Closed Bug 421673 Opened 16 years ago Closed 5 years ago

Another hang involving large column-count

Categories

(Core :: Layout, defect)

defect
Not set
critical

Tracking

()

RESOLVED DUPLICATE of bug 288704

People

(Reporter: j.moz, Unassigned)

References

Details

(Keywords: hang, testcase, Whiteboard: multicol)

Attachments

(2 files, 3 obsolete files)

Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9b5pre) Gecko/2008030704
Minefield/3.0b5pre

Several iExploder tests with large -moz-column-count hang the browser. This is similar to bug Bug 345583.
This is an example of a hanging iExploder test.

Changing "-moz-column-count: 46288386" to "-moz-column-count: 1" makes the hang go away.
The -moz-column-count is a large negative number in some hanging iExploder tests (for example -moz-column-count: -25084547 in test 40045453).

iExploder generates this hang a lot.
Is bug 345583 comment 12 relevant here?
Probably the same underlying issue as bug 421671 or bug 404140.
Depends on: 421671
Keywords: testcase
Product: Firefox → Core
QA Contact: general → general
Those bugs (and bug 444027) are now fixed, but this bug still hangs.

Also hangs on Windows.

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2a1pre) Gecko/20090210 Minefield/3.2a1pre
Attached file reduced test case (obsolete) —
Reduced from the previous test case:

<ul style="-moz-column-count: 99;">
followed by lots of <div>'s and 0x08 characters.

Note that there are 16 ASCII BS characters (0x08) on the last (non-empty) line.

I guess this isn't technically a "real" hang: the browser freezes completely, but it will eventually load the page (removing, for example, a couple of <div>'s will make it load faster).
Attachment #308135 - Attachment is obsolete: true
Possibly the same bug.  We're hanging inside reflow,
nsColumnSetFrame::ReflowChildren creates new next-in-flows... desired width
is 1 twip, it goes slower and slower, leaving it running for several hours
only creates 100 children or so... the reason for the slowness is
nsBlockFrame::RenumberLists ?  (if I interrupt it in a debugger I usually
have that at the top of the stack)
Status: UNCONFIRMED → NEW
Component: General → Layout
Ever confirmed: true
Keywords: hang
OS: Linux → All
QA Contact: general → layout
Hardware: x86 → All
Attached file reduced test case (v2) (obsolete) —
The original reduced test case doesn't hang anymore, but this one does.

I'm guessing it's still the same hang: the difference to the original (attachment 363300 [details]) is that there's some text in the middle and in the end, and the BS characters are gone.
Attachment #363300 - Attachment is obsolete: true
Still happens, even with patch for bug 730559.
Whiteboard: multicol
Removed the -moz prefix from the test case, because prefixed multi-column support will be removed (bug 1300895, bug 1308636).
Attachment #598910 - Attachment is obsolete: true
Summary: Another hang involving large -moz-column-count → Another hang involving large column-count

The test case in comment 10 is fixed by this push range, which contains mutiple bugs. However, based on comment 7, RenumberLists is what causes the hang, and Bug 288704 removes this function, so I think bug 288704 fixes this.

Status: NEW → RESOLVED
Closed: 5 years ago
Resolution: --- → DUPLICATE
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: