Should the Host Language Label be used if it's empty? #234
Comments
|
related w3c/html-aam#533 |
|
Not that it answers this question, but I wanted to note that |
The Host Language Label step [1] of the accname computation algorithm is unclear on whether the empty string should be used if that's what the native markup provides. I've filed an issue to clarify this at [2]. Meanwhile, other engines will allow a source later in the name computation steps to provide a name even if an empty `alt` attribute was set; see results for the "img with tooltip label with empty alt" case in accname/name/comp_tooltip.html [3]. So, switch Blink's behavior to match the other engines. However, we still ensure that `name_source` is still set to `kAttributeExplicitlyEmpty` returning from`AXNodeObject::TextAlternative` if `alt` was set to the empty string and a name was not provided from another source. This is necessary to preserve the behavior where setting `alt=""` causes the image to be treated as presentational and ignored. If a valid name is provided from another source, however, the other name overrides that behavior. [1] https://w3c.github.io/accname/#comp_host_language_label [2] w3c/accname#234 [3] https://wpt.fyi/results/accname/name/comp_tooltip.html?label=master&label=experimental&aligned&view=interop&q=label%3Ainterop-2024-accessibility Bug: 325612090 Change-Id: I38e510ce27319afad406a9c128a47288c2c087ee Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5382253 Commit-Queue: Aaron Leventhal <[email protected]> Reviewed-by: Aaron Leventhal <[email protected]> Cr-Commit-Position: refs/heads/main@{#1282227}
|
The test reflects the desired behavior but I have long been of the opinion that, since modern IDEs will pre-fill I do not think the specification should change here. If anything, we could add a note that this is a special case. |
The Host Language Label step E of the accname computation algorithm is unclear on whether the empty string should be used if that's what the native markup provides.
Going by a strict reading of the spec, I'd guess that an empty string should be used. However the
img with tooltip label with empty altsubtest of accname/name/comp_tooltip.html suggests that an emptyaltattribute should be ignored. Gecko and WebKit pass this test, while Blink fails.If that test reflects the intended behavior, should the spec be updated to something like this?
Also -- if the empty string is omitted, what about a whitespace-only string? (See #233).
The text was updated successfully, but these errors were encountered: