You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@reef.apache.org by Boris Shulman <sh...@gmail.com> on 2016/05/02 10:21:19 UTC
REEF 0.15
As REEF-1054 will take more time then anticipated I suggest to cur REEF
0.15 rigth after the following JIRAs will be in:
1) REEF-1354 <https://github.com/apache/reef/pull/966> - that allows using
multiruntime from c# by implementing the c# client side (without official
support in REEF .NET)
2) REEF-1375 <https://github.com/apache/reef/pull/983> - that fixes Java
multi runtime in 0.15
3) REEF-1243
Do we have any other JIRAs that are a must for 0.15 besides these?
@Dhruv, did we have any progress on REEF-1243?
Boris.
Re: REEF 0.15
Posted by Markus Weimer <ma...@weimo.de>.
Andrew, you had a hunch for this, right?
Markus
On Mon, May 9, 2016 at 8:53 PM, Boris Shulman <sh...@gmail.com> wrote:
> We are still blocked on REEF-1243. Any chance fixing it till end of this
> week, so we could cut 0.15 next week?
>
> On Thu, May 5, 2016 at 11:35 AM, Markus Weimer <ma...@weimo.de> wrote:
>
>> On Thu, May 5, 2016 at 11:12 AM, Boris Shulman <sh...@gmail.com> wrote:
>> > Should I remove from this JIRA all the links that we decided that are non
>> > blocking?
>>
>> Yes, and leave a comment that we decided so over here. Just in case we
>> question the wisdom of our present selves in the future :)
>>
>> Markus
>>
Re: REEF 0.15
Posted by Boris Shulman <sh...@gmail.com>.
We are still blocked on REEF-1243. Any chance fixing it till end of this
week, so we could cut 0.15 next week?
On Thu, May 5, 2016 at 11:35 AM, Markus Weimer <ma...@weimo.de> wrote:
> On Thu, May 5, 2016 at 11:12 AM, Boris Shulman <sh...@gmail.com> wrote:
> > Should I remove from this JIRA all the links that we decided that are non
> > blocking?
>
> Yes, and leave a comment that we decided so over here. Just in case we
> question the wisdom of our present selves in the future :)
>
> Markus
>
Re: REEF 0.15
Posted by Markus Weimer <ma...@weimo.de>.
On Thu, May 5, 2016 at 11:12 AM, Boris Shulman <sh...@gmail.com> wrote:
> Should I remove from this JIRA all the links that we decided that are non
> blocking?
Yes, and leave a comment that we decided so over here. Just in case we
question the wisdom of our present selves in the future :)
Markus
Re: REEF 0.15
Posted by Boris Shulman <sh...@gmail.com>.
Should I remove from this JIRA all the links that we decided that are non
blocking?
On Thu, May 5, 2016 at 11:10 AM, Markus Weimer <ma...@weimo.de> wrote:
> On Thu, May 5, 2016 at 10:46 AM, Mariia Mykhailova
> <ma...@microsoft.com> wrote:
> > Done. REEF-281 is our new issue for this work
>
> Thank you! I also took the liberty of assigning REEF-1215 to Boris :)
>
> Markus
>
Re: REEF 0.15
Posted by Markus Weimer <ma...@weimo.de>.
On Thu, May 5, 2016 at 10:46 AM, Mariia Mykhailova
<ma...@microsoft.com> wrote:
> Done. REEF-281 is our new issue for this work
Thank you! I also took the liberty of assigning REEF-1215 to Boris :)
Markus
RE: REEF 0.15
Posted by Mariia Mykhailova <ma...@microsoft.com>.
Done. REEF-281 is our new issue for this work
-Mariia
-----Original Message-----
From: Markus Weimer [mailto:markus@weimo.de]
Sent: Wednesday, May 4, 2016 4:46 PM
To: dev@reef.apache.org
Subject: Re: REEF 0.15
On 2016-05-04 15:31, Mariia Mykhailova wrote:
> [REEF-330] doesn't look like it will allow us to release any time
> soon.
+1 Let's remove it from the list of blockers.
> For [REEF-195], .nuspec files use $version$ for version, which uses
> assembly version from AssemblyInfo.cs. So dll and NuGet versions (as
> JIRA title and description require) should already be synchronized?
Fair enough. Maybe the issue is still open by accident? in which case we can close it.
> Comments on the issue discuss using version from .pom file, using one
> AssemblyInfo across all projects etc. But this is a different topic,
> with different amount of work associated.
Yes, maybe the right course of action is to close REEF-195 and file a new issue for this.
Markus
Re: REEF 0.15
Posted by Markus Weimer <ma...@weimo.de>.
On 2016-05-04 15:31, Mariia Mykhailova wrote:
> [REEF-330] doesn't look like it will allow us to release any time
> soon.
+1 Let's remove it from the list of blockers.
