Look at the testcase in the URL. I found this bug when fixing bug 44564. I am guessing the problem is that the parser sees an internal subset begin, and starts collecting the data in there. It then finds the comment, and sends it to the content sink. Then it finds the internal subset close, and sends the internal subset to content sink. So it looks like we would need to suppress the processing of comments while we are in the internal subset.
18 years ago
Status: NEW → ASSIGNED
Target Milestone: --- → mozilla0.9
This seems to be simple. In nsExpatTokenizer::HandleComment() method I check the parser state and if we are in the doctype just append the comment into the doctype text, and do not create a comment token. The patch also fixes some static global variables (to use the NS_GET_IID macro). While doing this I noticed nsITokenizer interface does not have the static GetIID() method so I added that as well.
18 years ago
Target Milestone: mozilla0.9 → mozilla0.8
sr=vidur. I'm assuming that the comment text passed to us by expat doesn't already include the delimiters.
Fix checked in. Oh, and by the way, this was r=harishd.
Status: ASSIGNED → RESOLVED
Last Resolved: 18 years ago
Resolution: --- → FIXED
Verified fixed per last comments.
Status: RESOLVED → VERIFIED
You need to log in before you can comment on or make changes to this bug.