Max, I confirm the test now also passes on mips32: https://uclibc-ng-ci.sionneau.net:8443/job/uclibc-ng-multiarch/arch=mips32r6... Now the CI uses the git sha1 of the commit that triggered the build :) Thanks! On 5/13/19 8:42 PM, Yann Sionneau wrote:
Hi Max,
Good catch!
Thanks for having a look, I'll fix the CI to use the SHA1 of last commit instead of the last release.
The CI is a work in progress and any improvement is welcome.
Le 13/05/2019 à 20:36, Max Filippov a écrit :
Hi Yann,
On Mon, May 13, 2019 at 8:55 AM Yann Sionneau <ysionneau@kalray.eu> wrote:
It has been merged, it trigged continuous integration on my side.
It seems tst-preadvwritev still passes on the following arches: armv7, aarch64
But this test was failing before your commit and still fails for mips32r6: https://uclibc-ng-ci.sionneau.net:8443/job/uclibc-ng-multiarch/arch=mips32r6...
Any idea why? Looking at the full log of your test here https://uclibc-ng-ci.sionneau.net:8443/job/uclibc-ng-multiarch/arch=mips32r6... I see that buildroot builds uclibc-ng 1.0.31 with two patches on top of it, but without the patch that fixes the preadv/pwritev issue.
I ran the test using fixed uclibc version, it passes for me on both little- and big-endian mips32r6 when run under linux-user qemu.
I'll submit a patch for the buildroot.
_______________________________________________ devel mailing list devel@uclibc-ng.org https://mailman.uclibc-ng.org/cgi-bin/mailman/listinfo/devel