NEW 119116
Typing style shouldn't inherit style from atomic elements 
https://bugs.webkit.org/show_bug.cgi?id=119116
Summary Typing style shouldn't inherit style from atomic elements
Ryosuke Niwa
Reported 2013-07-25 17:18:21 PDT
We need to fix the bug Blink attempted to fix in https://chromium.googlesource.com/chromium/blink/+/27371d2980715475f3443f7025141ebff8537179 which is that we sometimes compute the typing style inside an atomic element. Unfortunately, the Blink fix is wrong in that we end up not saving any typing style when the insertion position is inside an atomic element. Instead, we need to walk up the tree to get out of the atomic element and compute the typing style then.
Attachments
Add attachment proposed patch, testcase, etc.
Ahmad Saleem
Comment 1 2022-08-25 14:30:46 PDT
https://github.com/WebKit/WebKit/pull/3675
Ahmad Saleem
Comment 2 2023-05-22 14:39:46 PDT
Now since we have imported 'editing' test suites from WPT, I did this change in local build based on my previous PR and run test suite, it progress following tests: editing/run/delete.html?7001-last <--- 3 tests editing/run/forwarddelete.html?6001-last <--- 3 tests editing/run/inserttext.html?2001-last <---- 3 tests editing/run/multitest.html?9001-last <--- 18 tests ^ It progresses following without any regressions (only changes in failing expectations but we are failing them anyway. I think we have WPT coverage plus we are not failing any new additional test case, it should be good candidate to merge. @Darin & @Ryosuke - any input? Should I go ahead and do another PR.
Darin Adler
Comment 3 2023-05-22 15:08:07 PDT
Yes, that sounds good. I bet you can take it to completion with that kind of start. My question is how do we deal with this observation that Ryosuke made in 2013: "the Blink fix is wrong in that we end up not saving any typing style when the insertion position is inside an atomic element" Maybe Blink dealt with that later, maybe not. One way or another we need to make sure we have tests that cover that issue. If they are not already in WPT, we could add them there.
Ahmad Saleem
Comment 4 2023-05-22 16:16:08 PDT
(In reply to Darin Adler from comment #3) > Yes, that sounds good. I bet you can take it to completion with that kind of > start. > > My question is how do we deal with this observation that Ryosuke made in > 2013: > > "the Blink fix is wrong in that we end up not saving any typing style when > the insertion position is inside an atomic element" > > Maybe Blink dealt with that later, maybe not. One way or another we need to > make sure we have tests that cover that issue. If they are not already in > WPT, we could add them there. I tried to did stupid quick search using 'revision' (r154067) and 'review' (18452012) number on Monorail to see, if anything mentioned these for follow-up but it didn't seems. Also, we start matching exact same numbers of passing tests as of Blink on WPT, if we merge this except 1 or 2 cases. I'll wait for Ryosuke's input before opting to do PR but happy to do PR plus follow-up bug as well to track Ryousuke's concern.
Darin Adler
Comment 5 2023-05-22 16:17:38 PDT
There’s no guarantee that those test cases already exist. We might have to create them.
Note You need to log in before you can comment on or make changes to this bug.