This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: RFC: remove a bunch of testsuite/**/configure scripts
- From: Tom Tromey <tromey at redhat dot com>
- To: Joel Brobecker <brobecker at adacore dot com>
- Cc: gdb-patches at sourceware dot org
- Date: Tue, 01 Mar 2011 10:19:49 -0700
- Subject: Re: RFC: remove a bunch of testsuite/**/configure scripts
- References: <m3vd037t85.fsf@fleche.redhat.com> <20110301045303.GI30306@adacore.com>
>>>>> "Joel" == Joel Brobecker <brobecker@adacore.com> writes:
Joel> I'm in favor of at least removing the configure script. For
Joel> the Makefiles (would we keep the master one in gdb/testsuite/?,
Either way is ok.
I think there is probably enough code in testsuite/Makefile.in to
warrant keeping it around. Ordinarily I think it is better, for
parallelism, to have as few Makefiles as possible, but that doesn't seem
like a real consideration in this case.
Joel> I'd say yes, at least as a first step), it's a tougher one. I think
Joel> it depends on what the consequences in terms of funtionality would
Joel> be. I don't mind losing "make clean" capability in the testsuite
Joel> area, but that might be problems for those who build in-tree
Joel> (sacrilege! :-).
All these Makefiles exist only for "make clean".
Until Michael's patch last week, they didn't even successfully do that :-)
In fact, depending on the target, that may still be the case; I don't
know for certain but I suspect so.
I wonder whether we could arrange to invoke runtest in a subdirectory.
Then "make clean" would be "rm -rf testdir".
This would break "clean" for those people invoking runtest by hand though.
Personally I never run "make clean" for the testsuite and I really
couldn't care less if it stopped working. Is there anybody who does
care and rely on this? If you do use it -- why? I would rather not go
through a bunch of gyrations for a feature that no developer uses.
Tom