Re: [xml] SAX HTML still stuck.



At 04:36 PM 10/04/01 -0400, Daniel Veillard wrote:

 It's clear that there is yet another subtle bug in the push code,
check first that 

  if (!abort)
      htmlParseChunk(ctxt, buf, 0, 1);
is a workaround to your problem.

Yes, that makes it not hang, but I think that's just avoiding the code that
hangs.  I can use that when I know I'm aborting, but in some cases the
input stream may be truncated.

That is, if I'm parsing a file that's truncated right after the <title> (or
somewhere near, as I don't know what the parser does after calling my
endElement() callback) then we may see the same behavior.

I posted on Sept 29 a very similar problem ([xml] htmlParseChunk loop)
where I was aborting with I found a <meta name="robots" content="noindex>
-- but the odd think, IIRC, was that it only got stuck with a chunk size of
exactly 4096.  Odd.

Bill Brack sent a patch on Sept. 22nd for another (similar) loop I was
having.  I'm not sure if you ever saw that.

I wish I could offer more debugging help.

Thanks,


Bill Moseley
mailto:moseley hank org




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]