summaryrefslogtreecommitdiff
path: root/luaotfload-fonts.lua
Commit message (Collapse)AuthorAgeFilesLines
* Implement a --force optionKhaled Hosny2010-02-261-6/+12
|
* Checksum loaded fontsKhaled Hosny2010-02-261-9/+32
| | | | | | Store checksums of already loaded fonts in the database to avoids reloading them, this isn't much faster but it avoids reloading the fonts and the huge memory consumption. Is there a faster way to hash a file?
* Get rid of empty lines in the outputKhaled Hosny2010-02-261-3/+0
|
* making the end of the path (after TEXMFVAR) a variableeroux2010-02-261-1/+5
| | | | as it will change soon...
* Making the log function available elsewhereeroux2010-02-261-5/+3
| | | | + Ooops, I missed one commit...
* A safer path normalizationeroux2010-02-261-7/+16
|
* Remove the arg[0] checking hackKhaled Hosny2010-02-261-4/+0
| | | | The script is no longer to be run standalone.
* More robust progress bar buildingKhaled Hosny2010-02-261-3/+3
|
* Actually working checks for writabilityElie Roux2010-02-251-12/+19
| | | | | | | | | Now we scan the fonts only if we can write in the final file. There is currently (if I understand correctly) no clean way to check if a folder is writable or not in Lua: lfs.attributes will give you the string returned by ls -l, but there is no simple way to know if we can actually write, as we don't know if we are the owner, or if we have access to the group...
* Making the final directory changeableElie Roux2010-02-251-5/+14
| | | | Useful for the --sys option
* Fixing duplicate entries in the databaseElie Roux2010-02-251-7/+26
| | | | | | | If the texmf fonts are in fc-cache (which is the case in Debian's TeX Live 2009, there were two entries in the database, one with the filename and the other with the complete path, they were also scanned twice.
* Adding some comments, small changes in the logs.Elie Roux2010-02-251-6/+47
|
* Some README and author updatesKhaled Hosny2010-02-201-2/+2
|
* Fixing cygwin pathsElie Roux2010-02-161-1/+26
|
* Some cosmetic hacksKhaled Hosny2010-02-151-3/+4
|
* Disable progress bar in verbose modeKhaled Hosny2010-02-151-15/+17
|
* Guard against fonts with broken names tableKhaled Hosny2010-02-151-6/+14
| | | | | | Due to a bug in current luatex, some fonts would lack certain required entries in names table, now we'll just ignore such fonts. The real bug is fixed in luatex trunk anyway.
* Adding only modern fonts to the list of os fontsElie Roux2010-02-151-1/+5
|
* Fixing a small "graphical" bug when errors happenElie Roux2010-02-151-0/+7
|
* Implement a fancy progress barKhaled Hosny2010-02-151-7/+31
| | | | Now we have some estimate of when the script will finish.
* Use `fc-list` instead of `fc-cat`Khaled Hosny2010-02-141-126/+28
| | | | | `fc-list` returns a list of installed fonts, which simplify the parsing, and of course I broke non-unix systems!
* More compact layoutKhaled Hosny2010-01-281-12/+9
|
* Fix LM style name issueKhaled Hosny2010-01-281-8/+9
| | | | | | | According to OpenType spec, size feature can define a common style name for different optical sizes of the same font, now we use it. See http://www.microsoft.com/typography/otspec/features_pt.htm#size
* Fix font names preferenceKhaled Hosny2010-01-281-14/+11
| | | | | | Prefer different values of font names as suggested by TrueType specs. See http://developer.apple.com/textfonts/TTRefMan/RM06/Chap6name.html
* Make the font database dumbKhaled Hosny2010-01-251-156/+49
| | | | | Store the 'raw' values, the matching logic should be moved to the resolver function.
* More bugfixes in style guessingElie Roux2010-01-211-10/+19
| | | | | oblique is now considered different as italic, and fixing database shape for fonts with different fontfamilies.
* Improving style guessing based on real-life fontsElie Roux2010-01-211-36/+58
| | | | | | | | | | Now all the informations are printed in the guessed_style, for parsing facilities. Light and Heavy are not treated as weight anymore (some fonts are heavy regular or heavy bold in the same family). Treating the condensed style separately. If it could be tested on a large database it would help me improve it. Of course perverse styles like the ones of semaphor are not guessed, but they are not usual styles so it's not a problem.
