Hi Dmitry, On Sun, 13 Mar 2016 22:21:58 +0100 Dmitry Shachnev <mity...@debian.org> wrote: > (Also, if I just apply the change you suggested, sphinx will FTBFS on some > archs, because the tests use webkit which segfaults there — so it's a bit > more tricky).
I just built src:sphinx on all porter boxes of release architectures. The only FTBFS I observed was on mips and the problem on mips went away with the following patch: Index: sphinx-1.4.8/tests/test_intl.py =================================================================== --- sphinx-1.4.8.orig/tests/test_intl.py +++ sphinx-1.4.8/tests/test_intl.py @@ -442,10 +442,6 @@ def test_html_builder(app, status, warni updated = app.env.update(app.config, app.srcdir, app.doctreedir, app) yield assert_equal, len(updated), 0 - (app.srcdir / 'xx' / 'LC_MESSAGES' / 'bom.mo').utime(None) - updated = app.env.update(app.config, app.srcdir, app.doctreedir, app) - yield assert_equal, len(updated), 1 - @gen_with_intl_app('xml', freshenv=True) def test_xml_builder(app, status, warning): The problem is, that on mips, "len(updated)" evaluates to zero and not one. I was unable to reproduce this issue manually outside of the test suite as a lot of setup steps seem to be necessary before one can run above snippet. It is also important to notice, that the problem was not reproducible. When I built the package again (without modifications) the test suite finished successfully. I also don't know how serious this problem is and whether it merits its own bug against sphinx. But it seems that (unless the above is a result of said webkit segfault) the webkit issue went away and only this issue remains the sole blocker of making python-sphinx Architecture:any. Thanks! cheers, josch