On 2016-01-03, Scott Kostyshak wrote: > On Sun, Jan 03, 2016 at 09:02:40AM +0000, Guenter Milde wrote:
... >> >> The following patch will auto-select polyglossia with fontspec whenever >> >> possible. >> >> OK to commit? ... > I think there are some differences that you did not predict. Here are > the differences between a69d6496 and a69d6496+your_first_patch. ... > If you are OK with these other test failures and you think it is the right > thing to do, then correct the test labels such that there shouldn't be any > differences to test failures after your patch is applied. Post a new patch and > I will confirm that nothing changes. Actually, you list is good news: There are even more cases of test thats "fails to fail" than I expected. In all cases of new inversions, this is just extending an existing inversion for "pdf[45]_systemF" to ".*_systemF" - this is a hint to a problem with polyglossia (either the language nesting bug or some known incompatibility with some other package). I hope the attached patch addresses all changing test cases. All in all not only the logic is simpler but also the number of inversions is lowered. Günter
