bug-lilypond
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: PDF docs for 2.19.82 broken/missing fonts


From: Knut Petersen
Subject: Re: PDF docs for 2.19.82 broken/missing fonts
Date: Tue, 26 Jun 2018 15:02:34 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.8.0

Am 26.06.2018 um 13:50 schrieb Masamichi Hosoda:
Masamichi-san?  Any idea?

Knut Petersen <address@hidden> writes:

This problems seems to be related to https://codereview.appspot.com/314130043/

The file http://lilypond.org/doc/v2.19/Documentation/notation.pdf seems to be  
the intermediate version of notation.pdf.
pdfinfo shows:

    Creator:         XeTeX output 2018.06.24:1343
    Producer:       xdvipdfmx (0.7.9)
    CreationDate:   Sun Jun 24 14:43:19 2018 CEST

'Producer: xdvipdfmx' is unxpected and definitely broken.

The file http://lilypond.org/doc/v2.19/Documentation/notation.de.pdf is the 
expected final version of notation.de.pdf
  pdfinfo shows:

    Creator:         XeTeX output 2018.06.24:1347
    Producer:       GPL Ghostscript 9.21
    CreationDate:   Sun Jun 24 14:47:40 2018 CEST
    ModDate:        Sun Jun 24 14:47:40 2018 CEST

The intermediate pdfs should have been processed with extractpdfmark
and ghostscript according to this snippet from texinfo-rules.make:

    $(outdir)/%.pdf: $(outdir)/%.texi $(outdir)/version.itexi 
$(outdir)/%.pdf.omf $(outdir)/weblinks.itexi | $(OUT_TEXINFO_MANUALS)
    ifeq ($(WEB_VERSION),yes)
         PDFTEX=$(PDFTEX) PDFLATEX=$(PDFLATEX) $(buildscript-dir)/run-and-check "cd $(outdir); 
texi2pdf $(TEXI2PDF_FLAGS) -D web_version -I $(abs-src-dir) $(TEXINFO_PAPERSIZE_OPTION) $(<F) < 
/dev/null" "$*.texi2pdf.log"
    else
         PDFTEX=$(PDFTEX) PDFLATEX=$(PDFLATEX) $(buildscript-dir)/run-and-check "cd $(outdir); 
texi2pdf $(TEXI2PDF_FLAGS) -I $(abs-src-dir) $(TEXINFO_PAPERSIZE_OPTION) $(<F) < /dev/null" 
"$*.texi2pdf.log"
    endif
    ifeq ($(USE_EXTRACTPDFMARK),yes)
         $(EXTRACTPDFMARK) -o $(outdir)/$*.pdfmark $@
         $(GS920) -dBATCH -dNOPAUSE -sDEVICE=pdfwrite -dAutoRotatePages=/None 
-sOutputFile=$(outdir)/$*.final.pdf -c "30000000 setvmthreshold" -f 
$(top-build-dir)/out-fonts/*.font.ps $(outdir)/$*.pdfmark $@
         rm $@
         mv $(outdir)/$*.final.pdf $@
    endif

Apparently that worked for notation.de.pdf but not for notation.pdf. Why? At 
the moment I don't see the reason, maybe it's too late.

@Masamichi: Any idea ?
Any non-pattern rule overriding?  Stuff getting copied before being
finished because of bad parallelism?  $@ stripping .de. stuff and
similar so that everything is going through the same temporary file (not
likely I guess) At any rate, it doesn't seem to be related to your
recent patches.
If I understand correctly,

GUB generated correct PDFs for 2.19.81 in Jan. 2018.
extractpdfmark was used for generating PDFs of all languages including English.

http://lilypond.org/downloads/binaries/documentation/lilypond-2.19.81-1.documentation.tar.bz2

```
$ pdfinfo notation.pdf
Creator:         XeTeX output 2018.01.28:1325
Producer:       GPL Ghostscript 9.21
CreationDate:   Sun Jan 28 22:25:13 2018 JST
ModDate:        Sun Jan 28 22:25:13 2018 JST
...

$ pdfinfo notation.de.pdf
Creator:         XeTeX output 2018.01.28:1333
Producer:       GPL Ghostscript 9.21
CreationDate:   Sun Jan 28 22:33:44 2018 JST
ModDate:        Sun Jan 28 22:33:44 2018 JST
...
```

Yes.
However, GUB generated broken PDFs for 2.19.82 in Jun. 2018.
extractpdfmark was used for generating non-English PDFs.
For English PDFs, extractpdfmark was not used.

http://lilypond.org/downloads/binaries/documentation/lilypond-2.19.82-1.documentation.tar.bz2

```
$ pdfinfo notation.pdf
Creator:         XeTeX output 2018.06.24:1343
Producer:       xdvipdfmx (0.7.9)
CreationDate:   Sun Jun 24 21:43:19 2018 JST
...

$ pdfinfo notation.de.pdf
Creator:         XeTeX output 2018.06.24:1347
Producer:       GPL Ghostscript 9.21
CreationDate:   Sun Jun 24 21:47:40 2018 JST
ModDate:        Sun Jun 24 21:47:40 2018 JST
...
```

Yes. notation.pdf looks as if extractpdfmark/gs was not used. But the part of 
the build log Phil published clearly proves that extractpdfmark and ghostscript 
were used.

@Phil: If you have a look at the notation.pdf in the build tree: is that file 
ok or  is the same broken version as published on lilypond.org?

GUB has not been changed between Jan. and Jun.
Its last change is Aug. 2017.
https://github.com/gperciva/gub/commits/master

Yes.

Knut



reply via email to

[Prev in Thread] Current Thread [Next in Thread]