Closed
Bug 413274
Opened 18 years ago
Closed 18 years ago
"ASSERTION: How can this ever happen?: '*aUnconstrainedSize <= scriptLevelSize'"
Categories
(Core :: MathML, defect)
Core
MathML
Tracking
()
RESOLVED
FIXED
mozilla1.9beta3
People
(Reporter: jruderman, Assigned: vladimir.sukhoy)
References
Details
(Keywords: assertion, regression, testcase)
Attachments
(2 files, 2 obsolete files)
|
329 bytes,
application/xhtml+xml
|
Details | |
|
1.15 KB,
patch
|
roc
:
review+
roc
:
superreview+
beltzner
:
approval1.9+
|
Details | Diff | Splinter Review |
Loading the testcase triggers:
###!!! ASSERTION: How can this ever happen?: '*aUnconstrainedSize <= scriptLevelSize', file /Users/jruderman/trunk/mozilla/layout/style/nsRuleNode.cpp, line 1989
###!!! ASSERTION: scriptminsize should never be making things bigger: 'aFont->mScriptUnconstrainedSize <= aFont->mSize', file /Users/jruderman/trunk/mozilla/layout/style/nsRuleNode.cpp, line 2397
The first assertion is part of code added in bug 355548, so I'm assuming this is a regression from bug 355548.
This looks like an overflow issue.
vlad, I'd love it if you could take this, it shouldn't be too hard, just need some overflow checking in the new font code in nsRuleNode.cpp
|
Assignee | |
Comment 3•18 years ago
|
||
ok, I'll see what I can do..
|
|
Assignee | |
Comment 4•18 years ago
|
||
Since negative scriptsizemultipliers are not parsed, AFAUI we only need to check for overflow, not underflow in that code.
|
|
Assignee | |
Updated•18 years ago
|
Attachment #299085 -
Flags: review?(roc)
Looks fine but I'd actually prefer it if you didn't use a local variable, because the name 'x' isn't very useful
|
|
Assignee | |
Comment 6•18 years ago
|
||
Attachment #299085 -
Attachment is obsolete: true
Attachment #299116 -
Flags: review?(roc)
Attachment #299085 -
Flags: review?(roc)
You forgot the second NSToCoordRound
|
|
Assignee | |
Comment 8•18 years ago
|
||
Comment on attachment 299116 [details] [diff] [review]
without x
oops, nevermind
Attachment #299116 -
Attachment is obsolete: true
Attachment #299116 -
Flags: review?(roc)
|
|
Assignee | |
Comment 9•18 years ago
|
||
Sorry, messed up that one, was doing several things at a time :(
Attachment #299124 -
Flags: review?(roc)
Comment on attachment 299124 [details] [diff] [review]
with 2nd NSToCoordRound
:-)
Attachment #299124 -
Flags: superreview+
Attachment #299124 -
Flags: review?(roc)
Attachment #299124 -
Flags: review+
|
|
Assignee | |
Comment 11•18 years ago
|
||
Comment on attachment 299124 [details] [diff] [review]
with 2nd NSToCoordRound
this minor patch eliminates numeric overflow in script level size computations for mathml.
Attachment #299124 -
Flags: approval1.9?
|
||
Comment 12•18 years ago
|
||
Comment on attachment 299124 [details] [diff] [review]
with 2nd NSToCoordRound
a=beltzner for 1.9
Attachment #299124 -
Flags: approval1.9? → approval1.9+
|
|
Assignee | |
Updated•18 years ago
|
Keywords: checkin-needed
|
||
Updated•18 years ago
|
Assignee: nobody → vladimir.sukhoy
|
|
||
Comment 13•18 years ago
|
||
Checking in layout/style/nsRuleNode.cpp;
/cvsroot/mozilla/layout/style/nsRuleNode.cpp,v <-- nsRuleNode.cpp
new revision: 1.235; previous revision: 1.234
done
Status: NEW → RESOLVED
Closed: 18 years ago
Flags: in-testsuite?
Keywords: checkin-needed
OS: Mac OS X → All
Hardware: PC → All
Resolution: --- → FIXED
Target Milestone: --- → mozilla1.9 M11
You need to log in
before you can comment on or make changes to this bug.
Description
•