Open
Bug 745351
Opened 13 years ago
Updated 2 years ago
[inc][WinXP] After dynamic text insertion within a block, the final word is mispositioned (as compared to static tescase)
Categories
(Core :: Layout, defect)
Tracking
()
NEW
People
(Reporter: dholbert, Unassigned)
Details
Attachments
(3 files)
6.71 KB,
patch
|
Details | Diff | Splinter Review | |
30.30 KB,
text/plain
|
Details | |
203.48 KB,
application/gzip
|
Details |
(discovered this while writing flexbox reftests, and determined it to be a general issue that doesn't require a flexbox)
Patch (upcoming) includes 4 reftests that generate the following text:
> <u>abc def ghi</u> jkl <i>mno pqr stu</i>
using 10px "Ahem" font, wrapped in a 100px-wide block.
* The first two tests insert an (initially-missing) chunk of text after MozReftestInvalidate fires.
* The last two tests are the same, except they do the insertion a bit sooner -- in a function called synchronously from a <script> block.
* The reference case is just static -- no insertions.
The first two tests fail on WinXP for some reason, with the final word being shifted slightly to the right in the testcase with respect to the reference case. They pass on all other platforms.
Reporter | ||
Comment 1•13 years ago
|
||
Reporter | ||
Comment 2•13 years ago
|
||
Here's a Try run with the tests failing on WinXP Reftests, both opt and debug:
https://tbpl.mozilla.org/?tree=Try&rev=22a775212b17
This failure is 100% reproducible on TryServer WinXP, as far as I've seen -- not intermittent at all.
I'm attaching the log of the failure here.
Reporter | ||
Comment 3•13 years ago
|
||
Reporter | ||
Comment 4•13 years ago
|
||
Also, forgot to mention -- I named the tests "dynamic-text-insertion-1-*" in the Try run (so that's the name in the attached reftests logs), since that was before I filed this bug.
In the reftests-patch attached here, the tests are now named "745351-*" (after this bug)
Reporter | ||
Updated•13 years ago
|
OS: Linux → Windows XP
Hardware: x86_64 → x86
Reporter | ||
Comment 5•13 years ago
|
||
This seems to have become fixed recently. (not sure if it was intentionally or unintentionally, & if it was a test-machine-config change vs. a code change)
The attached reftests now pass on WinXP in a try push.
If I get a chance, I'd like to track down what fixed this (and possibly land the tests, if there aren't already similar tests included with whatever fixed this).
Updated•2 years ago
|
Severity: normal → S3
You need to log in
before you can comment on or make changes to this bug.
Description
•