September 2007 - dev-luatex - NTG Mailing Lists

Cron <luatex@server> /home/luatex/luatex-svn
by root＠aanhet.net 26 Sep '07

26 Sep '07

U luatex/trunk/src/texk/web2c/luatexdir/utils.c U luatex/trunk/src/texk/web2c/luatexdir/avlstuff.c U luatex/trunk/src/texk/web2c/luatexdir/lua/llanglib.c U luatex/trunk/src/texk/web2c/luatexdir/lua/luastuff.c U luatex/trunk/src/texk/web2c/luatexdir/ptexlib.h A luatex/trunk/src/texk/web2c/luatexdir/svnversion.h U luatex/trunk/src/texk/web2c/lib/texmfmp.c Checked out revision 758.

1 0

Collected notes
by Jonathan Sauer 25 Sep '07

25 Sep '07

Hello, some notes I collected while testing LuaTeX beta 0.11.2: - It should be noted in the manual that \luaescapestring completely expands its parameter. - When setting a \toks register in Lua using tex.toks['foo'] = 'bar' et.al., what catcodetable is used? From looking inside luatex.web, I would guess that, since str_toks is used, all characters get catcode 12, except for spaces, which get catcode 10. A quick test confirms this. In case this is neither a bug nor a temporary hack: This IMO severely limits the usefulness of the tex.toks table, since the toks registers written in Lua later have to be re-scanned in TeX. IMO, tex.settoks should be passed an optional parameter to set the catcode table to use. When using array addressing, the current catcode table should be used (just like tex.print/tex.sprint). - What exactly is meant by "Lua strings are converted to token lists using \the\toks style expansion" (pg. 23)? Jonathan

3 4

Cron <luatex@server> /home/luatex/luatex-svn
by root＠aanhet.net 25 Sep '07

25 Sep '07

U luatex/trunk/src/texk/web2c/luatexdir/lua/llanglib.c U luatex/trunk/src/texk/web2c/luatexdir/lua/hyphen.c U luatex/trunk/manual/luatexref-t.tex Checked out revision 748.

1 0

Cron <luatex@server> /home/luatex/luatex-svn
by root＠aanhet.net 25 Sep '07

25 Sep '07

U luatex/trunk/manual/luatexref-t.tex Checked out revision 747.

1 0

Cron <luatex@server> /home/luatex/luatex-svn
by root＠aanhet.net 25 Sep '07

25 Sep '07

U luatex/trunk/src/texk/web2c/luatexdir/lua/llualib.c Checked out revision 746.

1 0

Lua states and \dump
by Jonathan Sauer 25 Sep '07

25 Sep '07

Hello, (yes, it's me again ;-) Yesterday I noticed that when \dump-ing a format file, the Lua states are not dumped as well, discarding all initializations made while creating the format. This is not a problem, since the initializations can be put into \everyjob, and of course from a technical angle it would be really difficult to dump a Lua state to the format file (requiring access to the internals of the Lua VM which can change at any time), but IMO this should be noted in the manual. Jonathan

3 10

tex.print and \inputlineno
by Jonathan Sauer 25 Sep '07

25 Sep '07

Hello, while testing, I stumbled on a (then) strange phenomenon concerning line numbers in tex.print. It seems that \inputlineno is not equal to the number of the line of the tex.print statement (or the macro expanding to it), but of the virtual file tex.print uses internally. Example (PlainTeX): -------------------------------------------------------------------- % Output is `1' instead of `3': \directlua0{\unexpanded{tex.print('\\message{\\the\\inputlineno}')}} % Output is `1' instead of `6': \scantokens{\message{\the\inputlineno}} \bye -------------------------------------------------------------------- Results in: This is luaTeX, Version 3.141592-beta-0.11.2-2007091918 (Web2C 7.5.6) (format=luatex 2007.9.20) 24 SEP 2007 15:25 **LineNumberTest.tex (LineNumberTest.tex 1 1 ) No pages of output. As you can see, eTeX's \scantokens exhibits the same behaviour (most likely because it also uses a virtual file). But I think that tex.print will be used more often than \scantokens, since it is the only way to pass arbitrary tokens from Lua to TeX. This means that when TeX executes these tokens, log messages will use not the line number of the TeX macro that expanded to the Lua code that created new TeX tokens, but the "virtual line number" (as seen in the example above). These log messages (and most likely \errmessage-s as well) are not very useful. Jonathan

