This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: [RFA:] Stricter libstdc++ testsuite gate for working target file I/O
- From: Paolo Carlini <paolo dot carlini at oracle dot com>
- To: Hans-Peter Nilsson <hans-peter dot nilsson at axis dot com>
- Cc: hp at axis dot com, gcc-patches at gcc dot gnu dot org, libstdc++ at gcc dot gnu dot org, rdsandiford at googlemail dot com, nathan at codesourcery dot com
- Date: Thu, 07 Oct 2010 01:17:11 +0200
- Subject: Re: [RFA:] Stricter libstdc++ testsuite gate for working target file I/O
- References: <201010062305.o96N5jcD024839@ignucius.se.axis.com>
Hi,
>> Date: Wed, 06 Oct 2010 11:47:00 +0200
>> From: Paolo Carlini <paolo.carlini@oracle.com>
>>
>
>> The approach makes sense, in general. My only doubt
>>
> I take that as approval for the patch for the gate function!
>
Sure ;)
>
>> is that I don't
>> think we want to add the gate to a large part of the test files in
>> 27_io/basic_filebuf, essentially.
>>
> grep says that 105 of them have it, 110 lack it. So, maybe
> better eliminate the need for the gate in that subtree.
>
Excellent analysis.
>> Can't we skip the entire set at once,
>> like we do for wchar_t or thread testcases, and leave the gate for those
>> specific tests which are really sensitive to an lseek fully functional?
>>
> I think I see what you mean, hacking conformance.exp and filter
> by filename. As below, looking for the "_filebuf" substring
> perhaps? The base filenames themselves are rarely revealing
> enough and most tests in */*_filebuf seem to need it. Half the
> files in 27_io/basic_filebuf already have the
> // { dg-require-fileio "" }
> gate; I'm not completely sure how many of those without would
> actually need the gate, but of those that I checked, 9/10 seem
> to need it. A bit too many, so I appreciate been given an easy
> way out. :)
>
Sure, I think this is the way to go in this case. Note: there are
testcases using file i/o also *outside* basic_filebuf, I would guess not
too many, and a few more uses of the gate should do it.
> Tested in trees with/without a working simulator (together with
> the gate-function-patch), observing the right choice taken.
>
> Ok?
>
Looks good to me. Can you wait just, say, 12 h, in case of comments from
the aforementioned interested parties and the other C++ runtime maintainers?
Thanks,
Paolo.