Unreviewed testsuite patches

Unreviewed testsuite patches

[Rainer Orth]

The following patches have remained unreviewed for some time:

	[build, doc, libjava, testsuite] Fully support TLS on Solaris 8 and 9
        http://gcc.gnu.org/ml/gcc-patches/2010-07/msg01277.html

Most of this patch is Solaris-specific, but I may need approval for the
testsuite parts.  One might argue that they are obvious, though.

	[testsuite] Increase gcc.dg/pr43058.c timeout
        http://gcc.gnu.org/ml/gcc-patches/2010-07/msg01162.html

Another testsuite patch that might be obvious, unless the testcase isn't
supposed to take close to the default timeout to complete.

	Rainer

-- 
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University

Unreviewed testsuite patches

[Rainer Orth]

Two larger testsuite patches of mine have remained unreviewed for some
time:

	[ada, testsuite] gnat.dg cleanup
        http://gcc.gnu.org/ml/gcc-patches/2011-01/msg00564.html
        http://gcc.gnu.org/ml/gcc-patches/2011-01/msg01129.html

While Arno is happy with the semantics of this one, he doesn't feel
comfortable reviewing the DejaGnu parts of the patch.

        [testsuite, build] Convert boehm-gc testsuite to DejaGnu (PR boehm-gc/11412)
        http://gcc.gnu.org/ml/gcc-patches/2011-01/msg00244.html
	http://gcc.gnu.org/ml/gcc-patches/2011-01/msg01128.html
	http://gcc.gnu.org/ml/gcc-patches/2011-01/msg01638.html

Again, the patch received some support, but no one feeled qualified to
review it or even state if this could/should go into 4.6 or wait for
4.7.

So both patches seem to need a global reviewer for lack of a testsuite
maintainer.

Thanks.
	Rainer

-- 
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University

Re: Unreviewed testsuite patches

[Mike Stump]

On Jan 24, 2011, at 6:08 AM, Rainer Orth wrote:
> Two larger testsuite patches of mine have remained unreviewed for some
> time:

My take, once the technical side (non-dejagnu) is reviewed and approved, the work should just go in.  I'd look to the SC to elaborate on what the right approach is.  Sitting around for a global person for testsuite fixes and additions, is, well, silly.

Re: Unreviewed testsuite patches

[Gerald Pfeifer]

On Mon, 24 Jan 2011, Mike Stump wrote:
> My take, once the technical side (non-dejagnu) is reviewed and approved, 
> the work should just go in.  I'd look to the SC to elaborate on what the 
> right approach is.  Sitting around for a global person for testsuite 
> fixes and additions, is, well, silly.

The challenge is that of volunteer projects:  If there is someone who
starts reviewing patches, or actively volunteers to become a reviewer,
I feel the SC may be quite open to that.  On the other hand, the SC does
not have a magic hat which allows it to produce such volunteers with a
flick of a magic wand. :-)

Until and unless that happens, what are the alternatives to relying on
global reviewers and suggesting that maintainer of specific parts of the
compiler (frontends, optimizations, backends,...) should feel free to
make non-intrusive changes as necessary for their work?

Gerald (speaking for myself, not the SC)

Re: Unreviewed testsuite patches

[Mike Stump]

On Jan 25, 2011, at 2:54 PM, Gerald Pfeifer wrote:
> The challenge is that of volunteer projects:  If there is someone who
> starts reviewing patches, or actively volunteers to become a reviewer,
> I feel the SC may be quite open to that.  On the other hand, the SC does
> not have a magic hat which allows it to produce such volunteers with a
> flick of a magic wand. :-)

Yes, I've heard others say that as well...

> Until and unless that happens, what are the alternatives to relying on
> global reviewers and suggesting that maintainer of specific parts of the
> compiler (frontends, optimizations, backends,...) should feel free to
> make non-intrusive changes as necessary for their work?

Well, if there isn't a problem, then no solution is necessary.  So, the question is, do we have a problem?  I ask, because having a problem is probably a prerequisite to solving it.

Re: Unreviewed testsuite patches

[Gerald Pfeifer]

On Tue, 25 Jan 2011, Mike Stump wrote:
> Well, if there isn't a problem, then no solution is necessary.  So, the 
> question is, do we have a problem?  I ask, because having a problem is 
> probably a prerequisite to solving it.

Yes, we definitely have a problem.  And one possible volunteer who
raised his hand triggered by this e-mail exchange. ;-)  

Anyone else interested and knowledgable to help out here, drop me a
(private) e-mail, please!

Gerald

Re: Unreviewed testsuite patches

[Mike Stump]

On Jan 26, 2011, at 1:53 PM, Gerald Pfeifer wrote:
> On Tue, 25 Jan 2011, Mike Stump wrote:
>> So, the question is, do we have a problem?

> Yes, we definitely have a problem.

Ah, merely asking for people to step forward would be one of the solutions I'd recommend.

Another thing I might recommend would be to appoint the best person who stepped forward at the end of 20 days, provided at least 1 person stepped forward.  This is more controversial but I don't think badness would result.  If people thought they could do a better job, they would then be encouraged to step forward.  Also, one could try a slightly different policy, like anyone on the checkin after approval list can review testsuite work.  Those that possess the skill could then just approve the work, instead of saying, looks good to me, and still have the work twist in the wind.  Yet another approach would be review post checkin.  The work goes in, and then people review it and updates are made.  Another way would be to pick the top 5 contributors in that area and ask them privately if they would like to be reviewers or maintainers, and maybe 2-3 might say yes.  Another way might be to de-formalize the rules.  Another way might be to automatically migrate people into review status after some amount of time, or some amount of contributions.

I could probably think up more ways...  but I think that's a start, I'd like to think at least one of those might be better than not solving the problem.

unreviewed testsuite patches

[Janis Johnson]

After being away for a few weeks I'm way behind on reviewing
testsuite patches.  I'll make a real effort to catch up, but in
the meantime please send ping messages with links to patches I
might have missed.

Many thanks to all the other maintainers who have been approving
testsuite patches.

Janis