2 1

Bug with "\atop"
by Olivier Binda 23 Sep '07

23 Sep '07

Hello.... Luatex crashes when trying to compile the following minimal file (100% reproductible bug on my config) : ${1\atop 2}$\end I'm using W32TeX and the luatex version that comes bundled with it. Keep up the good work, Olivier

4 6

Cron <luatex@server> /home/luatex/luatex-svn
by root＠aanhet.net 21 Sep '07

21 Sep '07

U luatex/trunk/src/texk/web2c/luatexdir/Makefile.in A luatex/trunk/src/texk/web2c/luatexdir/lua/hnjalloc.h A luatex/trunk/src/texk/web2c/luatexdir/lua/hyphen.h U luatex/trunk/src/texk/web2c/luatexdir/lua/ltexlib.c A luatex/trunk/src/texk/web2c/luatexdir/lua/llanglib.c U luatex/trunk/src/texk/web2c/luatexdir/lua/luastuff.c A luatex/trunk/src/texk/web2c/luatexdir/lua/hnjalloc.c A luatex/trunk/src/texk/web2c/luatexdir/lua/hyphen.c Checked out revision 745.

1 0

Notes and comments on beta 0.11.2
by Jonathan Sauer 21 Sep '07

21 Sep '07

Hello, (complete this time, I accidentally pushed Ctrl-Enter) I tried out beta 0.11.2 yesterday, and it works great. Still, I noticed a few things that seem weird or wrong. First thing: I tried to use 'require', and the error message I got was slightly weird: >>> ! LuaTeX error [string "luas[0]"]:1: module '/Dokumente/Writing/trunk/Writing/T eX/tex/KTeX/lua/l-callback.klua' not found: no field package.p. \LU:require ...ed {require(kpse.find_file('#2'))}} <<< I look into Lua 5.1.2 and found this error message in loadlib.c: >>> lua_pushfstring(L, "\n\tno field package.preload['%s']", name); <<< A variation (shorter pathname, thus longer message): >>> ! LuaTeX error [string "luas[0]"]:1: module 'l-callback.klua' not found: no field package.preload['l-callback.klua'] no file './l-callback/k. <<< So it seems that Lua's error message is truncated by LuaTeX. Second thing: I typeset a document of mine and got quite a few overfull boxes (more on this later). I looked into the log file, which my editor warned me was not valid UTF-8. Following is an excerpt of said log file. It seems that an UTF-8 sequence is broken across two lines: >>> Overfull \hbox (1.53935pt too wide) in paragraph at lines 1493--1498 \FS:O@font@t1/lmr/r/r/n/n/10@font er-wi-der-te Kaya. Sie zog ih-re Freun-din nÃ ¤-her zu sich, zog Ataamis| <<< Third thing: I am currently typesetting documents with pdfTeX as well as with LuaTeX. The documents use Latin Modern with font expansion and protrusion. While pdfTeX typesets them without any overfull/underfull boxes, LuaTeX produces quite a few overfull boxes (one of them above). Before I investigate further: Is it possible that font expansion is currently disabled in LuaTeX? And a question regarding hyphenation trie size: pdfTeX, Version 3.141592-1.40.3 (Web2C 7.5.6): >>> Beginning to dump on file ../../web2c/ktex.fmt (format=ktex 2007.9.19) 3176 strings of total length 48723 7253 memory locations dumped; current usage is 64&7010 1616 multiletter control sequences \font\nullfont=nullfont 0 words of font info for 0 preloaded fonts 14 hyphenation exceptions Hyphenation trie of length 13987 has 388 ops out of 35111 207 for language 1 ^^^^^ 181 for language 0 No pages of output. Transcript written on ../../web2c/ktex.log. <<< luaTeX, Version 3.141592-beta-0.11.2-2007091918 (Web2C 7.5.6) >>> Beginning to dump on file ../../web2c/luaktex.fmt (format=luaktex 2007.9.19) 2992 strings of total length 48651 7223 memory locations dumped; current usage is 48&7048 1717 multiletter control sequences \font\nullfont=nullfont 0 preloaded fonts 0 words of active ocps \ocp\nullocp=nullocp 0 preloaded ocps \ocplist\nullocplist=[] 14 hyphenation exceptions Hyphenation trie of length 79256 has 388 ops out of 35111 207 for language 1 ^^^^^ 181 for language 0 No pages of output. Transcript written on ../../web2c/luaktex.log. <<< Is this difference due to Unicode? And finally: \catcode changes are not logged when using \tracingassigns. \pdfTeX logs these changes. These things do not diminish my excitement about LuaTeX. Thank you for your time for developing it! Jonathan

3 6