On 22 September 2017 at 09:25, Juan Quintela <quint...@redhat.com> wrote: > Peter Xu <pet...@redhat.com> wrote: >> On Thu, Sep 21, 2017 at 04:29:04PM -0700, no-re...@patchew.org wrote: >> >> [...] >> >>> /var/tmp/patchew-tester-tmp-b0y2ycie/src/util/bitmap.c: In function >>> ‘bitmap_to_from_le’: >>> /var/tmp/patchew-tester-tmp-b0y2ycie/src/util/bitmap.c:383:6: error: >>> "__WORD_SIZE" is not defined [-Werror=undef] >>> # if __WORD_SIZE == 64 >>> ^~~~~~~~~~~ >>> CC util/acl.o >>> cc1: all warnings being treated as errors >>> /var/tmp/patchew-tester-tmp-b0y2ycie/src/rules.mak:66: recipe for target >>> 'util/bitmap.o' failed >>> make: *** [util/bitmap.o] Error 1 >>> make: *** Waiting for unfinished jobs.... >>> === OUTPUT END === >>> >>> Test command exited with code: 2 >> >> This is the same error with previous one. We need to replace the >> single usage of __WORD_SIZE into either __WORDSIZE or HOST_LONG_BITS >> (as suggested by Dave). >> >> Again, please let me know if you want me to repost, and sorry for the >> trouble! > > I think we can let that in in the PULL request and fix it later if > needed, not?
No, a compile failure will always make me reject a merge. Otherwise I can't easily test anybody else's subsequent merge, and also patchew will send complaining emails to every patch that's posted to the list until the breakage in master is fixed... thanks -- PMM