On Mon, Nov 14, 2016 at 09:45:57AM +0000, Guenter Milde wrote: > On 2016-11-14, Scott Kostyshak wrote: > > > [-- Type: text/plain, Encoding: --] > > > After 0872bb78, the only failing tests of the recently committed > > doc/ar/Intro.lyx and doc/ar/Shortcuts.lyx are: > > > export/doc/ar/Intro_pdf4_texF (Failed) > > export/doc/ar/Shortcuts_pdf4_texF (Failed) > > This means, that 8-bit export (pdf2, ps, ...) as well as LuaTeX (pdf5_texF) > work?
Because I set the default output to pdf4, pdf2 and pdf5 are not tested. > > Note that the two documents are set to use non-TeX fonts. In such a > > case, should we still test TeX fonts? If so, which label should we > > invert them under? (i.e. is it LyX's fault that we can't export with TeX > > fonts?) > > It is not a "LyX bug" when documents intended for one export do not work > with others. > OTOH, we strife to make the documentation robust and working under the > widest sensibly possible range of export formats. OK. > There are several possible solutions: > > a) add patterns (and a comment) to ignoredTests (if we never expect to > overcome the issue) > > b) add patterns (and a comment) to ignoredTests (if we may overcome the issue > and keep testing) for b) do you mean "suspended"? Otherwise we might overcome the issue and never realize it. > c) If we know for sure a document does not work with non-TeX fonts, it should > set the default output format to one of XeTeX or LuaTeX. I did this. > As it seems to work with other "8-bit routes" but not to be easy to solve > for XeTeX, I vote for a). Sorry for the confusion on this. It does not work with 8-bit routes (see above). Scott
signature.asc
Description: PGP signature
