719014 - DeCOMtaminate nsParser instantiation

Status: NEW

(Core :: DOM: HTML Parser, enhancement, P5)

Description

[Henri Sivonen (:hsivonen)]

Instead of do_createInstance, nsParser should be instantiated with a static factory method (to avoid exporting nsParser.h).

Comment 1

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Attachment: Patch v1

Comment 2

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Comment on attachment 589857 [details] [diff] [review]
Patch v1

Bah, this doesn't work. nsParser::Init() isn't called in time. I'll try to figure something out over the weekend.

Comment 3

[Henri Sivonen (:hsivonen)]

Regarding my earlier IRC comment: I now see that there's no such things as nsParserModule.h, so I guess putting a static ::Create() on nsIParser itself is the least trouble.

It would probably be polite to comm-central to split the part that removed the ability to instantiate nsParser the old way into a separate patch so that the patch that adds nsIParser::Create() can land first, then comm-central can migrate to using it and only then the do_CreateInstance support goes away.

Comment 4

[Henri Sivonen (:hsivonen)]

(In reply to Ms2ger from comment #2)
> Bah, this doesn't work. nsParser::Init() isn't called in time.

You could add explicit InitializeStatics()/ReleaseStatics() calls to nsLayoutStatics instead of relying on the old parser module doing stuff.

(In reply to Henri Sivonen (:hsivonen) from comment #3)
> It would probably be polite to comm-central

The comm-central dependencies have gone away.

Comment 5

[Jens Stutte [:jstutte]]

Bulk-downgrade of unassigned, 4 years untouched DOM/Storage bugs' priority.

If you have reason to believe this is wrong (especially for the severity), please write a comment and ni :jstutte.