Closed strogdon closed 6 months ago
That's not ones I have seen. For the first one, it would look like a missing language pack. The second one is of interest because I believe that means you are on the vbraun branch. That missing file should have made its appearance in a merge a few hours ago that I have not tested yet.
I pulled from master
/var/db/repos/sage-on-gentoo # git status
On branch master
Your branch is up to date with 'origin/master'.
nothing to commit, working tree clean
I will check language packs. I was missing a few. I had to emerge texlive directly.
I have changed how I deal with the vbraun branch slightly. If you have installed the files from the vbraun.settings
folder, you are still pulling the vbraun branch even if the overlay has not been switched over. The overlay switch over means that you may have different patches, because it has just been merged for 10.3.beta4 the patches are identical.
If you have build logs, it should tell you which git repo you are pulling sage from at the beginning of the log.
So yes I had pulled from the vbraun
branch. However, I created no vbraun
symlinks. Is this working correctly? Must I remove the vbraun.settings
folder from the overlay?
I removed the vbraun.settings
folder and I'm still pulling from https://github.com/vbraun/sage.git
when building sage-doc
. In the sage-doc
ebuild there is EGIT_REPO_URI="https://github.com/vbraun/sage.git"
. Is this correct?
No it is not. It looks like a mistake made it to the sage-doc
ebuild on the master branch. The only ebuild that currently does not use the sage-git eclass. I should have paid more attention. I will review the vbraun settings and the ebuild shortly.
OK, fixed the repo source in the ebuild, it may take up to 15 mn for the gentoo mirror to synchronise.
I got through the html-docs but I have a SANDBOX VIOLATION
when building the pdf-docs which seems to originate with
LaTePackage fancyhdr Warning: \headheight is too small (12.0pt):
X Warn(fancyhdr) Make it at least 15.0pt, for example:
ing: H(fancyhdr) \setlength{\headheight}{15.0pt}.
yper ref(fancyhdr) You might also make \topmargin smaller to compensate:
e
rence (fancyhdr) \addtolength{\topmargin}{-3.0pt}.
`
functional_programming:functional-programming'
on page 346 undefined on input line 26844.
][73] * ACCESS DENIED: mkdir: /var/cache/fonts/pk/modeless/public
mkdir: cannot create directory ‘././var/cache/fonts/pk/modeless/public’: Permission denied
mktexpk: /usr/share/texmf-dist/web2c/mktexdir /var/cache/fonts/pk/modeless/public/tex-gyre failed.
Package fancyhdr Warning: \headheight is too small (12.0pt):
(fancyhdr) Make it at least 15.0pt, for example:
(fancyhdr) \setlength{\headheight}{15.0pt}.
(fancyhdr) You might also make \topmargin smaller to compensate:
(fancyhdr) \addtolength{\topmargin}{-3.0pt}.
[74] [163] [76kpathsea: Appending font creation commands to missfont.log.
dvipdfmx:warning: Could not locate a virtual/physical font for TFM "ec-qhvb".
dvipdfmx:warning: >> There are no valid font mapping entry for this font.
dvipdfmx:warning: >> Font file name "ec-qhvb" was assumed but failed to locate that font.
dvipdfmx:fatal: Cannot proceed without .vf or "physical" font for PDF output...
No output PDF file written.
]
Глава 6.
Latexmk: Errors, so I did not complete making targets
Collected error summary (may duplicate other messages):
dvipdf: Command for 'dvipdf' gave return code 256
Latexmk: Use the -f option to force complete processing,
unless error was exceeding maximum runs, or warnings treated as errors.
make[2]: *** [Makefile:30: a_tour_of_sage.pdf] Error 12
make[2]: Leaving directory '/var/tmp/portage/sci-mathematics/sage-doc-9999/work/build_doc/latex/ja/a_tour_of_sage'
Error building the documentation.
Have you had any issues? The folder /var/cache/fonts/pk/modeless/ exists but the public folder cannot be created. I am the owner of the modeless folder, not root.
This may not be the issue but dev-libs/kpathsea
has not been updated to the tex-overlay
version.
