On 2016-09-14, Kornel Benko wrote: > Am Mittwoch, 14. September 2016 um 19:25:43, schrieb Guenter Milde > <mi...@users.sf.net> >> On 2016-09-14, Scott Kostyshak wrote:
>> > I attach my complete list of failing tests (for git hash f2a263e3). The >> > ones that are interesting I think are listed below with some comments: >> > 206:INVERTED.LYXBUGS_export/export/LongestLabelWithUnderscore_lyx16 ... >> It looks like ctest is testing the backup files from lyx2lyx conversion. >> Do you have some *lyx21.lyx (etc.) files in the tested directories? > No backups, the test is exporting to lyx16 (e.g. lyx2lyx) format. I see, besides the lyx2lyx tests, there are also export tests for lyx2lyx. >> I hope these backups are marked as such with ~ soon. >> But we could also (at least for the meantime) add a clause to ignore them >> in "ignoredtests". > No need. The tests have a reason. I see. However, the inverted tests in autotests/export/ are dedicated LaTeX tests where we know that they currently fail with LaTeX and why. I did not think about the lyx2lyx export and hence used a too generic "inversion pattern". As testing these samples for other export routes or conversion or loading does not gain any new insights, I moved them to a new subdirectory autotests/export/latex/ and added "ignore" clauses for samples from this subdirectory. I hope this helps to keep the test system relatively clean and simple. Günter
