Fix transclude inefficiency #7647
Conversation
The transclusion target was sometimes being parsed twice when transcluding as text/plain Associated test results are also made more consistent
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
core/modules/widgets/transclude.js
Outdated
| // "text/plain" is the plain text result of wikifying the text | ||
| target = this.parseTransclusionTarget(); | ||
| this.sourceText = target.text; | ||
| this.parserType = target.type; |
There was a problem hiding this comment.
The 2 lines above seem to be the same code for all 3 states. They may be moved above the switch statement. Or is it intended to be that way?
There was a problem hiding this comment.
I think this.parseAsInline should also initialized and moved outside the cases.
There was a problem hiding this comment.
The 2 lines above seem to be the same code for all 3 states. They may be moved above the switch statement. Or is it intended to be that way?
Yes that can be made a bit more efficient.
I think
this.parseAsInlineshould also initialized and moved outside the cases.
I don't think so. It is not required on all code paths
| target = this.parseTransclusionTarget(parseAsInline); | ||
| this.sourceText = target.text; | ||
| this.parserType = target.type; | ||
| this.parseAsInline = target.parseAsInline; |
There was a problem hiding this comment.
It seems this.parseAsInline is only initialized if this case has been executed. Is that intentional? IMO undefined means "we do not know". The variable parseAsInline is defined as true or false. Just a gut feeling.
There was a problem hiding this comment.
The code relies on this.parseAsInline defaulting to false
Fixes #7592