That actually could be it. I did not test any other language than English and French. Did you install freefont
(not to be mistaken with freefonts
)?
I should note that I had to already add instructions in the ebuild for a sandbox violation where luatex check whether it can write on the system as part of a test looking for a location it can write to, and it is caught by the sandbox. So, I had to tell the sandbox to let it go, it fails anyway.
It is fine if we end up having to add some more of those.
I do have freefont
installed, kpathsea
and rebuilds are still building.
The doc build has gone back to be mostly single threaded I think. Parallelism has to be controlled with SAGE_NUM_THREADS
and SAGE_NUM_THREADS_PARALLEL
. I will try to introduce them in the environment in this cycle to see what happens.
After upgrading kpathsea
I still have a sandbox violation
:
LaTeX Warning: Citation `bibliography:pyt' on page 88 undefined on input line 6
366.
[6][88] * ACCESS DENIED: mkdir: /var/cache/fonts/pk/modeless/public
* ACCESS DENIED: mkdir: /var/cache/fonts/pk/modeless/public
mkdir: cannot create directory ‘././var/cache/fonts/pk/modeless/public’: Permission denied
mktexpk: /usr/share/texmf-dist/web2c/mktexdir /var/cache/fonts/pk/modeless/public/tex-gyre failed.
kpathsea: Appending font creation commands to missfont.log.
dvipdfmx:warning: Could not locate a virtual/physical font for TFM "ec-qhvb".
dvipdfmx:warning: >> There are no valid font mapping entry for this font.
dvipdfmx:warning: >> Font file name "ec-qhvb" was assumed but failed to locate that font.
dvipdfmx:fatal: Cannot proceed without .vf or "physical" font for PDF output...
The ec-qhvb
tfm font seems to be associated with the Japanese documentation. However there are situations where the font is generated correctly. As in
(/usr/share/texmf-dist/[9tex]/generic/iftex/ifvtex.sty (./a_tour_of_sage.aux
kpathsea: Running mktexpk --mfmode / --bdpi 600 --mag 2+293/600 --dpi 1493 ec-qhvb
)
OK the sandbox violation seems to have been me removing VARTEXFONTS
- unfortunately still needed when you deal with more than English and French. There is another failure lurking that I have not identified yet.
Document Class: sphinxmanual 2019/12/01 v2.3.0 Document class (Sphinx manual)
</usr/share/texmf-dist/fonts/type
1/public/amsfonts/cm/cmr10.pfb></usr/share/texmf-dist/fonts/type1/public/amsfon
ts/cm/cmsy10.pfb></usr/share/texmf-dist/fonts/type1/public/amsfonts/cm/cmsy5.pf
b(/usr/share/texmf-dist/tex/latex/base/report.c>ls
Document Class: report 2023/05/17 v1.4n Standard LaTeX document class
Output written on a_tour_of_sage.pdf (13 pages, 129696 bytes).
(/usr/share/texmf-dist/tex/latex/base/size10.clomktexpk: `gsftopk upgbm-h 1436' failed to make upgbm-h.1436pk.
Package fancyhdr Warning: \headheight is too small (12.0pt):
(fancyhdr) Make it at least 15.0pt, for example:
(fancyhdr) \setlength{\headheight}{15.0pt}.
(fancyhdr) You might also make \topmargin smaller to compensate:
(fancyhdr) \addtolength{\topmargin}{-3.0pt}.
[39]kpathsea: Appending font creation commands to missfont.log.
dvipdfmx:warning: Could not locate a virtual/physical font for TFM "upgbm-h".
dvipdfmx:warning: >> There are no valid font mapping entry for this font.
dvipdfmx:warning: >> Font file name "upgbm-h" was assumed but failed to locate that font.
dvipdfmx:fatal: Cannot proceed without .vf or "physical" font for PDF output...
No output PDF file written.
Transcript written on a_tour_of_sagLatexmk: Errors, so I did not complete making targets
Collected error summary (may duplicate other messages):
dvipdf: Command for 'dvipdf' gave return code 256
Latexmk: If appropriate, the -f option can be used to get latexmk
to try to force complete processing.
e.log.
make[2]: *** [Makefile:30: a_tour_of_sage.pdf] Error 12
make[2]: Leaving directory '/home/portage/sci-mathematics/sage-doc-9999/work/build_doc/latex/ja/a_tour_of_sage'
probably mixed with other stuff from parallel processing.
