[extra-cmake-modules] modules: Revert "use the NAME_PREFIX only for the test, not for the executable"

[Saro Engels]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1



Am 10.05.2014 22:24, schrieb Alex Merry:
> On 10/05/14 20:37, Patrick Spendrin wrote:
>> As you can see in my commits in ktextwidgets, kio, kxmlgui and
>> several other locations, only the add_test call had the name
>> prefix and not the executable. One occurance where the previous
>> (and now current) will fail is kxmlgui autotests, where dfaure
>> send me an email first without reverting my commits.
>> 
>> I am so sorry about not having put up a review request on your
>> project and will obviously do so in the future if I ever commit
>> in there again.
> 
> Sorry about reverting without contacting you first - it appeared to
> be breaking things, and I wasn't up to tracking down the problem
> properly right then.

Yes, I was a bit upset yesterday, so sorry for my harsh tone.

> 
> That said, I don't think your commit was the correct fix to the 
> situation. The breakage appeared to happen because the frameworks 
> weren't properly ported to ecm_add_tests - whoever did the porting
> (and I haven't looked to see who that was) assumed it worked like
> the copy-pasted macros that were in the autotests/CMakeLists.txt
> files, and it doesn't. But I think the existing behaviour of
> ecm_add_tests (to namespace the targets as well as the test names)
> is preferable.

I still think my commit was the correct one, as it sets up the
previous situation (no prefix for target, prefix for testname). As you
mentioned, problems did occur because not all frameworks had the same
commits (only after the first few commits, I talked to Alex Richardson
that a prefix in ecm_add_test/s would be an easier fix.).

what I think is more important now are two things:
1) are prefixes needed anywhere? They seem to be not needed because I
have seen only one repo (kio) where more than one prefix is used at all.
2) however the answer to q.1 might be, it would be useful for windows
to have all frameworks with the same code, e.g. ecm_add_test for
single tests, ecm_add_tests for multiple tests.
If somebody knows places where this isn't the case (e.g. ecm_add_test
in an additional macro, I would like to here about it so I can fix it).

> 
> Alex
> 

regards,
Patrick
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.21 (MingW32)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
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=Sxog
-----END PGP SIGNATURE-----