On Feb 3, 2004, at 5:56 PM, Neil Graham wrote:
Hi folks,
As folks who've been following this list for a while will have noticed,
there's at least one particularly nasty bug in 2.4.0 that's driving folks
crazy. That is, the parser sometimes reports attributes as being
duplicates when namespaces are disabled even when that's most certainly not
the case.
But there are some other rather severe problems with 2.4.0 as well:
- The parser will often read freed memory while looking for duplicate attributes; - There are a considerable number of bugs in the XML Schema API implementation. Most severe of these are a collection in the getCanonicalRepresentation code that sometimes causes it to overwrite random memory locations.
There have been myriad bugfixes applied recently as well, particularly to
improve thread safety, make sure the appropriate memory manager is used in
any context, and clean up compiler warnings.
So before we contemplate moving forward with development, perhaps it's a
good time to produce a bugfix release. Normally, one would like to
increment only the patch number and call this 2.4.1; but since so many
files have been touched since 2.4.0, it's unreasonable to assume we will
have binary compatibility on every platform. So that's why I'll propose to
call this 2.5.0.
I'll suggest that we should aim to publish this before the end of next
week, and in the mean time try and fix as many outstanding bugs (and commit
as many outstanding patches!) as possible. It would be wonderful if people
who have patches pending which have yet to be applied could retrofit their
patches to the current codebase, recognizing that it's unlikely any patches
made against the 2.4.0 source will apply cleanly now.
Here's my +1.
Cheers! Neil Neil Graham XML Parser Development IBM Toronto Lab Phone: 905-413-3519, T/L 969-3519 E-mail: [EMAIL PROTECTED]
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
