only the plain ajax inputsuggest and
partially the effects are broken.
The table suggest and the rest
still seems to work fine.
(all the dojo tests still run through)

I will commit the new codebase tonight
and will start to fix the affected components
tomorrow.

Re: status from the dojo update

[Werner Punz <we...@gmx.at>]

Werner Punz schrieb:
> Wendy Smoak schrieb:
>> On 8/18/06, Werner Punz <we...@gmx.at> wrote:
>>
>>> ok the update is done, i hope everything
>>> still works, as here....
>>> happy hacking, Gerald, check out
>>> your ajax auto suggest...
>>
>> Do we modify these files at all, or are they straight from the Dojo 
>> release?
>>
> ok now the update is fully done...
> the timed notifier also has been broken
> my code, but fixable ;-)
> 
> 
Btw. another note,
I already posted it, I am somewhat
sorry about this delayed update, it was
planned for june, but I am currently somewhat disabled
by an RSI which I am recovering from, so I had
to reduce my work on myfaces during the acute recovery period.
Still suffering on it, but it has been time to do it,
because there was a huge user demand and we cannot stay
on the old 0.2.2 for too long, because it was way to buggy.

Re: status from the dojo update

[Werner Punz <we...@gmx.at>]

Wendy Smoak schrieb:
> On 8/18/06, Werner Punz <we...@gmx.at> wrote:
> 
>> ok the update is done, i hope everything
>> still works, as here....
>> happy hacking, Gerald, check out
>> your ajax auto suggest...
> 
> Do we modify these files at all, or are they straight from the Dojo 
> release?
> 
ok now the update is fully done...
the timed notifier also has been broken
my code, but fixable ;-)

Re: status from the dojo update

[Mike Kienenberger <mk...@gmail.com>]

On 8/21/06, Werner Punz <we...@gmx.at> wrote:
> Also after that we should think about deprecation of some of the
> Tomahawk controls, we often have two controls doing the same
> and now a third dojo based implementation which often is better
> then ours.

We need to be careful about this.   We have end users who depend on
particular components, and having one that does "almost the same
thing" may not be sufficient.    When we combine components (like with
newspaperTable and dataTable), we need to make sure that the final
component has all of the functionality (and possibly look-and-feel) of
the original.

Re: status from the dojo update

[Werner Punz <we...@gmx.at>]

Wendy Smoak schrieb:
> On 8/18/06, Werner Punz <we...@gmx.at> wrote:
> 
>> Sorry I was not fully aware of that, I have been absent due
>> to health reasons for a while, until Gerald started
>> bugging me about the pending update.
> 
> I saw that-- you have my sympathies.  :(  If I had not been introduced
> to Kinesis keyboards and Cirque touchpads last summer, I probably
> wouldn't be typing now.
> 
> BTW, if anyone has been away for a while, we have a status document now:
>   http://svn.apache.org/repos/asf/myfaces/current/STATUS.txt
> 
>> Tomahawk core itself is not affected by those issues, since dojo is
>> sandbox only and nothing of the core references ít for now.
> 
> I just noticed that.  Okay, since the sandbox code isn't part of the
> release, we can still go with the latest trunk when we're ready.  It's
> probably at least a week away at this point.
> 
Ok, after the release, I will start to push dojo down into the Tomahawk
core.

My plan is...
new dojo release, wait 1-2 months for things to settle down
then upgrade in tomahawk after a major release.
If that is ok with everybody.
The thing is, the self written Dojo interfaces seem to be stable and
bugfree no significant changes since the last two months, and the
upgrade went mostly painless with only fixes on the javascript side.
We need the lib in the core now, because we then can start to remove a
lot of old not well working DHTML legacy code with Dojo controls or classes.

Also I will open a list in the wiki which will document which control
uses dojo and who the current core maintainer of the control is.
So that we will get some structure in.

Also after that we should think about deprecation of some of the
Tomahawk controls, we often have two controls doing the same
and now a third dojo based implementation which often is better
then ours.

Re: status from the dojo update

[Wendy Smoak <ws...@gmail.com>]

On 8/18/06, Werner Punz <we...@gmx.at> wrote:

> Sorry I was not fully aware of that, I have been absent due
> to health reasons for a while, until Gerald started
> bugging me about the pending update.

I saw that-- you have my sympathies.  :(  If I had not been introduced
to Kinesis keyboards and Cirque touchpads last summer, I probably
wouldn't be typing now.

BTW, if anyone has been away for a while, we have a status document now:
   http://svn.apache.org/repos/asf/myfaces/current/STATUS.txt

> Tomahawk core itself is not affected by those issues, since dojo is
> sandbox only and nothing of the core references ít for now.

I just noticed that.  Okay, since the sandbox code isn't part of the
release, we can still go with the latest trunk when we're ready.  It's
probably at least a week away at this point.

-- 
Wendy

Re: status from the dojo update

[Werner Punz <we...@gmx.at>]

Wendy Smoak schrieb:
> On 8/18/06, Werner Punz <we...@gmx.at> wrote:
> 
>> We already have a good plan on how to
>> get rid of the dojo codebase within myfaces
>> weblets probably will be the way to go ...
> 
> Sorry, it was probably discussed before I started paying attention. :)
> 
> Another question:  We branched for Tomahawk 1.1.4 a while ago, but
> Core has taken longer than expected.
> 
Sorry I was not fully aware of that, I have been absent due
to health reasons for a while, until Gerald started
bugging me about the pending update.

