This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: libstdc++ testsuite fixes, 1
- To: Zack Weinberg <zackw at stanford dot edu>
- Subject: Re: libstdc++ testsuite fixes, 1
- From: Benjamin Kosnik <bkoz at redhat dot com>
- Date: Tue, 15 May 2001 12:05:56 -0700 (PDT)
- cc: gcc-patches at gcc dot gnu dot org, libstdc++ at gcc dot gnu dot org
thanks Zack.
> If you build using a relative pathname to get from the object to the
> source directory, testsuite_flags --build-includes will generate -I
> switches which are relative to the top level of the libstdc++ build
> directory. runtest executes one level below (in the testsuite
> directory), therefore all the paths are wrong, and the test suite
> can't find any headers and blows up horribly.
kay. I didn't see this, but that just means I am configuring/building
differently. What's your setup?
> This patch corrects the problem. I chose to edit the result of
> testsuite_flags from libstdc++-v3-dg.exp, rather than change
> testsuite_flags, because there are other things (mkcheck.in, e.g.)
> which use testsuite_flags and may expect it to produce its old
> results.
Hmm. testsuite_flags should be giving non-relative output, so it needs to
get fixed at that point (which shouldn't prohibit mkcheck.in from working)
also the includes need to include the ext directory for tests using hash_map.
-benjamin