# Bug in linebreaking algorithm #523

Closed
opened this issue Jul 23, 2013 · 3 comments

Projects
None yet
3 participants
Member

### pkra commented Jul 23, 2013

 Here's a sample from OSA. I couldn't reduce it much further, I'm afraid.  1 2 ( | R R R ± | L L L ) ( | + | ) QD 6 1 2 ( | R R R | L L L ) ( | | ) QD 6 , 1 2 ( | R R R | L L L ) ( | + | ) QD 7 1 2 ( | R R R ± | L L L ) ( | | ) QD 7 , 1 2 ( | L R R ± | R L L ) ( | + | ) QD 6 1 2 ( | L R R ± | R L L ) ( | + | ) QD 6 , 1 2 ( | L R R ± | R L L ) ( | + | ) QD 7 1 2 ( | L R R ± | R L L ) ( | + | ) QD 7 . [/itex] 

### dpvc pushed a commit to dpvc/MathJax that referenced this issue Oct 28, 2013

 Don't let nesting count go below zero. Resolves issue mathjax#523. 
 cd0b88e 
Member

### dpvc commented Oct 28, 2013 • edited

 This turns out to be due to MathJax's attempts to determine the nesting level of "flat" MathML (i.e., MathML that doesn't use  to indicate nesting) by using open and close delimiters. When these were unmatched, as in this case (due to the right-angle-brackets), the nesting count was getting confused. The issue523 branch of my fork of MathJax includes a fix for this.

### dpvc pushed a commit to dpvc/MathJax that referenced this issue Oct 28, 2013

 More checks to make sure nesting level doesn't go below zero. Issue m… 
…athjax#523.
 08ec41c 
Contributor

### fred-wang commented Oct 30, 2013

 MathMLToDisplay/issue523.html => Ready for release, in testsuite

### dpvc pushed a commit that referenced this issue Oct 30, 2013

 Merge pull request #644 from dpvc/issue523 
Resolves issue #523
 1f013d7 
Member

### dpvc commented Oct 30, 2013

 => Merged