New variables structure #53
Conversation
Do not include storage modifiers in the variables matches because it's not great to re-use them in code blocks in that case. This also made it possible to put `#variables` as an include in `#code`, which removes the need for having both `#code` and `#variables` in many blocks. Fixes: atom#50
| 'include': '#code' | ||
| } | ||
| ] | ||
| 'include': '#anonymous-block-and-instance-initializer' |
There was a problem hiding this comment.
No actual changes, just some splitting up code to make it more readable.
| 'applyEndPatternLast': 1 | ||
| 'begin': ''' | ||
| (?x:(?= | ||
| (?: |
There was a problem hiding this comment.
Since we're in a lookahead all the non-capturing groups aren't needed. We're not capturing anything anyway.
There was a problem hiding this comment.
Right you are. Copied this from the previous definition, but might as well simplify!
| @@ -999,50 +1001,54 @@ | |||
| } | |||
| ] | |||
| 'variables': | |||
| 'applyEndPatternLast': 1 | |||
There was a problem hiding this comment.
I removed this. To be frank, I don't know how It worked with it before, since ; is matched by #code which was included before too.
There was a problem hiding this comment.
Surprisingly first-mate actually recognizes this. The super-sparse documentation seems to suggest that this only matters when the end patterns of a subpattern and the overarching end are exactly the same, but if everything already works, then no need to have it!
There was a problem hiding this comment.
Cool. In my manual tests it seems to work well, and the specs agrees.
|
NB. Added this to commit message
|
|
⛵ |
|
A multi-array variable still isn't highlighted correctly, eg |
|
Looks okay to me... what do you mean? |
Do not include storage modifiers in the variables matches
because it's not great to re-use them in code blocks in that
case.
This also made it possible to put
#variablesas aninclude in
#code, which removes the need for havingboth
#codeand#variablesin many blocks.One difference here is that storage modifiers (
private,final,static, etc) will no longer be tagged withmeta.definition.variable.java,and it shouldn't be IMO. It's not part of the definition, but rather a modification to
a definition. Syntax highlighting remains the same.
This also fixes my comment in #47, but not #47 itself.
Fixes: #50