[webkit-reviews] review denied: [Bug 192396] Web Inspector: Styles: toggling selected properties may cause data corruption : [Attachment 356890] WIP: single agent call
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Sat Dec 8 15:38:44 PST 2018
Nikita Vasilyev <nvasilyev at apple.com> has denied review:
Bug 192396: Web Inspector: Styles: toggling selected properties may cause data
corruption
https://bugs.webkit.org/show_bug.cgi?id=192396
Attachment 356890: WIP: single agent call
https://bugs.webkit.org/attachment.cgi?id=356890&action=review
--- Comment #12 from Nikita Vasilyev <nvasilyev at apple.com> ---
Created attachment 356890
--> https://bugs.webkit.org/attachment.cgi?id=356890&action=review
WIP: single agent call
I (In reply to Nikita Vasilyev from comment #9)
> (In reply to Devin Rousso from comment #8)
> > Comment on attachment 356785 [details]
> > Patch
> >
> > View in context:
> > https://bugs.webkit.org/attachment.cgi?id=356785&action=review
> >
> > r-, due to the potential double agent call
>
> Yes, two calls may happen but only when all these conditions meet:
> - when adding a new property,
> - appending it after the last property,
> - the last property isn't terminated with a semicolon.
I tried doing this in one call. The WIP is attached. I'm not going forward with
this approach.
The problem in the first place wasn't multiplying semicolons alone. That
wouldn't cause data corruption. The data corruption was caused because the
semicolons were a special case that didn't update ranges of the following
properties.
I don't want to make logic harder to follow. It's premature optimization.
I'll add tests for inserting new properties, which were long overdue, and I'll
continue with my previous approach.
More information about the webkit-reviews
mailing list