summaryrefslogtreecommitdiff
path: root/src/luaotfload-init.lua
Commit message (Collapse)AuthorAgeFilesLines
* [init] treat known fontloaders as special casePhilipp Gesang2015-11-181-3/+9
| | | | | Defaulting too soon will prevent loading other loaders like the TL 2014 one we ship for backwards compatibility.
* [init] handle case for reference loader firstPhilipp Gesang2015-11-181-2/+9
| | | | Since it’s the trivial case, expected in 99 % of runs.
* [init] fix pathless Context module loadingPhilipp Gesang2015-11-181-3/+7
|
* [main,init] implement path dependent loading of context modulesPhilipp Gesang2015-11-171-26/+35
|
* [init] unify both Context load branchesPhilipp Gesang2015-11-171-40/+63
|
* [init] rework fontloader choicesPhilipp Gesang2015-11-171-6/+29
|
* [init] access the context loader from initPhilipp Gesang2015-11-101-1/+36
| | | | Works fine now by choosing the “context” fontloader in luaotfloadrc.
* [init,conf] enable direct loading of fontloader componentsPhilipp Gesang2015-11-091-20/+33
| | | | | | | | | | | | | | For now, this pertains only to the modules we ship in Luaotfload. For loading from Context we still need a step to distinguish between the namespaced versions of the files, the pure ones in texmf/…/base, and the Luatex-Fonts ones. But yeah, setting [run] fontloader = unpackaged in the luaotfloadrc now works splendidly.
* [main,init] add no-op loader for fontloader filesPhilipp Gesang2015-11-091-15/+23
|
* [*] kill off file headersPhilipp Gesang2015-11-051-2/+0
| | | | We have the VCS info in the status file; these things are just silly.
* [main, *] convert for centralized initialization routinePhilipp Gesang2015-09-271-0/+2
|
* [main, parsers] prepare for deferred initializationPhilipp Gesang2015-09-271-2/+2
|
* [main,init,db,resolvers] separate resolvers from main into separate filePhilipp Gesang2015-07-231-0/+2
|
* [init,main,package] split initialization in early and late stagePhilipp Gesang2015-07-221-4/+14
| | | | | | | | | | | | | | | | | | | | | This is a big move. Due to the restructuring of the initialization we can now separate the early tasks -- mostly dirty stuff like writing globals -- from the later stage where we actually inject the actual font loader. As a consequence, the basic configuration is now present before the font loader is injected. It’s not complete, though, in that the reconfiguration step must still be executed after the font loader is there, since some config variables require values to be set up by the loader. A decision is pending regarding how we address the situation; possible approaches are, in order of increasing complexity: * reconfigure twice, skip some of the steps the first time through * split reconfiguration into two stages * make configuration independent of fontloader values Design-wise, the last point would be the most desirable, naturally. Though the fontloader has been known to require extensive static preparation which might very well make that strategy impossible or unjustifiably demanding. We’ll see how it plays out …
* [init] make saved globals inaccessiblePhilipp Gesang2015-07-221-10/+7
|
* [init,override] move AGL initialization into post fontloader hookPhilipp Gesang2015-07-161-0/+68
|
* [init] implement hook list for post-fontloader initializationPhilipp Gesang2015-07-161-5/+30
|
* [init] remove annyoing bannerPhilipp Gesang2015-06-211-16/+1
|
* [main, init] partition initialization into a series of routinesPhilipp Gesang2015-06-211-91/+110
|
* [init] move late initialization stage from toplevel into dedicated subroutinesPhilipp Gesang2015-06-201-36/+47
|
* [init] adapt log messagesPhilipp Gesang2015-06-201-12/+12
|
* [init] add stub for module sanitizationPhilipp Gesang2015-06-201-0/+8
|
* [main, init] move fontloader-specific stuff completely into initPhilipp Gesang2015-06-201-0/+252
|
* [main,init] separate loading of fontloader and lualibs packagesPhilipp Gesang2015-06-091-0/+21
|
* [init] prepare fontloader initialization componentPhilipp Gesang2015-05-261-0/+38