You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@qpid.apache.org by Marnie McCormack <ma...@googlemail.com> on 2008/12/02 13:58:33 UTC
M4 Timelines
Hi Rafi & All,
What are our current M4 timelines currently looking like now ?
I need to plan some testing on it and it'd be good to know what the dates
are looking like.
Thanks & Regards,
Marnie
Re: M4 Timelines
Posted by Carl Trieloff <cc...@redhat.com>.
Robert Godfrey wrote:
> 2008/12/2 Carl Trieloff <cc...@redhat.com>:
>
>>> We need to retain the 0-8 version as it is in use and may be upgraded.
>>> There
>>> were several fixes and enhancements to the 0-8 version in M3.
>>>
>> Can we move the 0-8 client into a subdir, any objections to that?
>>
>
> I think we should separate the 0-8 and 0-10 cleanly (i.e. neither one
> should be in a subdirectory of the other...) So if you mean that both
> 0-8 and 0-10 are in subdirs, then yes - I agree
yes, this is what I was suggesting.
dotnet
dotnet/client080
dotnet/client010
So it would be a very low / no risk change as what is in dotnet dir
would just be moved to client080 dir
Carl.
Re: M4 Timelines
Posted by Robert Godfrey <ro...@gmail.com>.
2008/12/2 Carl Trieloff <cc...@redhat.com>:
>
>> We need to retain the 0-8 version as it is in use and may be upgraded.
>> There
>> were several fixes and enhancements to the 0-8 version in M3.
>
> Can we move the 0-8 client into a subdir, any objections to that?
I think we should separate the 0-8 and 0-10 cleanly (i.e. neither one
should be in a subdirectory of the other...) So if you mean that both
0-8 and 0-10 are in subdirs, then yes - I agree
>
>>> 3. The Java build needs something in the README about Java 1.5 /1.6 or
>>> we
>>> need to fix it. It also needs to be updated
>>> with the ant instructions. I also think we should kill maven so people
>>> don't try and fail to build with Maven. (Hiram's
>>> issues trying to build)
>>>
>>>
>>
>> I think the issue with the 4? classes should be addressed - I think Martin
>> is going to raise a JIRA. Maven can be useful for generating the
>> IDEA/Eclipse project files so maybe we could just document the build
>> instructions for trunk ? I'm happy to do that.
>
> What if we just leave maven in for eclipse and pull it out for the rest of
> the Java. Some general
> cleanup is needed. Thoughts?
We need to revisit all the things that we were still relying on the
maven build for. IIRC it was not only the eclipse stuff, but also the
generation of the performance tests (not sure if someone fixed this in
the meantime)...
-- Rob
Re: M4 Timelines
Posted by Martin Ritchie <ri...@apache.org>.
2008/12/2 Carl Trieloff <cc...@redhat.com>:
>
>> We need to retain the 0-8 version as it is in use and may be upgraded.
>> There
>> were several fixes and enhancements to the 0-8 version in M3.
>
> Can we move the 0-8 client into a subdir, any objections to that?
>
>
>>> 3. The Java build needs something in the README about Java 1.5 /1.6 or
>>> we
>>> need to fix it. It also needs to be updated
>>> with the ant instructions. I also think we should kill maven so people
>>> don't try and fail to build with Maven. (Hiram's
>>> issues trying to build)
>>>
>>>
>>
>> I think the issue with the 4? classes should be addressed - I think Martin
>> is going to raise a JIRA. Maven can be useful for generating the
>> IDEA/Eclipse project files so maybe we could just document the build
>> instructions for trunk ? I'm happy to do that.
>
> What if we just leave maven in for eclipse and pull it out for the rest of
> the Java. Some general
> cleanup is needed. Thoughts?
General clean up of the maven would be good with the default maven
target giving details of how to build eclipse and intelij projects. It
would be good though to get the pom file generation in to the ant
build system, I recall the feature be discussed a while back but not
sure where that went.
Cheers
Martin
> Carl.
--
Martin Ritchie
Re: M4 Timelines
Posted by Carl Trieloff <cc...@redhat.com>.
> We need to retain the 0-8 version as it is in use and may be upgraded. There
> were several fixes and enhancements to the 0-8 version in M3.
Can we move the 0-8 client into a subdir, any objections to that?
