Am Freitag, 15. Februar 2019 06:47:00 CET schrieb Guenter Milde <mi...@users.sf.net>: > Dear Kornel, >
> > Am Donnerstag, 14. Februar 2019 00:06:19 CET schrieb Guenter Milde
> > <mi...@users.sf.net>:
...
> >> Thank you for testing. (Interestingly, there are failures in files
> >> that should not be affected by my recent commit.)
>
> > This is because you changed useSystemFonts.pl. The fonts there were
> > used _only_ in case of default setting. Now, in default case, we always
> > use Free(Mono|Sans|Serif) See lyxStatus.pm:219 ff
>
> This should only affect documents in language directories that had rules for
> replacing "default" with special font settings, i.e.
>
> if ($lang =~ /^(ru|uk|sk|el)$/) {
> elsif ($lang =~ /^(he)$/) {
> elsif ($lang eq "fa") {
> # elsif ($lang eq "ko" ) {
> elsif ($lang eq "ar" ) {
>
> rigth?
Yes. As a backstop.
> Many of the failing tests below do not belong to document in
> (ru|uk|sk|el|he|fa|ko|ar).
OK, but this _may_ have been because of other reasons.
> I agree that it was no good idea to do this changes without knowing which
> *.systemF tests failed prior to it. So I reverted the changes to
> useSystemFonts.py in 124e3c92e2.
>
> I apologize for the caused trouble.
>
No problem. In fact, I am happy that someone looks in the perl scripts.
...
> >> IMO we should ignore the warnings. This may also fix
> >> .*EmbeddedObjects_pdf4_systemF tests.
>
> > Mark, that it would not fix it for TL18.
>
> What are the errors with TL18?
Same as often reported by Scott and me, see
https://www.mail-archive.com/lyx-devel@lists.lyx.org/msg207755.html
>
> >> > 1771 - export/doc/attic/eu_Customization_dvi3_systemF (Failed)
> >> > 1776 - export/doc/attic/eu_Customization_pdf4_systemF (Failed)
> >> > 1778 - export/doc/attic/eu_Customization_pdf5_systemF (Failed)
>
> >> IMO we can ignore these tests (attic with non-default output).
>
> > Not in my opinion.
>
> >> > 2470 - export/doc/fr/Customization_dvi3_systemF (Failed)
> >> > 2475 - export/doc/fr/Customization_pdf4_systemF (Failed)
> >> > 2477 - export/doc/fr/Customization_pdf5_systemF (Failed)
>
> >> Missing character: There is no ― in font
> >> [lmroman10-regular]:mapping=tex-text;!
>
> >> Fixed by replacing Character '―' (8213, 0x2015) 2015 HORIZONTAL BAR
> >> with emdash (which is the correct character at the places of occurence and
> >> present in LatinModern).
>
> > OK
>
> >> > 5197 - export/examples/es/linguistics_dvi3_systemF (Failed)
> >> > 5202 - export/examples/es/linguistics_pdf4_systemF (Failed)
> >> > 5204 - export/examples/es/linguistics_pdf5_systemF (Failed)
>
> >> Fails here (like any linguistics.lyx export), because I don't have
> >> "covington.sty".
>
> > Many missing glyphs.
> > Missing character: There is no ʔ (U+0xxx) in font
> > [lmroman10-regular]:+tlig;!
> > where xxx is one of
> > 251
> > 261,26a
> > 283,28a
> > 294
> > 2c8
> > 2d0
> > 325,33c
>
> Is this problem only in Spanish or in all linguistics.*systemF tests?
>
Only in Spanish.
> >> > 6320 - export/examples/sr/Braille_dvi3_systemF (Failed)
> >> > 6325 - export/examples/sr/Braille_pdf4_systemF (Failed)
>
> >> "Hand compiling" works fine. ctests fail for strange reasons. Checking
> >> verbose log shows:
>
> >> [...]
> >> hyph-sh-latn.pat.txt,hyph-sh-cyrl.pat.txt
> >> Language serbian was not yet loaded; created with id 4
> >> Language serbian already loaded; id is 4
> >> Language serbian already loaded; id is 4
> >> Language serbian already loaded; id is 4
> >> Language serbian already loaded; id is 4
> >> support/Systemcall.cpp (291): Systemcall: 'dvilualatex
> >> "Braille_dvi3_systemF.tex"' finished with exit code 1
> >> Language serbian already loaded; id is 4-- Msg Summary:
> >> -- Error: Exporting
> >> "/usr/local/src/lyxtest/autotests/out-home/AbC_pGGdpH/sr/Braille_dvi3_systemF.lyx"
> >> to format dvi3
> >> -- Exporting sr/Braille.lyx to dvi3
> >> CMake Error at /usr/local/src/lyx/development/autotests/export.cmake:265
> >> (message):
> >> Export failed
>
> >> Could the "already loaded" warning trigger failure?
>
> > Sure, we agreed that warnings should be treated as errors in ctests.
> > Exceptions to that are listed in file filterCheckWarnings.
>
> I see. Not a problem of changed fonts, then.
> Could you please add an exception to this warning?
I could, but I wonder, why the tests succeed here.
...
> > With the _old_ useSystemFonts.pl, the remaining are only
> > EmbeddedObjects_.*_systemF (8 cases)
> > (And only with TL18)
> > Please restore useSystemFonts.pl.
>
> Done in 124e3c92e26.
> Do the remaining test still fail?
Yes for EmbeddedObjects.
> Thanks,
>
> Günter
>
Kornel
signature.asc
Description: This is a digitally signed message part.