$ mktexpk --mfmode / --bdpi 600 --mag 2+236/600 --dpi 1436 upgbm-h
mktexpk: Running gsftopk upgbm-h 1436
gsftopk(k) version 1.19.2
render.ps: font HaranoAjiGothic-Medium-UniJIS2004-UTF16-H is not defined.
GPL Ghostscript 10.02.1: Unrecoverable error, exit code 1
/10021
gs terminated abnormally with status 1
grep: upgbm-h.log: No such file or directory
mktexpk: `gsftopk upgbm-h 1436' failed to make upgbm-h.1436pk.
feels like a bug
$ cat /usr/share/texmf-dist/fonts/map/dvipdfmx/ptex-fontmaps/haranoaji/uptex-haranoaji-04.map
urml UniJIS2004-UTF16-H HaranoAjiMincho-Regular.otf
urmlv UniJIS2004-UTF16-V HaranoAjiMincho-Regular.otf
ugbm UniJIS2004-UTF16-H HaranoAjiGothic-Medium.otf
ugbmv UniJIS2004-UTF16-V HaranoAjiGothic-Medium.otf
uprml-h UniJIS2004-UTF16-H HaranoAjiMincho-Regular.otf
uprml-v UniJIS2004-UTF16-V HaranoAjiMincho-Regular.otf
upgbm-h UniJIS2004-UTF16-H HaranoAjiGothic-Medium.otf
upgbm-v UniJIS2004-UTF16-V HaranoAjiGothic-Medium.otf
uprml-hq UniJIS-UCS2-H HaranoAjiMincho-Regular.otf
upgbm-hq UniJIS-UCS2-H HaranoAjiGothic-Medium.otf
and upgbm-h.tfm
is on the system, and the appropriate .map
file is present.
And the issue seems to only be in Japanese. All other languages work.
I tried vanilla with -j1
and got the failure
grep: upgbm-h.log: No such file or directory
mktexpk: `gsftopk upgbm-h 1436' failed to make upgbm-h.1436pk.
kpathsea: Appending font creation commands to missfont.log.
dvipdfmx:warning: Could not locate a virtual/physical font for TFM "upgbm-h".
dvipdfmx:warning: >> There are no valid font mapping entry for this font.
dvipdfmx:warning: >> Font file name "upgbm-h" was assumed but failed to locate that font.
dvipdfmx:fatal: Cannot proceed without .vf or "physical" font for PDF output...
No output PDF file written.
Collected error summary (may duplicate other messages):
dvipdf: Command for 'dvipdf' gave return code 256
Latexmk: Use the -f option to force complete processing,
unless error was exceeding maximum runs, or warnings treated as errors.
Latexmk: Errors, so I did not complete making targets
make[5]: *** [Makefile:30: tutorial-jp.pdf] Error 12
Error building the documentation.
Traceback (most recent call last):
File "<frozen runpy>", line 198, in _run_module_as_main
File "<frozen runpy>", line 88, in _run_code
File "/local/sage-git/sage/src/sage_docbuild/__main__.py", line 508, in <module>
sys.exit(main())
^^^^^^
File "/local/sage-git/sage/src/sage_docbuild/__main__.py", line 504, in main
builder()
File "/local/sage-git/sage/src/sage_docbuild/builders.py", line 299, in pdf
raise RuntimeError(error_message % (command, tex_dir))
RuntimeError: failed to run $MAKE all-pdf in /local/sage-git/sage/local/share/doc/sage/latex/ja/tutorial
Not sure if something is missing in texlive
or if there is a coding problem.
That's exactly the same issue that I have. I suspect something subtle in the texlive install for japanese.
