386900 - Topic images rendered in wrong place on slashdot

Status: RESOLVED FIXED

(Core :: Layout: Floats, defect)

Description

[Warner Young]

Attachment: Screenshot of incorrect rendering

Starting with the July 2 nightly "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a6pre) Gecko/20070702 Minefield/3.0a6pre", some of the topic images on Slashdot are rendered in the wrong place.  The first 2 topics at the top don't show an image;  instead, 3 images are shown in the 3rd story (see attachment).

This worked correctly in the previous nightly "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a6pre) Gecko/20070701 Minefield/3.0a6pre"

Comment 1

[Steve Chapel]

It occurs in SeaMonkey, too, so it's a core bug.

Comment 2

[Martijn Wargers (dead)]

According to that regression range:
http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=all&branch=HEAD&branchtype=match&dir=&file=&filetype=match&who=&whotype=match&sortby=Date&hours=2&date=explicit&mindate=2007-07-01+04&maxdate=2007-07-02+09&cvsroot=%2Fcvsroot
Regression from bug 386147 or bug 343595?

Comment 3

[Eli Friedman]

I still haven't figured out what's going on, but it's definitely that patch.

Comment 4

[Eli Friedman]

Attachment: Testcase

Ugh, I didn't realize that a band with zero floats might actually have floats in it... patch coming up.

Comment 5

[Eli Friedman]

Attachment: Patch

Comment 6

[Eli Friedman]

Attachment: Patch (diff -w)

Comment 7

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 270995 [details] [diff] [review]
Patch

This might fix the underlying problem in bug 386182.

   PRBool result = PR_TRUE;
   if (0 != mBand.GetFloatCount()) {
     // XXX We should allow overflow by up to half a pixel here (bug 21193).
     if (mAvailSpaceRect.width < aFloatSize.width) {
       // The available width is too narrow (and its been impacted by a
       // prior float)
       result = PR_FALSE;
     }
-    else {
+  }
+
+  if (!result)
+    return result;

Couldn't this all be turned into "if (...) return PR_FALSE;"?

+      if (0 != mBand.GetFloatCount()) {
           if ((xa < mAvailSpaceRect.x) || (xb > mAvailSpaceRect.XMost())) {

Just use &&

Comment 8

[Eli Friedman]

(In reply to comment #7)
> This might fix the underlying problem in bug 386182.

That can't be the right bug number...

Comment 9

[Eli Friedman]

Checked in.