>> 3. The Java build needs something in the README about Java 1.5 /1.6 or we
>> need to fix it. It also needs to be updated
>> with the ant instructions. I also think we should kill maven so people
>> don't try and fail to build with Maven. (Hiram's
>> issues trying to build)
>>
>>
>
> I think the issue with the 4? classes should be addressed - I think Martin
> is going to raise a JIRA. Maven can be useful for generating the
> IDEA/Eclipse project files so maybe we could just document the build
> instructions for trunk ? I'm happy to do that.
What if we just leave maven in for eclipse and pull it out for the rest
of the Java. Some general
cleanup is needed. Thoughts?
Carl.
Re: M4 Timelines
Posted by Marnie McCormack <ma...@googlemail.com>.
Hi Carl,
My pref is strongly for c) at this point. After all, the 0-10 client doesn't
interop with the Java Broker - so it can't be the only option we point to.
Regards,
Marnie
On Tue, Dec 2, 2008 at 3:56 PM, Carl Trieloff <cc...@redhat.com> wrote:
>
>
>> I think it's too late to be moving around directories at this stage. IMHO
>> we should focus on testing, getting the READMEs up to date, and getting JIRA
>> cleaned up. It's always been the case that each sub-project in qpid has its
>> own private way of doing things, and our releases are a bit inconsistent as
>> a result. I don't think we're going to solve this issue for M4, but if we
>> make it a focus for the next devel cycle, there's a lot we could do to
>> improve things for M5. I think if we were to do this it would be reasonable
>> to drop the Milestone moniker from the release at that point.
>>
>>
> I did some README clean up, which brought the question up for me.
>
> My concern is that it is confusing. The issue is the RELEASE_NOTE.txt file
> in the top level talks about M2.
>
> options:
>
> a.) I could create a subdir for 0-8 and have top level REAME point to
> readme's in each subdir. But NOT move the
> 0-8 client into the subdir till after the release.
>
> b.) we can move 0-8 into a sub dir
>
> c.) we can do a bit more of what I have started -- add comments into the
> top level files to point to client010 files and
> strip out the out of date info around M2.
>
> any preferences?
>
> it can't go out with files that read like it is a M2 release.
> Carl.
>
Re: M4 Timelines
Posted by Aidan Skinner <ai...@apache.org>.
On Tue, Dec 2, 2008 at 3:56 PM, Carl Trieloff <cc...@redhat.com> wrote:
> I did some README clean up, which brought the question up for me.
> c.) we can do a bit more of what I have started -- add comments into the top
> level files to point to client010 files and
> strip out the out of date info around M2.
>
> any preferences?
I think this is the easiest route to go, I've attached updated README
and RELEASE_NOTES. I don't know much about the 0-10 client other than
it supports 0-10 so it'd be better if somebody who did could add some
more information. :)
- Aidan
--
Apache Qpid - World Domination through Advanced Message Queueing
http://cwiki.apache.org/qpid
"Have we anything resembling a plan?" "Mm-hm. Ride till we find
them... and kill them all." - The 13th Warrior
Re: M4 Timelines
Posted by Carl Trieloff <cc...@redhat.com>.
>
> I think (c) sounds like the appropriate measure at this point.
>
Seems the general preference, I'll hack at the implicated files a bit.
Carl.
Re: M4 Timelines
Posted by Rafael Schloming <ra...@redhat.com>.
Carl Trieloff wrote:
>
>>
>> I think it's too late to be moving around directories at this stage.
>> IMHO we should focus on testing, getting the READMEs up to date, and
>> getting JIRA cleaned up. It's always been the case that each
>> sub-project in qpid has its own private way of doing things, and our
>> releases are a bit inconsistent as a result. I don't think we're going
>> to solve this issue for M4, but if we make it a focus for the next
>> devel cycle, there's a lot we could do to improve things for M5. I
>> think if we were to do this it would be reasonable to drop the
>> Milestone moniker from the release at that point.
>>
>
> I did some README clean up, which brought the question up for me.
>
> My concern is that it is confusing. The issue is the RELEASE_NOTE.txt
> file in the top level talks about M2.
>
> options:
>
> a.) I could create a subdir for 0-8 and have top level REAME point to
> readme's in each subdir. But NOT move the
> 0-8 client into the subdir till after the release.
>
> b.) we can move 0-8 into a sub dir
>
> c.) we can do a bit more of what I have started -- add comments into the
> top level files to point to client010 files and
> strip out the out of date info around M2.
>
> any preferences?
>
> it can't go out with files that read like it is a M2 release.
I think (c) sounds like the appropriate measure at this point.
--Rafael
Re: M4 Timelines
Posted by Carl Trieloff <cc...@redhat.com>.