> For [REEF-195], .nuspec files use $version$ for version, which uses
> assembly version from AssemblyInfo.cs. So dll and NuGet versions (as
> JIRA title and description require) should already be synchronized?
Fair enough. Maybe the issue is still open by accident? in which case we
can close it.
> Comments on the issue discuss using version from .pom file, using
> one AssemblyInfo across all projects etc. But this is a different
> topic, with different amount of work associated.
Yes, maybe the right course of action is to close REEF-195 and file a
new issue for this.
Markus
RE: REEF 0.15
Posted by Mariia Mykhailova <ma...@microsoft.com>.
[REEF-330] doesn't look like it will allow us to release any time soon. We need to have a story for testing on YARN, but probably not yet
For [REEF-195], .nuspec files use $version$ for version, which uses assembly version from AssemblyInfo.cs. So dll and NuGet versions (as JIRA title and description require) should already be synchronized? During the release we only change version in AssemblyInfo.cs and .cpp files, not in any .nuspec files.
Comments on the issue discuss using version from .pom file, using one AssemblyInfo across all projects etc. But this is a different topic, with different amount of work associated.
Besides, change of version is already handled with a single script and does not slow down release process or require any extra effort from release manager. I agree that we should try and reduce the quantity of "version" values in our files, but this is a low priority work and should not block the release.
-Mariia
-----Original Message-----
From: Markus Weimer [mailto:markus@weimo.de]
Sent: Tuesday, May 3, 2016 5:36 PM
To: REEF Developers Mailinglist <de...@reef.apache.org>
Subject: Re: REEF 0.15
Looking at the things currently marked as blocked, the following should be blockers:
* [REEF-1354](https://na01.safelinks.protection.outlook.com/?url=https%3a%2f%2fissues.apache.org%2fjira%2fbrowse%2fREEF-1354&data=01%7c01%7cmamykhai%40microsoft.com%7cef0c06fbff354490e00308d373b43480%7c72f988bf86f141af91ab2d7cd011db47%7c1&sdata=0ViihURwmCgPeXo76GgHR9J%2baqtNxStRGlHjOYchIAM%3d), the Java side of the .NET support for multi-runtime, as it blocks an important user.
* [REEF-1375](https://na01.safelinks.protection.outlook.com/?url=https%3a%2f%2fissues.apache.org%2fjira%2fbrowse%2fREEF-1375&data=01%7c01%7cmamykhai%40microsoft.com%7cef0c06fbff354490e00308d373b43480%7c72f988bf86f141af91ab2d7cd011db47%7c1&sdata=wZGSj8nVEcYPEN0nB%2f%2fwPb9LayDNSRpWfxzbitqhI%2bc%3d) for the same reason.
* [REEF-1323](https://na01.safelinks.protection.outlook.com/?url=https%3a%2f%2fissues.apache.org%2fjira%2fbrowse%2fREEF-1323&data=01%7c01%7cmamykhai%40microsoft.com%7cef0c06fbff354490e00308d373b43480%7c72f988bf86f141af91ab2d7cd011db47%7c1&sdata=%2f3zBShmWM1DlFk2vVzZcUBm%2fnAO9gijHdU1CnggCHAA%3d) and [REEF-1243](https://na01.safelinks.protection.outlook.com/?url=https%3a%2f%2fissues.apache.org%2fjira%2fbrowse%2fREEF-1243&data=01%7c01%7cmamykhai%40microsoft.com%7cef0c06fbff354490e00308d373b43480%7c72f988bf86f141af91ab2d7cd011db47%7c1&sdata=UTSJkodZdlyqg%2f5hpf1NwxxOAMsjRC7O2KQ3TvX%2fc7k%3d), as they refer to known test failures we can fix.
I'm OK with postponing the following:
* [REEF-1230](https://na01.safelinks.protection.outlook.com/?url=https%3a%2f%2fissues.apache.org%2fjira%2fbrowse%2fREEF-1230&data=01%7c01%7cmamykhai%40microsoft.com%7cef0c06fbff354490e00308d373b43480%7c72f988bf86f141af91ab2d7cd011db47%7c1&sdata=NNVqsAqvUN7O2eFjOPW1zslZfo%2bOwf0fJ1%2bnzoKgH2g%3d), as we won't have the full multi-runtime in 0.15
* [REEF-400](https://na01.safelinks.protection.outlook.com/?url=https%3a%2f%2fissues.apache.org%2fjira%2fbrowse%2fREEF-400&data=01%7c01%7cmamykhai%40microsoft.com%7cef0c06fbff354490e00308d373b43480%7c72f988bf86f141af91ab2d7cd011db47%7c1&sdata=OgkJNpH4hUHIs5zCMOhHf6P33QNcQGe2wZ6SOGAJ0NY%3d), as we can't act on it (AVRO dependency)
Which leaves two issues I am unsure about:
* [REEF-330](https://na01.safelinks.protection.outlook.com/?url=https%3a%2f%2fissues.apache.org%2fjira%2fbrowse%2fREEF-330&data=01%7c01%7cmamykhai%40microsoft.com%7cef0c06fbff354490e00308d373b43480%7c72f988bf86f141af91ab2d7cd011db47%7c1&sdata=iZ3MFqeFMDM0Thf%2baP%2bTaeqBtIoU%2fI8nFqsJESk0bg4%3d) is about porting our .NET functional tests to the new client API, which in turn enables us to test on YARN. I'd love to see that sooner rather than later, but it also strikes me as a lot of work. WDYT?
