bug(parser): quoted strings should not allow interior adjacent spaces #636
Conversation
This doesn't catch all the ways that quoted strings should be treated as constrained quoted text, but does fix this non-pathological test, with fixed test cases. Also added quoted_strings_test for xhtml5 (forgot to add this earlier). Fixes bytesparadise#622
Codecov Report
@@ Coverage Diff @@
## master #636 +/- ##
=======================================
Coverage 86.09% 86.09%
=======================================
Files 70 70
Lines 4661 4661
=======================================
Hits 4013 4013
Misses 399 399
Partials 249 249 |
| return types.NewQuotedString(types.SingleQuote, elements.([]interface{})) | ||
| } | ||
|
|
||
| SingleQuotedStringElements <- elements:(SingleQuotedStringElement+) { | ||
| return types.NewInlineElements(elements) | ||
| } | ||
|
|
||
| SingleQuoteStringStart <- "'`" ![ \t\r\n] |
There was a problem hiding this comment.
you should be able to replace ![ \r\n\t] with [^ \r\n\t]
There was a problem hiding this comment.
No I think that is wrong. I don’t want to consume the token. This leaves the next tokens available in the input stream without consuming them.
There was a problem hiding this comment.
This is important for nesting quotes text for example.
There was a problem hiding this comment.
ah right! I think that &[^ \r\n\t] would work in that case then. I like the & syntax because it really says that it looks forward in the content without consuming the next character.
There was a problem hiding this comment.
I think the two syntaxes are basically the same. Using ! should convey that as well as & with one fewer characters. I guess it’s all what you’re used to.
There was a problem hiding this comment.
ok, I can try and refactor that later if you don't want to make the change here. It's not a big deal indeed.
There was a problem hiding this comment.
Actually there might be a difference - what happens with EOF. Though that difference won’t matter in this particular case it might in others.
More precisely the ! syntax says that the element is NOT followed by one of those characters whereas the & says that it IS followed by a character that is not in that list.
There was a problem hiding this comment.
yes, I believe it's fine here, but it matters for delimited blocks for which the ending delimiter can be omitted when the block is at the end of the document. (Asciidoc is permissive on that regards.)
| @@ -431,7 +435,11 @@ DoubleQuotedStringElement <- element:( | |||
| return element, nil | |||
| } | |||
|
|
|||
| DoubleQuotedStringFallbackCharacter <- ([^\r\n`] / "`" !"\"") { | |||
| DoubleQuoteStringStart <- "\"`" ![ \t\r\n] | |||
This doesn't catch all the ways that quoted strings should be treated
as constrained quoted text, but does fix this non-pathological test,
with fixed test cases.
Fixes #622