This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: test failures with gcc 4.4
- From: Chris Metcalf <cmetcalf at tilera dot com>
- To: GLIBC Devel <libc-alpha at sourceware dot org>
- Date: Sat, 6 Sep 2014 13:39:29 -0400
- Subject: Re: test failures with gcc 4.4
- Authentication-results: sourceware.org; auth=none
- References: <540A1ED2 dot 3040608 at tilera dot com>
I went ahead and created a new "Compiler-dependent" section on the 2.20 release wiki page for build and test failures, and a "gcc 4.4" section under that, and mentioned the ISO11 conform tests, the __SIG_ATOMIC_TYPE__ issue, and the build failures on the elf/tst-unique[34] tests.
On 9/5/2014 4:36 PM, Chris Metcalf wrote:
I notice that there are some issues with older gcc for the tests. We've already had a discussion about the failures building elf/tst-unique{3lib,4lib,3lib2}.so here:
https://sourceware.org/ml/libc-alpha/2012-06/msg00430.html
but I don't see those failures mentioned yet anywhere in the test results page.
Similarly, my conform results show a bunch of failures from using gcc 4.4, which I summarize under the TILE-Gx section (for now) as:
- The ISO11 conform tests fail because gcc 4.4 does not support "-std=c1x"
- The inttypes.h and stdint.h conform tests fail because gcc 4.4 does not define __SIG_ATOMIC_TYPE__
Would it make sense to create a new section on the 2.20 release wiki page for compiler-specific failures and put all of these in that section? Unfortunately the elf/tst-unique{3lib,4lib,3lib2}.so are build failures and so cause "make -k check" to fail to complete and as a result I just end up commenting them out in elf/Makefile, which isn't particularly clean.
--
Chris Metcalf, Tilera Corp.
http://www.tilera.com