* [REEF-195](https://na01.safelinks.protection.outlook.com/?url=https%3a%2f%2fissues.apache.org%2fjira%2fbrowse%2fREEF-195&data=01%7c01%7cmamykhai%40microsoft.com%7cef0c06fbff354490e00308d373b43480%7c72f988bf86f141af91ab2d7cd011db47%7c1&sdata=AYpM0zQWBmknRMzz8QmBYWt2MhfhV%2bAtiCMDOD1cTj4%3d) is about making sure our assemblies and the NuGets always have the same version. This makes the life of the release managers much easier.
Also, it should be almost simple to fix: IIRC, NuGet automatically picks up the version of the DLL. So all we need to do is delete all version statements from our nuspecs. This should be doable in the tradition of improving our release process in every release.
Markus
Re: REEF 0.15
Posted by Markus Weimer <ma...@weimo.de>.
Looking at the things currently marked as blocked, the following
should be blockers:
* [REEF-1354](https://issues.apache.org/jira/browse/REEF-1354), the
Java side of the .NET support for multi-runtime, as it blocks an
important user.
* [REEF-1375](https://issues.apache.org/jira/browse/REEF-1375) for
the same reason.
* [REEF-1323](https://issues.apache.org/jira/browse/REEF-1323) and
[REEF-1243](https://issues.apache.org/jira/browse/REEF-1243), as they
refer to known test failures we can fix.
I'm OK with postponing the following:
* [REEF-1230](https://issues.apache.org/jira/browse/REEF-1230), as
we won't have the full multi-runtime in 0.15
* [REEF-400](https://issues.apache.org/jira/browse/REEF-400), as we
can't act on it (AVRO dependency)
Which leaves two issues I am unsure about:
* [REEF-330](https://issues.apache.org/jira/browse/REEF-330) is
about porting our .NET functional tests to the new client API, which
in turn enables us to test on YARN. I'd love to see that sooner rather
than later, but it also strikes me as a lot of work. WDYT?
* [REEF-195](https://issues.apache.org/jira/browse/REEF-195) is
about making sure our assemblies and the NuGets always have the same
version. This makes the life of the release managers much easier.
Also, it should be almost simple to fix: IIRC, NuGet automatically
picks up the version of the DLL. So all we need to do is delete all
version statements from our nuspecs. This should be doable in the
tradition of improving our release process in every release.
Markus
Re: REEF 0.15
Posted by Byung-Gon Chun <bg...@gmail.com>.
No blocker on the REEF Java side, I believe.
Thanks.
-Gon
On Wed, May 4, 2016 at 8:35 AM, Boris Shulman <sh...@gmail.com> wrote:
> Sure.
>
> Besides REEF-1354 <https://github.com/apache/reef/pull/966> and REEF-1375
> <https://github.com/apache/reef/pull/983> I would add REEF-1323
> <https://issues.apache.org/jira/browse/REEF-1323> (It might be a duplicate
> of REEF-1243 <https://issues.apache.org/jira/browse/REEF-1243> at this
> stage).
>
> I don't think anything else should block the release.
>
> Thoughts?
>
> Boris.
>
> On Tue, May 3, 2016 at 4:04 PM, Markus Weimer <ma...@weimo.de> wrote:
>
> > On Tue, May 3, 2016 at 1:39 PM, Boris Shulman <sh...@gmail.com>
> wrote:
> > > Can we finalize what is required for 0.15 and what can be moved to
> future
> > > releases? I prefer having this discussion here.
> >
> > Sure. Can you start with a proposal?
> >
> > Markus
> >
>
--
Byung-Gon Chun
Re: REEF 0.15
Posted by Boris Shulman <sh...@gmail.com>.
Sure.
Besides REEF-1354 <https://github.com/apache/reef/pull/966> and REEF-1375
<https://github.com/apache/reef/pull/983> I would add REEF-1323
<https://issues.apache.org/jira/browse/REEF-1323> (It might be a duplicate
of REEF-1243 <https://issues.apache.org/jira/browse/REEF-1243> at this
stage).
I don't think anything else should block the release.
Thoughts?
Boris.
