This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: [PATCH C++/testsuite] Remove pchtest check objects and compile with current tool
- From: Mike Stump <mikestump at comcast dot net>
- To: Bernhard Reutner-Fischer <rep dot dot dot nop at gmail dot com>
- Cc: Jason Merrill <jason at redhat dot com>, Janis Johnson <janis_johnson at mentor dot com>, Paolo Carlini <paolo dot carlini at oracle dot com>, GCC Patches <gcc-patches at gcc dot gnu dot org>, libstdc++ at gcc dot gnu dot org
- Date: Sat, 9 Nov 2013 09:15:22 -0800
- Subject: Re: [PATCH C++/testsuite] Remove pchtest check objects and compile with current tool
- Authentication-results: sourceware.org; auth=none
- References: <20131030095618 dot GB22154 at nbbrfq dot cc dot univie dot ac dot at> <C08D96A0-F9FD-4820-ABBA-6F2A4DA9B900 at comcast dot net> <CAC1BbcT5=uFQhfXgGcWCaYjmzjT8W5j0MAdjQtv0iQqwFTF0mg at mail dot gmail dot com> <C4F14FCB-5EDB-481F-BBA7-1A508046A532 at comcast dot net> <CAC1BbcSe0oufDykFaX+cRNyoqi3CN3XO=qjeA7uEZupAYUumGw at mail dot gmail dot com> <DAC22B43-236B-4191-86D2-F883B1BD3E5B at comcast dot net> <CAC1BbcT==xfHu6LpU_AOO8qp+GBZkfYLMwVQaXiG5zsNpJP-+g at mail dot gmail dot com>
On Nov 9, 2013, at 5:34 AM, Bernhard Reutner-Fischer <rep.dot.nop@gmail.com> wrote:
> Ok for trunk?
> 2013-10-12 Bernhard Reutner-Fischer <aldot@gcc.gnu.org>
>
> * lib/dg-pch.exp (pch-init): Remove pchtest check objects.
Ok.
The code that does set chk_lang "c++-header", but just not the code that puts in the comment, is fine to repost as well.
> Oh, and while double-checking the auto-wipe patch that i just
> simplified, i encountered an inconvenient side-effect of checking
> pristine trunk with plain
> $ make -k check -j4
> versus my local tree with
> $ make -k check -j4 RUNTESTFLAGS="-v -v -v -v "
>
> This bug/feature was added in a0d20ccbd97fde9c2af1f7345e3eb1313dea570f
> aka svn r141337 .
> From the sounds, if --target_board= is seen, then run specific test, serial.
> Otherwise do the parallel stuff, i.e. check_DEJAGNU_normal_targets.
> .
> Still i want -v to not behave differently as not passing -v in
> RUNTESTFLAGS especially regarding parallelism, so what about doing:
> diff --git a/libstdc++-v3/testsuite/Makefile.am
> b/libstdc++-v3/testsuite/Makefile.am
> index 4c92fef..6edc410 100644
> --- a/libstdc++-v3/testsuite/Makefile.am
> +++ b/libstdc++-v3/testsuite/Makefile.am
> @@ -108,7 +108,7 @@ $(check_DEJAGNU_normal_targets):
> check-DEJAGNUnormal%: normal%/site.exp
> check-DEJAGNU $(check_DEJAGNU_normal_targets): check-DEJAGNU%: site.exp
> AR="$(AR)"; export AR; \
> RANLIB="$(RANLIB)"; export RANLIB; \
> - if [ -z "$*$(filter-out --target_board=%, $(RUNTESTFLAGS))" ] \
> + if [ -z "$*$findstring --target_board=,$(RUNTESTFLAGS))" ] \
> && [ "$(filter -j, $(MFLAGS))" = "-j" ]; then \
> $(MAKE) $(AM_MAKEFLAGS) $(check_DEJAGNU_normal_targets); \
> for idx in 0 1 2 3 4 5 6 7 8 9; do \
>
> This may run non-parallel if somebody has RUNTESTFLAGS=" -v -v
> --target_board=" but at least we don't have -v --verbose -d --debug in
> the filter-out.
> Mike?
Head, hurts; can you discuss with jakub. I just want to get out of the way and let you two decide what is best.