Daniel Veillard wrote:
It always amazes me how monstruous bugs like this manage to fell through the cracks and stay unnoticed for so long, impressive ! Can be trivially reproduced with xmllint --push ... If you have found other problems like that, please post them :-)
You kidding, rather have the bugs on someone elses list than mine :)
That change also works for me. Wasnt exactly sure about that so had just pulled it from the XML_PARSER_END_TAG case.Didn't work fine for me, I made a slightly different (and I hope correct) using nsPop(ctxt, ctxt->nsNr - nsNr) . I also added the test case to the regression tests,
Rob