>
> I think it's too late to be moving around directories at this stage.
> IMHO we should focus on testing, getting the READMEs up to date, and
> getting JIRA cleaned up. It's always been the case that each
> sub-project in qpid has its own private way of doing things, and our
> releases are a bit inconsistent as a result. I don't think we're going
> to solve this issue for M4, but if we make it a focus for the next
> devel cycle, there's a lot we could do to improve things for M5. I
> think if we were to do this it would be reasonable to drop the
> Milestone moniker from the release at that point.
>
I did some README clean up, which brought the question up for me.
My concern is that it is confusing. The issue is the RELEASE_NOTE.txt
file in the top level talks about M2.
options:
a.) I could create a subdir for 0-8 and have top level REAME point to
readme's in each subdir. But NOT move the
0-8 client into the subdir till after the release.
b.) we can move 0-8 into a sub dir
c.) we can do a bit more of what I have started -- add comments into the
top level files to point to client010 files and
strip out the out of date info around M2.
any preferences?
it can't go out with files that read like it is a M2 release.
Carl.
Re: M4 Timelines
Posted by Marnie McCormack <ma...@googlemail.com>.
<snip>
>>
>
> I don't know if you're referring to some comments I made on IM, but my
> concerns are not with the 0-10 .net code per se. In fact 0-10 .net code is a
> direct translation of the 0-10 Java code, both of which have received more
> review in general than the 0-8 code which predates our review process.
>
> My concern is with the general approach to the .net client. Both the 0-8
> .net client and the 0-10 .net client are simply translated snapshots of the
> Java codebase at a particular point in time. I don't believe this approach
> is maintainable. The 0-8 .net client is years behind the Java 0-8 client,
> and it shows in terms of bugs and features that were fixed/added long ago
> for Java, but never made it to .net. Similarly the 0-10 .net client is
> already behind the Java 0-10 code (both in bugs and features) despite being
> only a month or so old.
>
> As you say though this is separate from the release discussion.
Yes, my comments weren't really directed at the IM discussions. Just that we
need to discuss the .Net outside the scope of the M4 release.
<snip>
> Maven can be useful for generating the
>> IDEA/Eclipse project files so maybe we could just document the build
>> instructions for trunk ? I'm happy to do that.
>>
>
> I don't think the ability to generate IDEA/Eclipse project files is worth
> the confusion of keeping around the pom files. Isn't there some way it could
> be done without having broken pom files sitting around?
I'm not sure it's so confusing really - they are really handy if you have to
blat and redo your project. maybe we could go back to having a template
.ipr/.iml files etc. We've had maven & ant a long time :-)
>
>
> I am still checking licensing, but that is the punch list I have found so
>>> far on the last spin.
>>>
>>> Volunteers welcome to help clean up.
>>> Carl.
>>>
>>>
>>> I also sent some feedback on the alpha & beta. Be good to get Rafi's view
>> of
>> what he thinks should be done and we can pitch in.
>>
>
> I think it's too late to be moving around directories at this stage. IMHO
> we should focus on testing, getting the READMEs up to date, and getting JIRA
> cleaned up. It's always been the case that each sub-project in qpid has its
> own private way of doing things, and our releases are a bit inconsistent as
> a result. I don't think we're going to solve this issue for M4, but if we
> make it a focus for the next devel cycle, there's a lot we could do to
> improve things for M5. I think if we were to do this it would be reasonable
> to drop the Milestone moniker from the release at that point.
I'd be happy for this to go as is now, and agree about the moving around.
Let's discuss/change the component structures for M5 as suggested.
>
> --Rafael
>
>
Re: M4 Timelines
Posted by Hiram Chirino <hi...@hiramchirino.com>.
Even the eclipse project files that the poms generate are not very
accurate. Perhaps you should kill the poms and just check in the
eclipse project files?
On Tue, Dec 2, 2008 at 1:49 PM, Carl Trieloff <cc...@redhat.com> wrote:
>
>>
>>> IDEA/Eclipse project files so maybe we could just document the build
>>> instructions for trunk ? I'm happy to do that.
>>
>> I don't think the ability to generate IDEA/Eclipse project files is worth
>> the confusion of keeping around the pom files. Isn't there some way it could
>> be done without having broken pom files sitting around?
>
> +1.
>
> Any concern taking this approach? i.e. leave pom files for eclipse - kill
> the rest.
>
> Carl.