> I was planning to delete the Tomahawk 1.1.4 branch and re-copy it from
> the latest trunk when we're ready to start on the release.
> 
> Does that still make sense, or should we copy from a revision prior to
> the Dojo upgrade, and go with that for 1.1.4?

Tomahawk core itself is not affected by those issues, since dojo is 
sandbox only and nothing of the core references ít for now.
(which will be the next step to get dojo into tom...)

As for Tom...
I cannot say that for now, lets play it save
use a revision prior to the dojo update is the safe bet.
But...

I just noticed three sandbox components which were not working fully
which was way better than I even expected in my most positive dreams.

I expect them to work again within a few days timeframe.
But who knows for now.

If everything works as expected by mid next week, you probably
can add dojo 0.3.1 to the 1.14 sandbix if not keep it safe.

Re: status from the dojo update

[Wendy Smoak <ws...@gmail.com>]

On 8/18/06, Werner Punz <we...@gmx.at> wrote:

> We already have a good plan on how to
> get rid of the dojo codebase within myfaces
> weblets probably will be the way to go ...

Sorry, it was probably discussed before I started paying attention. :)

Another question:  We branched for Tomahawk 1.1.4 a while ago, but
Core has taken longer than expected.

I was planning to delete the Tomahawk 1.1.4 branch and re-copy it from
the latest trunk when we're ready to start on the release.

Does that still make sense, or should we copy from a revision prior to
the Dojo upgrade, and go with that for 1.1.4?

Thanks,
Wendy

Re: status from the dojo update

[Werner Punz <we...@gmx.at>]

Wendy Smoak schrieb:
> On 8/18/06, Werner Punz <we...@gmx.at> wrote:
> 
>> in my case I was using some old effects stuff
>> which has been deprecated and is moved
>> into an animation framework, not a big issue
>> in the case of the autosuggest, a lot was
>> added due to requests of a project
>> which paid a lot of the dev for some
>> of the components, so yes, some stuff has
>> been added.
> 
> Added to what was in Tomahawk before, or added to Dojo itself ?
> 
> What I'm wondering is if we can get Dojo into a Maven repo, and simply
> depend on a jar or zip file, which we expand at build time.  That
> would save us having to add it all to the MyFaces svn repo, but it
> will not work if we're modifying the Dojo source code.
> 
The code affected is our code,
as for your request, the long term goal
is to move dojo out of our codebase
but lets do one step after the other...


We already have a good plan on how to
get rid of the dojo codebase within myfaces
weblets probably will be the way to go
to achieve that, but that will require
some infrastructural overhaul.
Lets live for now with what we have
and do one step after the other...
The most important thing was
to get to the current stable version
next step is to get dojo into tomahawk
so that we can move some components off
the sandbox, which has become too big.
Then we can think on replacing the current
loader infrastructure of dojo with weblets...


> Also, is there a JIRA issue for this upgrade?  If not, please open one
> so it will show up in the release notes.
> 
ok....
will do that

Re: status from the dojo update

[Wendy Smoak <ws...@gmail.com>]

On 8/18/06, Werner Punz <we...@gmx.at> wrote:

> in my case I was using some old effects stuff
> which has been deprecated and is moved
> into an animation framework, not a big issue
> in the case of the autosuggest, a lot was
> added due to requests of a project
> which paid a lot of the dev for some
> of the components, so yes, some stuff has
> been added.

Added to what was in Tomahawk before, or added to Dojo itself ?

What I'm wondering is if we can get Dojo into a Maven repo, and simply
depend on a jar or zip file, which we expand at build time.  That
would save us having to add it all to the MyFaces svn repo, but it
will not work if we're modifying the Dojo source code.

Also, is there a JIRA issue for this upgrade?  If not, please open one
so it will show up in the release notes.

Thanks,
-- 
Wendy

Re: status from the dojo update

[Werner Punz <we...@gmx.at>]

Wendy Smoak schrieb:
> On 8/18/06, Werner Punz <we...@gmx.at> wrote:
> 
>> ok the update is done, i hope everything
>> still works, as here....
>> happy hacking, Gerald, check out
>> your ajax auto suggest...
> 
> Do we modify these files at all, or are they straight from the Dojo 
> release?
> 
sorry I was too fast, I am still updating
(subclipse has missed some important files)

in my case I was using some old effects stuff
which has been deprecated and is moved
into an animation framework, not a big issue
in the case of the autosuggest, a lot was
added due to requests of a project
which paid a lot of the dev for some
of the components, so yes, some stuff has
been added.
Gerald who did the autosuggest is aware of that
and will fix his stuff in the next days.
The same goes for me.

Re: status from the dojo update

[Wendy Smoak <ws...@gmail.com>]

On 8/18/06, Werner Punz <we...@gmx.at> wrote:

> ok the update is done, i hope everything
> still works, as here....
> happy hacking, Gerald, check out
> your ajax auto suggest...

Do we modify these files at all, or are they straight from the Dojo release?

-- 
Wendy

Re: status from the dojo update

[Werner Punz <we...@gmx.at>]

Werner Punz schrieb:
> Hi good news,
> I did a test on the dojo stuff.
> It looks good,
> so far only the plain ajax inputsuggest and
> partially the effects are broken.
> The table suggest and the rest
> still seems to work fine.
> (all the dojo tests still run through)
> 
> I will commit the new codebase tonight
> and will start to fix the affected components
> tomorrow.
> 
> 
ok the update is done, i hope everything
still works, as here....
happy hacking, Gerald, check out
your ajax auto suggest...


Cheers

Werner