[testsuite] Include gdc.test prefix in test names (PR testsuite/88041)

Rainer Orth ro@CeBiTec.Uni-Bielefeld.DE
Thu Dec 13 15:01:00 GMT 2018


Hi Iain,

>> I'm omitting the changes (also mentioned in the PR) to
>> compilable/ddoc9676a.d and compilable/depsOutput9948.d which contain
>> absolute path names in EXTRA_SOURCES because those need to go upstream
>> first.
>>
>
> This has been committed upstream
> (https://github.com/dlang/dmd/pull/9051) - I just need to trickle it
> down, among other patches that are in my queue.

great, thanks.

> I have no problem however if you go ahead commit the changes with this patch.

Good: that's what I did.

>> Bootstrapped without regressions on i386-pc-solaris2.11 and
>> x86_64-pc-linux-gnu.  Ok for mainline?
>
> OK on my side, just one mental note.
>
>> @@ -360,6 +361,9 @@ proc gdc-do-test { } {
>>      # Initialize `dg'.
>>      dg-init
>>
>> +    # Create gdc.test link so test names include that subdir.
>> +    catch { file link $subdir . }
>> +
>
> I guess the proper fix would be to split this into three files, right?
>
>   - compilable/compilable.exp
>   - fail_compilation/fail_compilation.exp
>   - runnable/runnable.exp
>
> Then move the common bits into lib/gdc-dmd.exp

While that's possible, we usually only have per-subdir .exp files when
e.g. there are massive/systematic differences in the set of compilation
flags used.  In the case of those three subdirs, you could indeed split
into per-subdir files and a common one, but it's also fine as is.

The $subdir link above (where subdir is gdc.test, the directory below
.../testsuite) only achieves that both (say)
gdc.test/compilable/99bottles.d as passed to dg-runtest and
imports.jsonimport1 (without the prefixe and referenced in the sources
and -I flags) work just the same.  So this will remain necessary even if
you decide to split gdc-test.exp into three .exp files.

	Rainer

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



More information about the Gcc-patches mailing list