>
--
Regards,
Hiram
Blog: http://hiramchirino.com
Open Source SOA
http://open.iona.com
Re: M4 Timelines
Posted by Carl Trieloff <cc...@redhat.com>.
>
>> IDEA/Eclipse project files so maybe we could just document the build
>> instructions for trunk ? I'm happy to do that.
>
> I don't think the ability to generate IDEA/Eclipse project files is
> worth the confusion of keeping around the pom files. Isn't there some
> way it could be done without having broken pom files sitting around?
+1.
Any concern taking this approach? i.e. leave pom files for eclipse -
kill the rest.
Carl.
Re: M4 Timelines
Posted by Rafael Schloming <ra...@redhat.com>.
Marnie McCormack wrote:
> Hi Carl,
>
> Comments inline ....
>
> On Tue, Dec 2, 2008 at 1:47 PM, Carl Trieloff <cc...@redhat.com> wrote:
>
>> I looked at the last spin artifacts. I think we have some cleanup /
>> decision to make before the next spin
>>
>> 1. Ruby & Python README's need some, how to use comments (I have pinging
>> Rafi)
>> 2. .NET ( we have a 0-8 client that has no changed since M2 and the new 010
>> client. However the 010 client
>> is in a subdir, 0-8 is not. Should we move the 0-8 client into a subdir or
>> just delete it, as it can be obtained from M3
>> or M2? It is problematic that the RELEASE_NOTES talk about M2 at the top
>> level and confusing
>
>
> We need to retain the 0-8 version as it is in use and may be upgraded. There
> were several fixes and enhancements to the 0-8 version in M3. I think a
> separate discussion on the 0-10 .Net client would be useful as I know some
> concerns existed around the review.
I don't know if you're referring to some comments I made on IM, but my
concerns are not with the 0-10 .net code per se. In fact 0-10 .net code
is a direct translation of the 0-10 Java code, both of which have
received more review in general than the 0-8 code which predates our
review process.
My concern is with the general approach to the .net client. Both the 0-8
.net client and the 0-10 .net client are simply translated snapshots of
the Java codebase at a particular point in time. I don't believe this
approach is maintainable. The 0-8 .net client is years behind the Java
0-8 client, and it shows in terms of bugs and features that were
fixed/added long ago for Java, but never made it to .net. Similarly the
0-10 .net client is already behind the Java 0-10 code (both in bugs and
features) despite being only a month or so old.
As you say though this is separate from the release discussion.
>> 3. The Java build needs something in the README about Java 1.5 /1.6 or we
>> need to fix it. It also needs to be updated
>> with the ant instructions. I also think we should kill maven so people
>> don't try and fail to build with Maven. (Hiram's
>> issues trying to build)
>>
>
> I think the issue with the 4? classes should be addressed - I think Martin
> is going to raise a JIRA.
There is still the double build issue even if the overrides are removed.
I believe this issue is really what permitted the @Overrides to sneak
in, since it's fairly inconvenient now to use 1.5 as your primary devel
platform, so we don't really check that we compile on 1.5 very often.
Maven can be useful for generating the
> IDEA/Eclipse project files so maybe we could just document the build
> instructions for trunk ? I'm happy to do that.
I don't think the ability to generate IDEA/Eclipse project files is
worth the confusion of keeping around the pom files. Isn't there some
way it could be done without having broken pom files sitting around?
>> I am still checking licensing, but that is the punch list I have found so
>> far on the last spin.
>>
>> Volunteers welcome to help clean up.
>> Carl.
>>
>>
> I also sent some feedback on the alpha & beta. Be good to get Rafi's view of
> what he thinks should be done and we can pitch in.
I think it's too late to be moving around directories at this stage.
IMHO we should focus on testing, getting the READMEs up to date, and
getting JIRA cleaned up. It's always been the case that each sub-project
in qpid has its own private way of doing things, and our releases are a
bit inconsistent as a result. I don't think we're going to solve this
issue for M4, but if we make it a focus for the next devel cycle,
there's a lot we could do to improve things for M5. I think if we were
to do this it would be reasonable to drop the Milestone moniker from the
release at that point.
--Rafael
Re: M4 Timelines
Posted by Marnie McCormack <ma...@googlemail.com>.
Hi Carl,
Comments inline ....