* Changing the database shapeElie Roux2010-01-211-7/+9
| | | | | | Now keys are font file names, in order to have all fonts listed (it was not the case all the time). The values are raw_style and guessed_style. Wdyt?
* Improvements in the style guessing systemElie Roux2010-01-211-5/+16
|
* Mouahaha, I got you Windows!Elie Roux2010-01-201-4/+3
|
* A bit of adaptation to the buggy OSElie Roux2010-01-201-5/+75
| | | | | | | | Windows is a pain, as usual... I'm blocked on a bug where the command line passed to popen is reshaped to contain linebreaks instead of spaces, so the fc-cat of TeXLive cannot be called if TeXLive is under C:/Program Files/ (grrr). I should report it to LuaTeX or TeXLive, more certainly to MS but...
* Merge branch 'master' of github.com:mpg/luaotfloadElie Roux2010-01-191-1/+1
|\
| * Missing log levelKhaled Hosny2010-01-191-1/+1
| |
* | Exploring the fontconfig directoriesElie Roux2010-01-191-19/+51
| | | | | | | | | | | | Looking for the directories in fc-cat that actually contain fonts and adding it to the directories to explore. Now system fonts are available even on systems where OSFONTDIR is not set in texmf.cnf.
* | Not relying on require()Elie Roux2010-01-191-5/+1
|/ | | | require() doesn't work with LuaTeX 0.50 from Debian.
* Differenciation between texmf and non-texmf dirsElie Roux2010-01-191-4/+15
| | | | The paths are now absolute for files that are not in the texmf tree.
* Adding strong euristics to guess the styleElie Roux2010-01-191-1/+68
|
* Several improvements:Elie Roux2010-01-191-18/+37
| | | | | | | | | | | | | - making the script work under Windows (cygwin at least) - creating several levels of verbosity - logging explored directories by default - fixing a bug with the . directory (at least under windows) - not exploring same directory twice (the same directory can be in OPENTYPEFONTS and TTFONTS) - using dofile on old luatex versions instead of require (which raises an error) - adding checks before mkdir and file writing (should be tested under Unix though, not really testable under Windows...)
* Prefer names.prefmodifiers over names.subfamilyKhaled Hosny2010-01-161-1/+1
| | | | | This fixes style checking for many correctly named fonts, but breaks crappy fonts like LM and MnSymbol, both aren't a priority for me.
* CleanupKhaled Hosny2010-01-161-22/+11
|
* Drop otfl-luat-dum.lua dependencyKhaled Hosny2010-01-161-8/+7
|
* Remove some no longer needed codeKhaled Hosny2010-01-161-26/+3
| | | | | Now LuaTeX can return proper TTF names, we no longer need these dirty hacks.
* Make non-subfonts entries dictionaries as wellKhaled Hosny2010-01-151-2/+2
| | | | | Even if it is a single value (filename), to make our life easier in matching process.
* Update database version numberKhaled Hosny2010-01-151-1/+1
| | | | | We now use our own version number since we no longer compatible with luatex-fonts' version.
* Sanitize font namesKhaled Hosny2010-01-151-2/+5
| | | | | | | | Remove spaces, hyphens etc. so that we don't have to deal with many irregularities. Now I seem to process all fonts included with TeXlive correctly, and my +2000 files font collection with two errors only.
* Fix more buggy fonts (Linux Libertine Bold et al.)Khaled Hosny2010-01-151-8/+13
| | | | Now this code is getting really ugly!
* Add support for subfonts (ttc and dfont)Khaled Hosny2010-01-151-15/+22
|
* New table structureKhaled Hosny2010-01-151-18/+25
| | | | | | | | | | | | | | | | | | | | Now database looks like: mappings = { families = { ["familyname"] = { ["regular] = filename-regular.ttf ["bold"] = filename-bold.ttf ... } ... } psnames = { ["psnameregular"] = filename-regular.ttf ["psnamebold"] = filename-bold.ttf ... } }
* 'scan_os_fonts' is uselessKhaled Hosny2010-01-141-1/+2
| | | | | | If $OSFONTDIR is set, then it'll be appended to $OPENTYPEFONTS and $TTFONTS anyway, so we are just rescanning fonts. To be replaced by fontconfig based implementation.
* Our own font names extraction implementationKhaled Hosny2010-01-141-2/+50
| | | | | | Instead of using (the faster) fontloader.info, we now fully open the font to get more usable data, my be it is worth extending fontloader.info to cover our needs?