Verbatim parsing #1083
Verbatim parsing #1083
Conversation
This commit introduces the foundations for a more precise HTML parsing mode that preserves the original document structure more faithfully, as much as possible, and instead let the browser fix invalid HTML itself.
…eriment/verbatim-parsing
This commit makes the previously implemented exact HTML parsing mode an opt-in experimental feature rather than the default parsing mode. The new implementation: 1. Adds an `experimentalExactParsingThingy` flag to control whether to use the more precise HTML parsing mode. Will use a more formal name once we think it through 2. Reverts the default insertion mode from `initialIMExact` back to `initialIM` 3. Updates all WASM bindings to pass the experimental flag through to the parser 4. Adds extensive test coverage with duplicated previously failing test cases that verify the behavior with exact parsing enabled The exact parsing mode preserves the original HTML document structure more faithfully, allowing invalid HTML to pass through to the browser rather than having the parser attempt to normalize it. (Commit written with the assistance of AI)
|
Add `ExperimentalExactParsingThingy: true` to test cases from #1070
|
@MoustaphaDev what needs to be done to get this across the line? |
|
Not much really, IIRC only fixing a regression that rendered the head in the wrong position and add tests to cover all issues caused by the current HTML correction behavior. |
|
Btw I let this sit without an update for way too long—owning that. I’ll resume work by Thursday and plan to wrap within two weeks (if the RFC process follows smoothly). Remaining work is what I outlined above; I’ll post an update if anything changes and then take this out of draft. |
No description provided.