You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Daniel Shahaf <d....@daniel.shahaf.name> on 2009/03/02 21:25:36 UTC
Patch management (was: Re: [PATCH] patch for links.html, added
svn2svn)
Gavin Baumanis wrote on Mon, 2 Mar 2009 at 23:49 +1100:
> A patch is proposed that is deemed to be of a benefit to the
> subversion project - but is hamstrung from being fully welcomed /
> supported because it doesn't meet appropriate use-case tests.
>
> Eg. "this"patch is useful only for Windows users.
>
On links.html this is quite acceptable (as Daniel Doubrovkine pointed
out). We have Mac-only clients, pay-only clients, etc., there.
> So in the sense of my patch manager role is there anything more that I
> can do?
> an example might be to request some "other" people have a look at the
> patch proposal - specifically looking at the mentioned shortfalls - so
> that a "complete" patch might be created.
>
Yes. Pointing out specific issues that stand in the way of the patch to
be applied is one way you (as patch manager) can advance the discussion.
(Just screaming "Ping!" repeatedly and threatening to archive a patch in
the tracker usually won't be as effective. :-))
This is equally true for technical issues (bugs in the code: e.g., not
portable) and formal issues (bugs in the submission: e.g., not sent as
a unidiff) --- fixing issues of either kind makes the committer's life
easier when he comes to apply the patch.
> or,
>
> Is it simply a case of - we're all volunteers (primarily) so if
> someone wants to make a *nix version of the patch they will - and in
> the meantime the OP is more than supported by the group providing
> assistance peer-review / assistance in patching their own installs and
> thus there is nothing more for me to do, aside from checking if there
> is any life left in the patch proposal?
>
I don't understand the question. If we asked the OP to make some
changes (e.g.: fix a portability bug) and resubmit the patch, and he
hasn't, you can ping the thread and ask for an update. Is that what you
asked?
> Afterall, - the mailing list serves as an archive should someone else
> on Windows have the same requirement - or a *nix coder wish to create
> a non-windows version of the patch?
>
Patches can also be archived in the issue tracker. (In general, where
there's already an issue, the patch should always be linked from it.)
Daniel
>
> Thanks.
>
> Beau.
>
> On 26/02/2009, at 5:45 AM, Daniel Doubrovkine wrote:
>
> >
> >
> >
> >
> > -dB.
> >
> > dblock.org / foodcandy.com
> >
> >
> >
> > <svn2svn.patch>
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1256168
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1258343
Re: Patch management (was: Re: [PATCH] patch for links.html, added
svn2svn)
Posted by Daniel Shahaf <d....@daniel.shahaf.name>.
Gavin 'Beau' Baumanis wrote on Tue, 3 Mar 2009 at 09:12 +1100:
...
> If there was a requirement for a *nix port of my patch someone else would have
> to do it - other than me.
>
> It gets reviewed by someone on the list - who points out that it will only
> work on Windows, after which there is no more discussion.
> No one has volunteered to do the *nix work and I can't do it for myself.
>
> Thus, as the patch manager - do I bring it back to people's attention and see
> if "I" can drum up support for someone to do the required work - or is it a
You can raise the thread again. But if no one expresses interest in
finishing the patch or helping the OP do so, archive the patch and
move on. Not all patches get committed.
> case of , well we're volunteering our time - so if it gets done - well that's
> great ... but if it doesn't too bad and there is nothing more for me to do (as
> Patch Manager) - obviously no one else in interested / worried about the
> Windows only patch)?
>
As a rule, we expect the OP to be responsible for writing the missing parts.
That said, there are exceptions: e.g., in your example, we don't expect
*nix devs to fix bugs in the Windows build system; and we shouldn't
reject a bindings patch because it unearthed a segfault in the core
C libraries.
> Now saying all that , by answering some of the earlier paragraphs, it would
> seem to be appropriate for me to bring it back into the foreground for some
> further discussion - even if it turned out to be along the lines of;
> "This really isn't an issue for anyone else - and we have higher priorities to
> address - archive it off, for now at least, anyway"
>
> And again thanks for the help.
>
>
> Beau,.
(And I probably said this already, but you can check the archives for how
we did it in the past.)
Daniel
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1262566
RE: Patch management (was: Re: [PATCH] patch for links.html, added
svn2svn)
Posted by Daniel Doubrovkine <db...@dblock.org>.
