convert-grammars now supports a few flags that we can use to make it
dump out the YAML just for the local grammar submodules. We can then
compare this to the YAML that's actually in grammars.yml to check that
they're the same. If they aren't, grammars.yml needs to be updated.
This will help catch mistakes like using the wrong scope name.
We tell apt-get to download it to vendor/apt and then install it into a
vendor/icu directory.
We should be able to just specify --with-icu-dir, but apparently Ruby
2.0 (but not 1.9 or 2.1 or 2.2) has a bug that requires us to use
--with-icu-include/--with-icu-lib instead. Otherwise it can't find the
ICU libraries.
This runs 8 `git submodule update` processes in parallel, speeding up
bootstrap from 2 minutes to 30 seconds for me. (Obviously this is
dependent on bandwidth.)
This makes it so we don't have to redownload all the grammars every time
we build the grammars gem. It will also let us verify that grammars.yml
is accurate in the future by checking it against the submodules on disk.
script/bootstrap now updates the submodules.
The purpose of this gem is to package up the language grammars that are
used for syntax highlighting on github.com. The grammars are TextMate,
Sublime Text, or Atom language grammars, converted to JSON and given the
filename SCOPE.json, where SCOPE is the language scope that the grammar
defines.
The github-linguist-grammars gem packages up all the grammars, and also
exports a Linguist::Grammars.path method to locate the directory
containing the grammars.
To build the gem, simply run `rake build_grammars_gem`. The grammars.yml
file lists all the repositories we download grammars from, as well as
which scopes are defined by each repository. The
script/download-grammars script takes that list and downloads and
processes the grammars into the format expected by the gem.