On Tue, May 3, 2016 at 4:04 PM, Markus Weimer <ma...@weimo.de> wrote:
> On Tue, May 3, 2016 at 1:39 PM, Boris Shulman <sh...@gmail.com> wrote:
> > Can we finalize what is required for 0.15 and what can be moved to future
> > releases? I prefer having this discussion here.
>
> Sure. Can you start with a proposal?
>
> Markus
>
Re: REEF 0.15
Posted by Markus Weimer <ma...@weimo.de>.
On Tue, May 3, 2016 at 1:39 PM, Boris Shulman <sh...@gmail.com> wrote:
> Can we finalize what is required for 0.15 and what can be moved to future
> releases? I prefer having this discussion here.
Sure. Can you start with a proposal?
Markus
Re: REEF 0.15
Posted by Boris Shulman <sh...@gmail.com>.
Can we finalize what is required for 0.15 and what can be moved to future
releases? I prefer having this discussion here.
Boris.
On Mon, May 2, 2016 at 11:26 AM, Boris Shulman <sh...@gmail.com> wrote:
> I don't think REEF-1230 <https://issues.apache.org/jira/browse/REEF-1230> should block
> the release.
>
> On Mon, May 2, 2016 at 10:38 AM, Markus Weimer <ma...@weimo.de> wrote:
>
>> Hi,
>>
>> the release JIRA [REEF-1215] mentions a bunch more, some of which we
>> habitually move from release to release. Ironically, the ones you mentioned
>> weren't even on that list. I fixed that :)
>>
>> Let's figure out what to do with all these blockers, either here or on
>> the JIRA.
>>
>> Markus
>>
>>
>>
>> On 5/2/2016 1:21 AM, Boris Shulman wrote:
>>
>>> As REEF-1054 will take more time then anticipated I suggest to cur REEF
>>> 0.15 rigth after the following JIRAs will be in:
>>> 1) REEF-1354 <https://github.com/apache/reef/pull/966> - that allows
>>> using
>>> multiruntime from c# by implementing the c# client side (without official
>>> support in REEF .NET)
>>> 2) REEF-1375 <https://github.com/apache/reef/pull/983> - that fixes
>>> Java
>>> multi runtime in 0.15
>>> 3) REEF-1243
>>>
>>> Do we have any other JIRAs that are a must for 0.15 besides these?
>>>
>>> @Dhruv, did we have any progress on REEF-1243?
>>>
>>> Boris.
>>>
>>>
>
Re: REEF 0.15
Posted by Boris Shulman <sh...@gmail.com>.
I don't think REEF-1230
<https://issues.apache.org/jira/browse/REEF-1230> should block
the release.
On Mon, May 2, 2016 at 10:38 AM, Markus Weimer <ma...@weimo.de> wrote:
> Hi,
>
> the release JIRA [REEF-1215] mentions a bunch more, some of which we
> habitually move from release to release. Ironically, the ones you mentioned
> weren't even on that list. I fixed that :)
>
> Let's figure out what to do with all these blockers, either here or on the
> JIRA.
>
> Markus
>
>
>
> On 5/2/2016 1:21 AM, Boris Shulman wrote:
>
>> As REEF-1054 will take more time then anticipated I suggest to cur REEF
>> 0.15 rigth after the following JIRAs will be in:
>> 1) REEF-1354 <https://github.com/apache/reef/pull/966> - that allows
>> using
>> multiruntime from c# by implementing the c# client side (without official
>> support in REEF .NET)
>> 2) REEF-1375 <https://github.com/apache/reef/pull/983> - that fixes Java
>> multi runtime in 0.15
>> 3) REEF-1243
>>
>> Do we have any other JIRAs that are a must for 0.15 besides these?
>>
>> @Dhruv, did we have any progress on REEF-1243?
>>
>> Boris.
>>
>>
Re: REEF 0.15
Posted by Markus Weimer <ma...@weimo.de>.
Hi,
the release JIRA [REEF-1215] mentions a bunch more, some of which we
habitually move from release to release. Ironically, the ones you
mentioned weren't even on that list. I fixed that :)
Let's figure out what to do with all these blockers, either here or on
the JIRA.
Markus
On 5/2/2016 1:21 AM, Boris Shulman wrote:
> As REEF-1054 will take more time then anticipated I suggest to cur REEF
> 0.15 rigth after the following JIRAs will be in:
> 1) REEF-1354 <https://github.com/apache/reef/pull/966> - that allows using
> multiruntime from c# by implementing the c# client side (without official
> support in REEF .NET)
> 2) REEF-1375 <https://github.com/apache/reef/pull/983> - that fixes Java
> multi runtime in 0.15
> 3) REEF-1243
>
> Do we have any other JIRAs that are a must for 0.15 besides these?
>
> @Dhruv, did we have any progress on REEF-1243?
>
> Boris.
>