I think you should simply reject a patch that's not cross-platform for
components of subversion that are cross-platform.
-----Original Message-----
From: Gavin 'Beau' Baumanis [mailto:gavinb@thespidernet.com]
Sent: Monday, March 02, 2009 5:12 PM
To: Daniel Shahaf
Cc: dB.; dev@subversion.tigris.org; John Peacock
Subject: Re: Patch management (was: Re: [PATCH] patch for links.html,
added svn2svn)
Hi Daniel- and everyone else , of course!
>
>> or,
>>
>> Is it simply a case of - we're all volunteers (primarily) so if
>> someone wants to make a *nix version of the patch they will - and in
>> the meantime the OP is more than supported by the group providing
>> assistance peer-review / assistance in patching their own installs
>> and
>> thus there is nothing more for me to do, aside from checking if there
>> is any life left in the patch proposal?
>>
>
> I don't understand the question. If we asked the OP to make some
> changes (e.g.: fix a portability bug) and resubmit the patch, and he
> hasn't, you can ping the thread and ask for an update. Is that what
> you
> asked?
>>
Ultimately, you already answered my question in a previous part of the
post but just to clarify;
(and while similar to the original posting of Daniel's - it turns out
it is different - and it may well be that it is something that's not
likely to ever happen)
Let's assume, I have written a patch to address an issue that I am
having in subversion.
But it only works for me, using my peculiar build / flavour of SVN -
let's just go with Windows for this example.
If there was a requirement for a *nix port of my patch someone else
would have to do it - other than me.
It gets reviewed by someone on the list - who points out that it will
only work on Windows, after which there is no more discussion.
No one has volunteered to do the *nix work and I can't do it for myself.
Thus, as the patch manager - do I bring it back to people's attention
and see if "I" can drum up support for someone to do the required
work - or is it a case of , well we're volunteering our time - so if
it gets done - well that's great ... but if it doesn't too bad and
there is nothing more for me to do (as Patch Manager) - obviously no
one else in interested / worried about the Windows only patch)?
Now saying all that , by answering some of the earlier paragraphs, it
would seem to be appropriate for me to bring it back into the
foreground for some further discussion - even if it turned out to be
along the lines of;
"This really isn't an issue for anyone else - and we have higher
priorities to address - archive it off, for now at least, anyway"
And again thanks for the help.
Beau,.
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1258481
Re: Patch management (was: Re: [PATCH] patch for links.html, added
svn2svn)
Posted by Gavin Baumanis <ga...@thespidernet.com>.
Hi Daniel- and everyone else , of course!
>
>> or,
>>
>> Is it simply a case of - we're all volunteers (primarily) so if
>> someone wants to make a *nix version of the patch they will - and in
>> the meantime the OP is more than supported by the group providing
>> assistance peer-review / assistance in patching their own installs
>> and
>> thus there is nothing more for me to do, aside from checking if there
>> is any life left in the patch proposal?
>>
>
> I don't understand the question. If we asked the OP to make some
> changes (e.g.: fix a portability bug) and resubmit the patch, and he
> hasn't, you can ping the thread and ask for an update. Is that what
> you
> asked?
>>
Ultimately, you already answered my question in a previous part of the
post but just to clarify;
(and while similar to the original posting of Daniel's - it turns out
it is different - and it may well be that it is something that's not
likely to ever happen)
Let's assume, I have written a patch to address an issue that I am
having in subversion.
But it only works for me, using my peculiar build / flavour of SVN -
let's just go with Windows for this example.
If there was a requirement for a *nix port of my patch someone else
would have to do it - other than me.
It gets reviewed by someone on the list - who points out that it will
only work on Windows, after which there is no more discussion.
No one has volunteered to do the *nix work and I can't do it for myself.
Thus, as the patch manager - do I bring it back to people's attention
and see if "I" can drum up support for someone to do the required
work - or is it a case of , well we're volunteering our time - so if
it gets done - well that's great ... but if it doesn't too bad and
there is nothing more for me to do (as Patch Manager) - obviously no
one else in interested / worried about the Windows only patch)?
Now saying all that , by answering some of the earlier paragraphs, it
would seem to be appropriate for me to bring it back into the
foreground for some further discussion - even if it turned out to be
along the lines of;
"This really isn't an issue for anyone else - and we have higher
priorities to address - archive it off, for now at least, anyway"
And again thanks for the help.
Beau,.
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1258469