pug-lexer loc with source endings
#2957
Conversation
packages/pug-lexer/index.js
Outdated
| var key = ''; | ||
| var i; | ||
|
|
||
| characterParser.defaultState(); |
There was a problem hiding this comment.
This line is a no-op. You don't need to call defaultState unless you are going to use the state that gets returned.
packages/pug-lexer/index.js
Outdated
| var col = this.colno; | ||
| var line = this.lineno; | ||
|
|
||
| characterParser.defaultState(); |
There was a problem hiding this comment.
Again, this line is a no-op
packages/pug-lexer/index.js
Outdated
| this.colno = col; | ||
|
|
||
| tok.val = val; | ||
| tok.mustEscape = escapeAttr; |
There was a problem hiding this comment.
Rather than taking and mutating a token. I feel if would be clearer if this function returned {val, mustEscape, remainingSource}
packages/pug-lexer/index.js
Outdated
| , str = this.input.substr(1, index-1); | ||
|
|
||
|
|
||
| attributeName: function(str){ |
There was a problem hiding this comment.
This function should be called attribute since calling it has the effect of lexing both the name and the value. In the future, we might split out a separate attributeName function.
packages/pug-lexer/index.js
Outdated
| this.incrementColumn(1); | ||
| this.tokEnd(tok); | ||
| this.tokens.push(tok); |
There was a problem hiding this comment.
In most places you have done this.tokens.push(this.tokEnd(tok)); - I think we should be consistent here.
|
I added some comments to the commit. I think this looks really good, but it's a bit tricky to review because there's so much git merging happening and it makes it hard to figure out exactly what's changed. If you could rebase everything off master so there's only a few commits. Ideally have the commit that updates all the snapshots be a separate commit from the commit with your manual changes to the source code, that would really help, but if that's too much work, then that's ok. |
jbsulli
force-pushed
the
pug-lexer-source-endings
branch
from
2b6ac09
to
7eacbab
Compare
|
Just updated the commits and removed the merge commits. I will update with your suggestions when I get a chance later today. |
|
Ping me when you get a chance to do those updates. |
- remove no-ops - rename `attributeName` to `attribute` - attributeName no longer takes a tok and returns a reponse object - minor cleanup
|
@ForbesLindesay, I just finished the updates you requested. For PR #2956, it's easy to default |
|
Thanks 👍 |
line/col, there is now alocobject on all tokens that matches the babellocobject (ex:{start:{line:1,column:1},end:{line:1,column:10},filename:"test.pug"}mustEscape: trueto all boolean attributes (ex:input(disabled))attrsinto three separate functions:attrs,attributeNameandattributeValuepug-parserandpugto use newlocfrom lexer tokens