On Wed, Apr 29, 2015 at 03:06:48PM +0200, Kornel Benko wrote: > Am Mittwoch, 29. April 2015 um 14:57:29, schrieb Enrico Forestieri > <for...@lyx.org> > > On Wed, Apr 29, 2015 at 02:30:47PM +0200, Kornel Benko wrote: > > > Am Mittwoch, 29. April 2015 um 13:56:38, schrieb Enrico Forestieri > > > <for...@lyx.org> > > > > On Wed, Apr 29, 2015 at 01:34:33PM +0200, Kornel Benko wrote: > > > > > Am Mittwoch, 29. April 2015 um 13:27:12, schrieb Enrico Forestieri > > > > > <for...@lyx.org> > > > > > > On Wed, Apr 29, 2015 at 01:22:14PM +0200, Kornel Benko wrote: > > > > > > > Am Mittwoch, 29. April 2015 um 12:33:00, schrieb Enrico > > > > > > > Forestieri <for...@lyx.org> > > > > > > > > lib/scripts/lyxpreview2bitmap.py > > > > > > > > b/lib/scripts/lyxpreview2bitmap.py > > > > > > > > > > > > > > What was wrong with my patch? > > > > > > > > > > > > > > I posted it on the bug tracker > > > > > > > > > > > > Sincerely, I don't know what you are talking about. > > > > > > > > > > > > > > > > I mean Ticket #6369. Sorry. > > > > > > > > Ah, that one... I had forgotten it. As I already said, if you discuss > > > > these things on the tracker, they risk to go unnoticed and the > > > > probability of stepping on each other toes becomes high. > > > > > > > > That said, the patch I committed is more general. > > > > > > > > > > OK, > > > your newest commit produces now plenty of warnings > > > > > > Warning: pdftocairo failed on page 1, file lyxpreviewP29083_legacy.001 > > > ... > > > Warning: pdftocairo failed on page 31, file lyxpreviewP29083_legacy.031 > > > Warning: Now trying to obtain failed previews through pdflatex > > > > > > if testes with doc/Math.lyx > > > > Yes, that's normal. It first tries to convert the postscript files > > generated by dvips. If that fails (for the math manual the responsible > > for the failure is hyperref), it goes through the pdflatex route. > > You can avoid those warnings by setting the default output format for > > the document to some pdf type. Then, dvips is not even attempted. > > > > Yes, OK, but that is our official file. And it takes now long to process the > previews. > Why is pdftocairo trying to interpret PS files? According to 'pdftocairo > --help' it expects PDF file > as input.
And it gets a PDF file using epstopdf. The reason it is about 2 times slower is that it first tries the postscript route. However, the real culprit for the slowness is hyperref. Take it away and the conversion will be lightning fast. -- Enrico