FS#25000 - [fontforge] lilypond notation bug

Attached to Project: Community Packages
Opened by Frederik Vos (inktvis75) - Sunday, 03 July 2011, 14:45 GMT
Last edited by Sergej Pupykin (sergej) - Friday, 20 January 2012, 16:49 GMT
Task Type Bug Report
Category Packages
Status Closed
Assigned To Sergej Pupykin (sergej)
Gaetan Bisson (vesath)
Architecture All
Severity Medium
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 1
Private No

Details

Description:

If you compile a lilypond file, the half note is not correctly written (it's written like a fat quarter note)
The problem exist with the 2.14.1 version and with the version in AUR.
Using the binary file from the lilypond website, everything is ok.
This task depends upon

Closed by  Sergej Pupykin (sergej)
Friday, 20 January 2012, 16:49 GMT
Reason for closing:  Fixed
Comment by Sergej Pupykin (sergej) - Wednesday, 06 July 2011, 15:03 GMT
I have no ideas about lilypond.
Gaetan, could you please check - can it be fontforge (the same as https://bbs.archlinux.org/viewtopic.php?id=103644 ) issue?
Comment by Gaetan Bisson (vesath) - Wednesday, 06 July 2011, 16:47 GMT
Well, the spine patch was integrated into upstream fontforge a while ago, and our fontforge package has not been updated in months...
When did you start noticing this problem? Can you reproduce it with earlier versions of our fontforge and/or lilypond packages?
Comment by Sergej Pupykin (sergej) - Thursday, 07 July 2011, 10:59 GMT
Frederik,

what if you build previous version of lilypond on fresh system?

http://projects.archlinux.org/svntogit/community.git/tree/lilypond/trunk?id=42bce223d693c58ab1162f30bf299d818970b452

contains 2.12.3-7 pkgbuild
Comment by Andrew Geng (pteromys) - Thursday, 07 July 2011, 18:26 GMT
The same symptom has been reported upstream as http://code.google.com/p/lilypond/issues/detail?id=1717 , where they seem to suspect problems with ghostscript 9.
Comment by Frederik Vos (inktvis75) - Thursday, 07 July 2011, 19:21 GMT
I'll do some tests this weekend, but the problem is recent: i wrote scores on 9 may, without a problem. Ghostscript is updated after that date, so maybe Andrew's suggestion is right. Also the screenshot from that issue 1717 is the same problem. I can confirm that the version from Aur (git) has the same problem.
Comment by Frederik Vos (inktvis75) - Thursday, 07 July 2011, 21:23 GMT
I update the git version, and this is what happened:

Layout output to `t10.ps'...
Converting to `./t10.pdf'...
`(gs -q -dSAFER -dDEVICEWIDTHPOINTS=419.53 -dDEVICEHEIGHTPOINTS=595.28 -dCompatibilityLevel=1.4 -dNOPAUSE -dBATCH -r1200 -sDEVICE=pdfwrite -sOutputFile=./t10.pdf -c.setpdfwrite -ft10.ps)' failed (256)

but:
lilypond --format=png t10.ly
gives a correct png file and a correct ps file
Comment by Frederik Vos (inktvis75) - Thursday, 07 July 2011, 21:24 GMT
tried it manually:
s -dSAFER -dDEVICEWIDTHPOINTS=419.53 -dDEVICEHEIGHTPOINTS=595.28 -dCompatibilityLevel=1.4 -dNOPAUSE -dBATCH -r1200 -sDEVICE=pdfwrite -sOutputFile=./t10.pdf -c.setpdfwrite -ft10.ps
GPL Ghostscript 9.02 (2011-03-30)
Copyright (C) 2010 Artifex Software, Inc. All rights reserved.
This software comes with NO WARRANTY: see the file PUBLIC for details.
Error: /ioerror in --showpage--
Operand stack:
1 true
Execution stack:
%interp_exit .runexec2 --nostringval-- --nostringval-- --nostringval-- 2 %stopped_push --nostringval-- --nostringval-- --nostringval-- false 1 %stopped_push 1894 1 3 %oparray_pop 1893 1 3 %oparray_pop 1877 1 3 %oparray_pop 1771 1 3 %oparray_pop --nostringval-- %errorexec_pop .runexec2 --nostringval-- --nostringval-- --nostringval-- 2 %stopped_push --nostringval-- 1777 0 3 %oparray_pop --nostringval-- --nostringval--
Dictionary stack:
--dict:1157/1684(ro)(G)-- --dict:0/20(G)-- --dict:120/200(L)--
Current allocation mode is local
Last OS error: 28
Current file position is 236615
GPL Ghostscript 9.02: Unrecoverable error, exit code 1
Comment by Frederik Vos (inktvis75) - Thursday, 07 July 2011, 21:34 GMT
ok, this is weird: by mistake, i runned lilypond as root, now it's ok
revert back to lilypond from extra repo, runned as root, it's also ok!
as normal user, i'll get the same error as the git version

btw: i didnt had that error before, just the wrong notation
Comment by Thomas Schuetz (Murray_B) - Sunday, 10 July 2011, 13:00 GMT
I got into lilypond-trouble, too. There are some more things that are wrong. The #-sign hasn't solid lines any longer, the pause sign have wrong positions and the notes below or above the normal system have no horizontal line through them. Ah, and there are no vertical bars between the notes. Short: The notation is completely broken! And there are no problems with the binary from the lilypond-website! I attached a little example: the line on the top are done with the binary from lilypond.org, the line at the bottome is done with the community-package.
Comment by Frederik Vos (inktvis75) - Sunday, 10 July 2011, 15:03 GMT
did a clean install, the same problem as first reported. An older version of fontforge didn't fix the problem, also an older version of ghostscript didn't fix it.

