GNOME Bugzilla – Bug 128620
scrollkeeper complains on malformed XML
Last modified: 2004-12-22 21:47:04 UTC
making a "scrollkeeper-rebuilddb", it complained : /usr/share/gnome/help/rhythmbox/C/legal.xml:1: parser error : Missing encoding in text declaration <?xml version="1.0" ?> ^ /usr/share/gnome/help/rhythmbox/C/rhythmbox.xml:72: error: Failure to process entity legal &legal; ^ /usr/share/gnome/help/rhythmbox/C/rhythmbox.xml:72: parser error : Entity 'legal' not defined &legal; ^ /usr/share/gnome/help/rhythmbox/C/authors.xml:1: parser error : Missing encoding in text declaration <?xml version="1.0" ?> ^ /usr/share/gnome/help/rhythmbox/C/rhythmbox.xml:213: error: Failure to process entity authors &authors; ^ /usr/share/gnome/help/rhythmbox/C/rhythmbox.xml:213: parser error : Entity 'authors' not defined &authors; ^ /usr/share/gnome/help/rhythmbox/C/user-guide.xml:1: parser error : Missing encoding in text declaration <?xml version="1.0" ?> ^ /usr/share/gnome/help/rhythmbox/C/rhythmbox.xml:215: error: Failure to process entity user-guide &user-guide; ^ /usr/share/gnome/help/rhythmbox/C/rhythmbox.xml:215: parser error : Entity 'user-guide' not defined &user-guide; ^ /usr/share/gnome/help/rhythmbox/C/quick-reference.xml:1: parser error : Missing encoding in text declaration <?xml version="1.0" ?> ^ /usr/share/gnome/help/rhythmbox/C/rhythmbox.xml:216: error: Failure to process entity quick &quick; ^ /usr/share/gnome/help/rhythmbox/C/rhythmbox.xml:216: parser error : Entity 'quick' not defined &quick; ^ Document is not well-formed XML: /usr/share/gnome/help/rhythmbox/C/rhythmbox.xml hopefully, rhythmbox isn't the only bad boy for which scrollkeeper complained. others were gnome-applets-2.4.1 and anjuta-1.1.98 gentoo x86 system, pretty up to date i'm open to any question / testing bye
change : anjuta is no longer a bad boy (1.2.0 fixed this). only rhythmbox-0.6.1 and gnome-applets-2.4.1 anyway, it's not the most anoying problem i've encountered in my life. better concentrate on gnome 2.5 / gstreamer 0.7
This patch should fix, please apply. (Debian bugs #223538 and #220747).
Created attachment 22288 [details] [review] Docs patch
the patch fixes everything but "authors.xml" adding UTF-8 encoding to this file fixes the last bit
Woops. Silly autogenerated file. :) This patch is better then.
Created attachment 22290 [details] [review] Docs patch, take #2
I'm confused by this. The XML standard explicity says that the encoding= field is optional. It then later very strongly implies that if it is not supplied, implementations should assume UTF-8. http://www.w3.org/TR/REC-xml#NT-XMLDecl and: http://www.w3.org/TR/REC-xml#charencoding So it appears to me that this is a bug in Scrollkeeper. I'll leave this as NEW pending your reply...
Should you disagree with my assessment, please re-add the PATCH keyword.
Hi walters, Right, this appears to be a bug somwhere else (scrollkeeper, libxml2, I don't know). The patch does workaround it, and doesn't harm at all. I suggest it is applied, the real bug will surely be fixed where it needs to be fixed. If you don't agree, feel free to close the bug (I just wrote the patch to stop the stream of Debian bugs coming in about the issue, and one of the GNOME maintainers already applied it).
scrollkeeper-rebuilddb doesn't give this error on my system. I'm using 0.3.12. Perhaps it's a bug introduced after that? By the way, the proper name is "UTF-8", although probably most implementations will accept "utf-8" too. So I fixed that and applied your patch. But could you forward this issue on to the scrollkeeper upstream? Thanks!