Issue 1424171: patch for etree cdata and attr quoting (original) (raw)

Created on 2006-02-04 18:23 by chrism, last changed 2022-04-11 14:56 by admin. This issue is now closed.

Messages (6)
msg27438 - (view)	Author: Chris McDonough (chrism)	Date: 2006-02-04 18:23
Attached is a patch for ElementTree (based on a checkout from the SVN trunk's xmlcore.etree) that seems to perform better escaping of cdata and attribute values. Instead of replacing, for example ""e;" with ""e;" or "&" with "&", it tries to avoid requoting ampersands in things that look like entities. Sorry, I haven't tested this with anything except Python 2.4, I'm not quite sure what to do about _encode_entity, and I haven't patched any tests or written a new one for this change. Consider this more of a RFC than a patch ready-for-submission as a result.
msg27439 - (view)	Author: Chris McDonough (chrism)	Date: 2006-02-04 18:26
Logged In: YES user_id=32974 Sorry, the tracker doesn't seem to want to allow me to upload the file. See http://www.plope.com/static/misc/betterescape.patch for the patch.
msg27440 - (view)	Author: Georg Brandl (georg.brandl) *	Date: 2006-02-04 18:29
Logged In: YES user_id=1188172 OP: You did check the box?
msg27441 - (view)	Author: Chris McDonough (chrism)	Date: 2006-02-04 20:23
Logged In: YES user_id=32974 Egads, I did this time.
msg27442 - (view)	Author: Fredrik Lundh (effbot) *	Date: 2006-02-06 21:55
Logged In: YES user_id=38376 I'm not sure I follow. ET works on the infoset side of things, where everything is decoded into Unicode strings (or compatible ASCII strings). If you set an attribute to "&" in the infoset, it *must* be encoded on the way out. If you want an ampersand, use "&".
msg27443 - (view)	Author: Chris McDonough (chrism)	Date: 2006-02-08 17:00
Logged In: YES user_id=32974 Doh, of course. Sorry, I suspect I'll need to go to remedial XML class. Ignore this.