One thing is certainly incorrect
$ ls -al /usr/share/texmf-dist/dvipdfmx/
total 2
drwxr-xr-x 2 root root 1024 Dec 27 16:38 .
drwxr-xr-x 30 root root 1024 Dec 27 16:38 ..
lrwxrwxrwx 1 root root 46 Dec 27 16:38 dvipdfmx.cfg -> '../../../../etc/texmf/dvipdfmx).d/dvipdfmx.cfg
The symlink should point to ../../../../etc/texmf/dvipdfmx.d/dvipdfmx.cfg
.
This is interesting indeed!
Vanilla builds with the appropriate fix.
I think it should go in bugzilla. I'll see if I can figure which package dvipdfmx.cfg belongs to. I think it is texlive-core but I am not 100% per cent certain.
texlive-core but I think the error may be a corner case in one of the eclass.
This is done by the texlive-common.eclass
^[[32m*^[[0m Moving (and symlinking) /usr/share/texmf-dist/web2c/updmap-hdr.cfg to /etc/texmf/web2c.d
^[[32m*^[[0m Moving (and symlinking) /usr/share/texmf-dist/web2c/fmtutil-hdr.cnf to /etc/texmf/web2c.d
^[[32m*^[[0m Moving (and symlinking) /usr/share/texmf-dist/dvipdfmx/dvipdfmx.cfg to /etc/texmf/dvipdfmx.d
>>> Completed installing app-text/texlive-core-2023 into /home/portage/app-text/texlive-core-2023/image
and they are all affected
fbissey@tarazed ~ $ ll /usr/share/texmf-dist/web2c/updmap-hdr.cfg
lrwxrwxrwx 1 root root 45 Dec 28 14:05 /usr/share/texmf-dist/web2c/updmap-hdr.cfg -> '../../../../etc/texmf/web2c).d/updmap-hdr.cfg'
fbissey@tarazed ~ $ ll /usr/share/texmf-dist/web2c/fmtutil-hdr.cnf
lrwxrwxrwx 1 root root 46 Dec 28 14:05 /usr/share/texmf-dist/web2c/fmtutil-hdr.cnf -> '../../../../etc/texmf/web2c).d/fmtutil-hdr.cnf'
fbissey@tarazed ~ $ ll /usr/share/texmf-dist/dvipdfmx/dvipdfmx.cfg
lrwxrwxrwx 1 root root 46 Dec 28 14:05 /usr/share/texmf-dist/dvipdfmx/dvipdfmx.cfg -> '../../../../etc/texmf/dvipdfmx).d/dvipdfmx.cfg'
The links are fine for my texlive-2019
. So the eclass has changed since it was installed (12/20/2023).
Found the error. texlive-common_handle_config_files
function in texlive-common.eclass
Gentoo
dosym8 -r /etc/texmf/$(dirname ${f}).d/$(basename ${f}) ${TEXMF_PATH}/${f}
Texlive-overlay
dosym -r "/etc/texmf/${rel_dir}).d/$(basename "${f}")" "${texmf_path}/${f}"
Note the stray )
Added a comment to the texlive 2023 bug https://bugs.gentoo.org/836779
Fixing the eclass and re-installing texlive-core-2023 fixed the links. Next step, all languages sage-doc build.
I was able to build the docs for all langs.
How long does it take on your system? I am doing around 30mn for English + French and I estimate I will do 35 mn for all languages (based on 34mn for everything but Japanese). This is up from 20mn for English + French before the switch to luatex. So, it is slower.
Success in 33mn and 54s. Keeping the issue open until we have a fix in the texlive overlay.
How time this? I've not done any timings.
genlop -t sage-doc
from the gentoolkit I think.
Building all langs:
merge time: 53 minutes and 6 seconds.
and
Tue Dec 19 16:59:03 2023 >>> sci-mathematics/sage-doc-9999
merge time: 31 minutes.
So, yes the build seems to take longer.
https://gitweb.gentoo.org/proj/tex-overlay.git/commit/eclass/texlive-common.eclass?h=main&id=02a45ff97b50eb0ecffa7532bf1697b065b71925 I will officially close this issue
First, I have not been able to build the above with vanilla. The stopper appears to be
which suggest something is amiss with texlive. However, when building the docs for s-o-g I'm unable to get past the html-docs with failure
I'm unable to find
arith_sage
.