santuario-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Raul Benito <>
Subject java xml-sec 1.3 release plan
Date Tue, 26 Jul 2005 21:46:07 GMT
I am planning to do the 1,3 release of the java xml security. The
changelog so far of the release is stated below:

But I want to concentrate in bugfixes for the time is left till the release.

So if anyhow has any bug that isn't fixed or any one not new
reported... You know send an email(or write a bugzilla entry)....

I am going to do a release candidate jar to facilitate the testing.

Anyhow You can begin to send bugs(and better patches)



Changelog for "Apache xml-security" <>
New in v1.2.1-99
	Init-Don't fail if a transformation don't have all of its dependecies.
		Remove XPath initialization from Init and do only when xpath is needed.
    Resolv-Removed the use of xpath expressions to search the elements
to sign/verify, now use only plain DOM searching.
	Resolvers-Remove wantsOctectStream wantsNodeSet and his returns pair
they are not used, right now and some are incorrect.
	Remove the Use of xalan or xerces class URI
	Removed the expandSystemId
	Changed from Vector<String> to List<Class>, so we don't need to use
classForName everytime and used it just the first time.
	Removed PRNG,HexDump,Version, X509CertificateValidator
	Added an unsync buffer outputstream.
	Changed Symbol table to a more efficient and simple structure
	Fixed bug 34743 , Submitted by: Lee Coomber <>
	Minor speedups in b64, Halved the table lookups.
	Reduce Object creation during c14n, from one to level to one per c14n.
	Change all Vector to List(ArrayList), we don't need synchronization safety.
	*Refactor the way we handle c14n of nodesets:
	    Before this patch every transformation creates a set with the
nodes that should
	    be outputed. Every set is obtaining visiting the whole dom tree every time,
	    and then do it other time at c14n time. So it does <number of
	    visitings, very slow and memory costly.
	    Now every transformation just return a NodeFilter that tells if
the node is included or not.
	    So only one visiting is done.
	Unified and transformation
	Canonicalization tree travesing is not recursive. it gives better
memory handling and performance.


View raw message