On Tue, Dec 2, 2008 at 1:47 PM, Carl Trieloff <cc...@redhat.com> wrote:
>
> I looked at the last spin artifacts. I think we have some cleanup /
> decision to make before the next spin
>
> 1. Ruby & Python README's need some, how to use comments (I have pinging
> Rafi)
> 2. .NET ( we have a 0-8 client that has no changed since M2 and the new 010
> client. However the 010 client
> is in a subdir, 0-8 is not. Should we move the 0-8 client into a subdir or
> just delete it, as it can be obtained from M3
> or M2? It is problematic that the RELEASE_NOTES talk about M2 at the top
> level and confusing
We need to retain the 0-8 version as it is in use and may be upgraded. There
were several fixes and enhancements to the 0-8 version in M3. I think a
separate discussion on the 0-10 .Net client would be useful as I know some
concerns existed around the review.
>
> 3. The Java build needs something in the README about Java 1.5 /1.6 or we
> need to fix it. It also needs to be updated
> with the ant instructions. I also think we should kill maven so people
> don't try and fail to build with Maven. (Hiram's
> issues trying to build)
>
I think the issue with the 4? classes should be addressed - I think Martin
is going to raise a JIRA. Maven can be useful for generating the
IDEA/Eclipse project files so maybe we could just document the build
instructions for trunk ? I'm happy to do that.
>
> I am still checking licensing, but that is the punch list I have found so
> far on the last spin.
>
> Volunteers welcome to help clean up.
> Carl.
>
>
I also sent some feedback on the alpha & beta. Be good to get Rafi's view of
what he thinks should be done and we can pitch in.
Bye for now,
Marnie
>
> Marnie McCormack wrote:
>
>> Hi Rafi & All,
>>
>> What are our current M4 timelines currently looking like now ?
>>
>> I need to plan some testing on it and it'd be good to know what the dates
>> are looking like.
>>
>> Thanks & Regards,
>> Marnie
>>
>>
>>
>
>
Re: M4 Timelines
Posted by Carl Trieloff <cc...@redhat.com>.
I looked at the last spin artifacts. I think we have some cleanup /
decision to make before the next spin
1. Ruby & Python README's need some, how to use comments (I have pinging
Rafi)
2. .NET ( we have a 0-8 client that has no changed since M2 and the new
010 client. However the 010 client
is in a subdir, 0-8 is not. Should we move the 0-8 client into a subdir
or just delete it, as it can be obtained from M3
or M2? It is problematic that the RELEASE_NOTES talk about M2 at the
top level and confusing
3. The Java build needs something in the README about Java 1.5 /1.6 or
we need to fix it. It also needs to be updated
with the ant instructions. I also think we should kill maven so people
don't try and fail to build with Maven. (Hiram's
issues trying to build)
I am still checking licensing, but that is the punch list I have found
so far on the last spin.
Volunteers welcome to help clean up.
Carl.
Marnie McCormack wrote:
> Hi Rafi & All,
>
> What are our current M4 timelines currently looking like now ?
>
> I need to plan some testing on it and it'd be good to know what the dates
> are looking like.
>
> Thanks & Regards,
> Marnie
>
>
Re: M4 Timelines
Posted by Gordon Sim <gs...@redhat.com>.
Rafael Schloming wrote:
> Marnie McCormack wrote:
>> Hi Rafi,
>>
>> Where are we at now do you think ?
>>
>> It'd be great to have a plan of attack for getting M4 out there, to which
>> people could chip in ?
>
> I'm waiting for the last M4 blocker to be cleared (QPID-1494: perf core
> on exit). As soon as this is done I'm going to roll an M4 RC.
I've moved that to the next release, I don't believe it should block M4.
Re: M4 Timelines
Posted by Rafael Schloming <ra...@redhat.com>.
Marnie McCormack wrote:
> Hi Rafi,
>
> Where are we at now do you think ?
>
> It'd be great to have a plan of attack for getting M4 out there, to which
> people could chip in ?
I'm waiting for the last M4 blocker to be cleared (QPID-1494: perf core
on exit). As soon as this is done I'm going to roll an M4 RC.
--Rafael
Re: M4 Timelines
Posted by Marnie McCormack <ma...@googlemail.com>.
Hi Rafi,
Where are we at now do you think ?
It'd be great to have a plan of attack for getting M4 out there, to which
people could chip in ?
Bye for now,
Marnie
On Tue, Dec 2, 2008 at 12:58 PM, Marnie McCormack <
marnie.mccormack@googlemail.com> wrote:
> Hi Rafi & All,
>
> What are our current M4 timelines currently looking like now ?
>
> I need to plan some testing on it and it'd be good to know what the dates
> are looking like.
>
> Thanks & Regards,
> Marnie
>