- dev-context - NTG Mailing Lists

\nonknuthmode breaks _ in protected mode
by Wolfgang Schuster 02 Feb '11

02 Feb '11

Hi Hans, the \nonknuthmode in core-def.mkiv breaks _ in modules: \starttext \the\catcode`\_ \unprotect \the\catcode`\_ \protect \the\catcode`\_ \stoptext It helps when you enable \nonknuthmode with \everystarttext (or \everyjob), e.g. \appendtoks \nonknuthmode \to \everystarttext Wolfgang

2 4

beta
by Hans Hagen 30 Jan '11

30 Jan '11

Hi, [for those with access to the ftp server] There have been some changes to the resolver code. In order to make trees relocatable (handy if you run from a usb stick or so), the texroot is now more abstract. The final implemention might change but it seems to work ok (here). It can be tested with an existing minimal: - delete the context path under texmf-cache - unzip cont-tmf.zip in texmf-context - copy mtxrun.lua from the scripts path to the bin path - update the texmfcnf.lua file (just copy contextcnf.lua) - remake the file database In logs you will notice prefixes like selfautoparent:texmf which comes from the texmfcnf.lua file (plus a preset TEXMF). That abstraction is used in cached names and is resolved at runtime. It might change to an url like syntax at some point. I hope that Mojca does not get too fed up after this change -) Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------

2 7

Help regarding Kpsewhich and mtxrun
by Julius Canute 23 Jan '11

23 Jan '11

Please correct me if i am wrong. If kpsewhich is not used and replaced by mtxrun why is it included in context minimals?? Doubt: Inside texexec file i find this line: mtxrun --usekpse --execute texexec "$@" Suppose say i type the following command. texexec --pdf sample.tex It will somehow call mtxrun which in turn will invoke 'texmfstart.rb' i guess. The above command somehow invokes 'pdftex' which processes sample.tex file. But pdftex is a binary how does it come to know about location of fonts and all other information unless there is some interface that is hard coded???

3 2

\usemodule and prtcatcodes
by Wolfgang Schuster 18 Jan '11

18 Jan '11

Hi, why does \usemodule load the requested file with “prtcatcodes” settings when nearly each module has \unprotect and \protect at the begin and end of the file? I ask this because of a recent question [1] on the list where a user defined shortcuts for math. Maybe a module isn’t the perfect thing for these things and he can put them in a environment but now comes the next problem. When you load a environment in this way: \environment myenvironment \starttext … \stoptext the file is loaded again with prtcatcodes while in the following form \startproduct myproduct \environment myenvironment … \stopproduct the environment is loaded with ctxcatcodes. While prtcatcodes make sense for a module i can’t understand why they are used for environments. [1] http://www.ntg.nl/pipermail/ntg-context/2011/055915.html Wolfgang

2 5

Help regarding context architecture and compilation
by Julius Canute 17 Jan '11

17 Jan '11

Hi, I am doing a study about context. I wanted to understand context and its interaction with other components in the system. Is there a architecture diagram that could describe it ? Also i want to make context available through fink. For which i need to understand the dependencies of context. Looking for a way to build context from the source manually without using minimal context? Can anyone help me do the compilation manually from source or at-least point to some references to do that? Regards A.Julius Canute

5 9

new beta
by Hans Hagen 10 Jan '11

10 Jan '11

Hi, I uploaded a new beta: The character definitions have been updated (esp math, thanks to Aditya). After some testing by Wolfgang the decision was made to permit _ in macronames in modules. So, the following works: \unprotect \def\test_test_test#1#2{#1+#2=whatever} \test_test_test{1}{2} \protect In principle we can also support ^. The mkiv code base no longer uses _ and ^ in math in modules and those who write modules should avoid them as well. So, when writing macros in unprotected mode, one schould use: $x\normalsuperscript2\normalsubscript n$ Taco and I think that we can safely switch to \nonknuthmode in mkiv so this might happen soon (which means that _ and ^ become normal characters in text mode). Actually, being permanently in \nonknuthmode would permit _ and ^ in modules again. We could even make _ and ^ always letters outside math in which case \text_test etc would become possible everywhere but that needs some more consideration. In that case \def\make_subscript#1#2{\mathematics{#1_{#2}}} \make_subscript{x}{2} becomes valid everywhere. We can discuss this in detail at the dev list. Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------

2 1

Bugs regarding module documentation
by Wolfgang Schuster 06 Jan '11

06 Jan '11

Hi Hans, 1. Typo (s-mod-01.mkii|mkiv) Line 40: two ! in the stretch value [\v!verytolerant,\v!!stretch] 2. It’s not possible to produce a pdf from mkvi file (mtx-modules.lua) Line 121: Add “mkvi” to the list of suffixes local suffixes = table.tohash { 'tex','mkii','mkiv','mp' } Kind regards, Wolfgang

2 1

contextgarden.net
by Patrick Gundlach 17 Dec '10

17 Dec '10

Hi, as I have mentioned on the conference: I am looking for a person (or a group) to take over the server contextgarden.net. Not only the wiki or the modules section or ..., but the complete server. I have started everything more then seven years ago and we have all made it possible that contextgarden.net is now used by more then 500 unique visitors each day (the wiki alone). The average time a user spends on the wiki is three and a half minute (according to the stats, but this is inaccurate, of course), I think that is pretty good! But as usual in a setting like this, I have shifted my focus. This does not mean I like ConTeXt less then several years ago! By no means! But I am too busy with my new job and my family so there is pretty much no server maintenance (except for a reboot once in a while). I really hope that someone steps in and I can hand over all the files/databases and the domain name before the next conference. Do you think that this is worth posting on the main mailing list as well? Patrick

