Issue 2732: XML-illegal characters in darcs changes are not escaped - Darcs bug tracker

Title	XML-illegal characters in darcs changes are not escaped
Priority	bug	Status	unknown
Milestone		Resolved in
Superseder		Nosy List	bortzmeyer
Assigned To		Topics

Created on 2024-09-02.11:28:16 by bortzmeyer, last changed 2024-11-29.12:08:51 by bfrk.

Files
File name	Uploaded	Type	Edit	Remove
resolve-issue2732_-generate-xml-using-the-xml-package.dpatch	bfrk, 2024-11-11.20:33:18	application/darcs-patch

Messages
msg24079 (view)	Author: bortzmeyer	Date: 2024-09-02.11:28:11
Testing the tool darcs-to-git, it crashes on one repository because: Illegal character "\b" in raw string "Toujours en DISCUSS a l'IEG\bSG" This is apparently because it relies on darcs changes --xml and this one does not escape the characters like \b when they are in the commit messages, therefore proudcing illegal XML.
msg24094 (view)	Author: bfrk	Date: 2024-11-06.08:42:45
I have attached a patch bundle, to be applied to current screened (`darcs clone https:/darcs.net/screened`). This should fix your problem. Could you please try and see if that works better? If it does, I am willing to make a new bug-fix release (2.18.5). Attachments lower-the-upper-bound-for-http_client_tls.dpatch
msg24097 (view)	Author: bortzmeyer	Date: 2024-11-11.16:47:58
Sorry, it did not help. Testing the output of 'darcs changes --xml' with xmllint: /tmp/ch.xml:50015: parser error : xmlParseCharRef: invalid xmlChar value 56548 <name>P&#56548;s approuve</name> Indeed, the commit messages were in Latin-1 at the beginning of the repository and are in UTF-8 now. All commit messages with non-ASCII Latin-1 characters experience the above problem. Suggestion: invalid UTF-8 sequences could be replaced by ??? or something similar. At least, the XML would be legal.
msg24100 (view)	Author: bfrk	Date: 2024-11-11.20:33:17
I agree. The patch I made did not yet do the correct conversion for all patch metadata. Please try with newly attached bundle instead. Note: you should obliterate the two patches from previous bundle and also pull. Attachments resolve-issue2732_-generate-xml-using-the-xml-package.dpatch
msg24101 (view)	Author: bfrk	Date: 2024-11-11.20:36:09
Or apply to a fresh clone.
msg24108 (view)	Author: bfrk	Date: 2024-11-29.12:08:50
Could you please try if the attached bundle fixes your issue?

History
Date	User	Action	Args
2024-09-02 11:28:16	bortzmeyer	create
2024-11-06 08:42:47	bfrk	set	files: + lower-the-upper-bound-for-http_client_tls.dpatch messages: + msg24094
2024-11-11 16:47:59	bortzmeyer	set	messages: + msg24097
2024-11-11 20:33:19	bfrk	set	files: + resolve-issue2732_-generate-xml-using-the-xml-package.dpatch messages: + msg24100
2024-11-11 20:34:38	bfrk	set	files: - lower-the-upper-bound-for-http_client_tls.dpatch
2024-11-11 20:36:09	bfrk	set	messages: + msg24101
2024-11-29 12:08:51	bfrk	set	messages: + msg24108