Regex in Sublime Text tmLanguage file doesn't use multiline - regex

Question

I'm trying to create a custom syntax language file to highlight and help with creating new documents in Sublime Text 2. I have come pretty far, but I'm stuck at a specific problem regarding Regex searches in the tmLanguage file. I simply want to be able to match a regex over multiple lines within a YAML document that I then convert to PList to use in Sublime Text as a package. It won't work.
This is my regex:
/(foo[^.#]*bar)/
And this is how it looks inside the tmLanguage YAML document:
patterns:
- include: '#test'
repository:
test:
comment: Tester pattern
name: constant.numeric.xdoc
match: (foo[^.#]*bar)
If I build this YAML to a tmLanguage file and use it as a package in Sublime Text, I create a document that uses this custom syntax, try it out and the following happens:
This WILL match:
foo 12345 bar
This WILL NOT match:
foo
12345
bar
In a Regex tester, they should and will both match, but in my tmLanguage file it does not work.
I also already tried to add modifiers to my regex in the tmLanguage file, but the following either don't work or break the document entirely:
match: (/foo[^.#]*bar/gm)
match: /(/foo[^.#]*bar/)/gm
match: /foo[^.#]*bar/gm
match: foo[^.#]*bar
Note: My Regex rule works in the tester, this problem occurs in the tmLanguage file in Sublime Text 2 only.
Any help is greatly appreciated.
EDIT: The reason I use a match instead of begin/end clauses is because I want to use capture groups to give them different names. If someone has a solution with begin and end clauses where you can still name 'foo', '12345' and 'bar' differently, that's fine by me too.

Answer 1

I found that this is impossible to do. This is directly from the TextMate Manual, which is the text editor Sublime Text is based on.
12.2 Language Rules
<...>
Note that the regular expressions are matched against only a single
line of the document at a time. That means it is not possible to use a
pattern that matches multiple lines. The reason for this is technical:
being able to restart the parser at an arbitrary line and having to
re-parse only the minimal number of lines affected by an edit. In most
situations it is possible to use the begin/end model to overcome this
limitation.
My situation is one of the few in which a begin/end model cannot overcome the limitation. Unfortunate.

Answer 2

Long time since asked, but are you sure you can't use begin/end? I had similar problems with begin/end until I got a better grasp of the syntax/logic. Here's a rough example from a json tmLanguage file I'm doing (don't know the proper YAML syntax).
"repository": {
"foobar": {
"begin": "foo(?=[^.#]*)", // not sure about what's needed for your circumstance. the lookahead probably only covers the foo line
"end": "bar",
"beginCaptures": {
"0": {
"name": "foo"
}
},
"endCaptures": {
"0": {
"name": "bar"
}
},
"patterns": [
{"include": "#test-after-foobarmet"}
]
},
"test-after-foobarmet": {
"comment": "this can apply to many lines before next bar so you may need more testing",
"comment2": "you could continue to have captures here that go to another deeper level...",
"name": "constant.numeric.xdoc",
"match": "anyOtherRegexNeeded?"
}
}
I didn't follow your
"i need to number the different sections between the '#' and '.'
characters."
, but you should be able to have a test in test-after-foobarmet with more captures if needed for naming different groups between foo bar.
There's are good explanation of TextMate Grammar here. May still suffer from some errors but explains it in a way that was helpful for me when I didn't know anything about the topic.