2 1

Luatex 0.65.0 announcement
by Taco Hoekwater 13 Dec '10

13 Dec '10

Hi, I have just uploaded the archives for a new luatex release, 0.65.0. This release mostly fixes a number of bugs, but it also deprecates a number of existing lua and macro commands. Please read the News section carefully if you are the current maintainer of a luatex macro package. News: * The Aleph and Omega version primitives are deprecated. * The Apple Snow Leopard Truetype font collections now work properly. * Libpng has been updated to 1.4.4, and zlib to 1.2.5. * New lua functions node.next() and node.prev(). * The already deprecated for some time node.protrusion_skipable() function has been removed. * fontloader.apply_featurefile() and fontloader.apply_afmfile() now return nil,<errortable> on failure. * The various node list fields in the lua node interface have been renamed from 'list' to 'head', e.g. tex.box[0].head, and the 'list' name is deprecated. For now, the deprecation is silent, as 'list' will stay for quite some time. Please update your lua source as soon as possible, though. * The call node.type(<node>) now returns "node". The existing functionality is otherwise unchanged. * Luafilesystem has been updated to 1.5.0. NOTE: there is an incompatible change in the directory iterator method. * The restriction on linebreak_filter has been eased: at least one horizontal box should be returned, but that is not necessarily the last node any more. * Backward compatibility has been restored for \meaning\mathchar * node.first_character() has been renamed to node.first_glyph() to better match its actual functionality. * The \Umathaccent primitive now supports both/bottom keywords, and \Umathbotaccent / \Umathaccents are now deprecated. * The \Umathaccent primitive now also supports the fixed keyword, which supports non-growing math accents even if the font says there are extensible variants. * Ongoing work on the epdf library. Bugfixes: * Assignment to \dimen65535 was broken earlier. * <box>.head.prev is now always initialized to null (0). * luatex -lua=<script> '\relax' no longer exits immediately, instead going into interactive mode as it should. * The \jobname was initialized wrongly for input file names with multiple dots in them. * The synctex file is now written in the right place when --output-directory is given. * \letterspacefont could cause crashes, especially when combined with font expansion. * Luatex would crash on linux32 systems if it was build in an environment where LFS (>2GB files) was enabled by default. * Node lists created by node.copy_list() could not be hyphenated due to an assertion error. * Synctex support was partially broken since 0.60.0. * With OpenType math fonts, accents with multiple parts (like the \vec arrow) forgot to include all parts if the accentee was too narrow. * With OpenType math fonts, accents over any class except 0 and 7 were misplaced. * kpse.lookup("plain", {all = true}) was broken in kpathsea itself (no extension given as well as --all specified). * A FontForge bug made Cambria Math accents look bad, and the placement of math top accents did not allow a marker at the origin (patches by Stefan Philipp) * Cut&Paste from PDF was sometimes broken due to bad ToUnicode map. * Many small manual improvements (thanks to Paul Isambert) * An other bugfix to lpeg 0.9 (from Roberto Ierusalimsky) * Some small fixes to writet1.w (from Han The Thanh) * A build failure because of bad libz configuration on Win32 The archives can be downloaded from supelec as usual: http://foundry.supelec.fr/gf/project/luatex/ You could also check out the sources via anonymous svn: svn co http://foundry.supelec.fr/svn/luatex/tags/beta-0.65.0 Bugs and feature requests can be added to the issue tracker at http://tracker.luatex.org Have fun, Taco

1 0

Luatex 0.64.0 announcement
by Taco Hoekwater 12 Nov '10

12 Nov '10

Hi, I have just uploaded the archives for a new luatex release, 0.64.0. This is a development release, but it also fixes a few rather annoying bugs in 0.63.0. News: * New lua functions pdf.refobj(), pdf.maxobjnum() and pdf.objtype() * New lua function node.is_node() * The max_strings configuration variable's maximum value is raised to be identical to texlive's maximum (2097151), and same for param_size (32767). * The embedded mplib is now 1.503. * Minor cleanups in the output of fontloader.to_table(): - removed out-of-range altuni's from the output - don't dump the extra encodings, just the actual font encoding - export the encoding's name also as 'enc_name' at one level up - no longer exporting unset macstyles, nor zero uniqueids - fix creationtime and modification time * Ongoing work on the epdf library Bugfixes: * Fix a bug where an explicit hyphen prevented hyphenation for the whole rest of the paragraph * Internal font copying forgot to copy the last font dimen * os.selfdir() is now documented * patch quad (\fontdimen6) in \letterspacefont to be compatible with pdftex * Multi-page pdf inclusion was very slow in 0.63.0 * Some extra checks added to the handling of glue_spec node objects in lua code. * Some node.prev pointers have been corrected (but this is not quite finished). * A small fix to lpeg 0.9. The archives can be downloaded from supelec as usual: http://foundry.supelec.fr/gf/project/luatex/ You could also check out the sources via anonymous svn: svn co http://foundry.supelec.fr/svn/luatex/tags/beta-0.64.0 Bugs and feature requests can be added to the issue tracker at http://tracker.luatex.org Have fun, Taco

1 0