Closed OrKoN closed 1 month ago
@OrKoN do we have any blockers for this PR?
@sadym-chromium Mozilla's review I believe
There is still an open question how to deal with request with document as a start node, and maxDepth equals zero:
{
"type": "innerText",
"maxDepth": 0,
"value": "BAR"
}
As long as document does not have innerText
, we can:
@OrKoN @jgraham
@sadym-chromium I think #2 would be something I would expect. I.e., the document does match and the max depth is 0, therefore, no recursive calls.
@sadym-chromium I think #2 would be something I would expect. I.e., the document does match and the max depth is 0, therefore, no recursive calls.
Currently spec says in this case the search should go to all the depth.
I feel like users are probably going to expect not decreasing depth on document elements i.e. passing in document
is equivalent to document.documentElement
, but we should at least be consistent between different selector types. That would also be inconsistent with the way that node serialization works, but maybe that is less important?
@OrKoN are you going to create / update wpt tests for this change?
I feel like users are probably going to expect not decreasing depth on document elements i.e. passing in
document
is equivalent todocument.documentElement
, but we should at least be consistent between different selector types. That would also be inconsistent with the way that node serialization works, but maybe that is less important?
Here is the spec change for this option: https://github.com/w3c/webdriver-bidi/pull/713
@whimboo @sadym-chromium will update it after the remaining issue is resolved
This PR changes the innerText locator to handle non-Element nodes correctly.
Preview | Diff