i am not able to fix the older version:
Traceback (most recent call last):
File "out/midi2ly", line 50, in <module>
import midi
ImportError: /home/frederik/AUR/lilypond/src/lilypond-2.12.3/python/out/midi.so: undefined symbol: Py_InitModule
help2man: can't get `--help' info from out/midi2ly
make[1]: *** [out/midi2ly.1] Error 1
make[1]: Leaving directory `/home/frederik/AUR/lilypond/src/lilypond-2.12.3/scripts'
make: *** [all] Error 2
==> ERROR: A failure occurred in build().
Aborting...
Comment by Frederik Vos (inktvis75) - Sunday, 10 July 2011, 15:40 GMT
Ok, a temporary solution:
1) download binary from lilypond website
2) sh lilypond-2.14.1-1.linux-64.sh --tarball
3) tar xjvf lilypond-2.14.1-1.linux-64.tar.bz2
4) cp -R usr/share/lilypond/current/fonts/{otf,svg} /usr/share/lilypond/2.14.1/fonts/

So something is wrong in the fonts in the lilypond package, and the same problem is there if you compile it from source (using the devel or git version from AUR)
Comment by Frederik Vos (inktvis75) - Sunday, 10 July 2011, 16:34 GMT
ps: the reason that it worked temporary with the root account: a mix between the website version and the arch versions
Comment by Sergej Pupykin (sergej) - Monday, 11 July 2011, 09:42 GMT
fontforge reports errors during build like this:

Invoking "fontforge -script parmesan18.pe parmesan18.pfb"...
Copyright (c) 2000-2011 by George Williams.
Executable based on sources from 13:48 GMT 22-Feb-2011-ML.
Library based on sources from 13:48 GMT 22-Feb-2011.
Internal Error (overlap) in noteheads.sM2neomensural: Mismatched intersection.
(522.624,-427.312)->(500,-179.022) ends at (-999999,-999999) while (500,-179.022)->(462.232,-180.56) starts at (500,-179.022)
Internal Error (overlap) in noteheads.sM2neomensural: Winding number did not return to 0 when x=499.999
Internal Error (overlap) in noteheads.sM2neomensural: Winding number did not return to 0 when x=499.999
Internal Error (overlap) in noteheads.sM2neomensural: monotonic is both needed and unneeded (500,-179.022)->(462.232,-180.56). x=499.999 (prev=469.804)
Internal Error (overlap) in noteheads.sM2neomensural: Winding number did not return to 0 when x=499.999
Internal Error (overlap) in noteheads.sM2neomensural: Humph. This monotonic leads nowhere (499.998,-179.023)->(499.998,-179.022).
Internal Error (overlap) in noteheads.sM2neomensural: Expected needed monotonic @(500,-179.022) (500,-179.022)->(462.232,-180.56).
Internal Error (overlap) in noteheads.sM2neomensural: Closing contour with unneeded path
Internal Error (overlap) in noteheads.sM2neomensural: Expected needed monotonic @(462.232,-180.56) (481.093,-197.93)->(500,-179.022).
Comment by Sergej Pupykin (sergej) - Monday, 11 July 2011, 09:59 GMT
Thanks, Frederik, I upload workaround (lilypond-2.14.1-2 with fonts replaced from upstream binary build). It should work.
Comment by Frederik Vos (inktvis75) - Monday, 11 July 2011, 16:29 GMT
Thanks Sergej, it works like a charm now !
Comment by Thomas Schuetz (Murray_B) - Wednesday, 10 August 2011, 18:53 GMT
Okay, new version of lilypond, old bugs again. New bugs, too. I am using lilypond 2.14.2-1 and there are no horizontal bars again. It's a bit difficult to describe all the bugs, because lilypond often (most of the time...) crashes with "Vorverarbeitung der grafischen Elemente...throw from within critical section.". Still everything works alright with the binary from lilypond.org
Comment by Sergej Pupykin (sergej) - Wednesday, 14 December 2011, 09:13 GMT
please try lilypond-2.14.2-2
Comment by Sergej Pupykin (sergej) - Wednesday, 14 December 2011, 09:39 GMT
Gaetan, I rebuilt lilypond-2.14.2-2 with fontforge-git-20111214 from aur. Lilypond looks fixed now. Please consider upgrade fontforge from git.
Comment by Gaetan Bisson (vesath) - Wednesday, 14 December 2011, 11:02 GMT
Sergej: I have put fontforge-20111214-1 in [testing]; could you confirm that lilypond has no issue with it?
Before moving it to [extra], I will wait a few days for fontforge developers to reply to my query about whether a new release is coming soon.
Comment by Sergej Pupykin (sergej) - Wednesday, 14 December 2011, 11:45 GMT
I put community-testing/lilypond-2.14.2-3 - it looks good.

Frederik, could you please try it too?
Comment by Thomas Schuetz (Murray_B) - Tuesday, 20 December 2011, 19:33 GMT
I just tested lilypond-2.14.2-3 and there are still no vertical bars.
Comment by Sergej Pupykin (sergej) - Tuesday, 20 December 2011, 20:20 GMT
it is my test:

\version "2.14.2"

%the following gives no barlines in pdf output
\relative c'' {
c1 | c1 | c1 |
}

is it ok?
Comment by Sergej Pupykin (sergej) - Monday, 16 January 2012, 21:09 GMT
I hope this bug fixed too by last update?
Comment by Thomas Schuetz (Murray_B) - Friday, 20 January 2012, 16:47 GMT
Okay, I tried a few songs and everything looks alright here. Thank you!

Loading...