Re: [cdt-dev] Patches

["James Blackburn" <jamesblackburn@xxxxxxxxx>]

On Mon, Nov 10, 2008 at 8:35 AM, Schorn, Markus
<Markus.Schorn@xxxxxxxxxxxxx> wrote:
> My experience is that submitters of patches are willing to provide
> test-cases, when I ask for it in the bugzilla and give a hint
> on where similar testcases can be found.

I think Elena's comment was half-aimed at me.  It's true, I am happy
to provide unit tests -- I created tests for the Alias issues I
encountered in core.resources, and some additional concurrency tests
for ICDescriptorManager as part of project storage work.

However a new test will only prevent that  issue/regression from
reapparing, it won't validate that I haven't broken any existing
functionality.  There are 3 bugs I've filed with patches for
linked-resource issues with the debugger, as far as I can see there
are no existing debug.core.tests...

It's a chicken and egg problem.  If the code isn't owned by anyone and
there are no existing tests, but there are bugs, is there value in
fixing the bugs?  As Markus said contributors who've made the effor to
file a patch are usually happy to provide a test as well.  They're
generally not so happy to write the entire testsuite.

Cheers,

James


>
> Markus.
>
>> -----Original Message-----
>> From: cdt-dev-bounces@xxxxxxxxxxx
>> [mailto:cdt-dev-bounces@xxxxxxxxxxx] On Behalf Of John Cortell
>> Sent: Saturday, November 08, 2008 12:20 AM
>> To: CDT General developers list.; CDT General developers list.
>> Subject: RE: [cdt-dev] Patches
>> Importance: Low
>>
>> Mixed feelings on this, bordering on -1. Great in theory, bad
>> in practice. As long as we never flat out reject a patch for
>> lack of a test, I'm +1.
>>
>> I've submitted patches to projects outside CDT. Taking the
>> time to learn their test suite/infrastructure and update that
>> can be a fairly complex process. Take a look at the CDT
>> debugger junit tests and that will become obvious. I would be
>> very disappointed if a project would refuse my patch solely
>> on the grounds that I haven't submitted a test. So, I don't
>> want to create language that may lead CDT committers to take
>> that route. I guess a submitter could say "can't provide a
>> test case because I don't have the time to learn the tests".
>> As long as we accept that as a valid excuse, then I'm good
>> with this. This reduces your proposal to a suggestion.
>>
>> We should be thankful that someone has made the effort to
>> learn the code and provide a fix rather than just open a bug.
>> Let's not add additional barriers.
>>
>> One thing is for sure, submissions MUST be accompanied by
>> product-agnostic reproducibility steps, allowing for
>> exceptions where the bug is very obvious and providing
>> agnostic repro steps is not feasible.
>>
>> John
>>
>>
>> At 09:12 AM 11/7/2008, Schaefer, Doug wrote:
>> >Emphatic +1.
>> >
>> >Thanks Elena.
>> >
>> > > -----Original Message-----
>> > > From: cdt-dev-bounces@xxxxxxxxxxx
>> > > [mailto:cdt-dev-bounces@xxxxxxxxxxx] On Behalf Of Elena Laskavaia
>> > > Sent: Friday, November 07, 2008 9:48 AM
>> > > To: CDT General developers list.
>> > > Subject: [cdt-dev] Patches
>> > >
>> > > Regarding applying patches from contributors:
>> > > If you submit a patch to cdt,
>> > > please create junit test if you can or explain why it cannot be
>> > > created, it would greatly simplify patch applying process.
>> > > Otherwise commiter has to try to reproduce (which sometime is not
>> > > easy at all) to re-test and create tests themselves which
>> is not as
>> > > much fun.
>> > > _______________________________________________
>> > > cdt-dev mailing list
>> > > cdt-dev@xxxxxxxxxxx
>> > > https://dev.eclipse.org/mailman/listinfo/cdt-dev
>> > >
>> >_______________________________________________
>> >cdt-dev mailing list
>> >cdt-dev@xxxxxxxxxxx
>> >https://dev.eclipse.org/mailman/listinfo/cdt-dev
>>
>>
>> _______________________________________________
>> cdt-dev mailing list
>> cdt-dev@xxxxxxxxxxx
>> https://dev.eclipse.org/mailman/listinfo/cdt-dev
>>
> _______________________________________________
> cdt-dev mailing list
> cdt-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/cdt-dev
>