[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#402692: Status of KeyFile regressions in glib2.0



On Sun, 2006-12-31 at 21:10 +0100, Loïc Minier wrote:
>  Finally, since glib's upstream added support for key names with MIME
>  types chars such as "/" and "+" for gnome-vfs2, I also added a patch to
>  silently support key names with spaces in the middle of the name (not
>  at the beginning or at the end) as I think this might be relatively
>  common.  (Attached as well.)

If such a patch is in, and can be verified to work by the people who
reported the bug against gnucash, then I have no objections (on this
score) to the migration of glib into testing.  Can you let me know when
an upload is made so that it can be conveniently tested?

The upstream bug in bugzilla is looking like upstream is not going to
allow spaces in keynames, though this may change.  I understand that
gnucash upstream plans to talk directly to them.

>  It would be best if Gnucash switched as soon as possible to stricter
>  key names.

gnucash upstream is working on the problem.  However, a serious
deficiency with this particular change was that it did not make any
sensible *conversion* possible.  When new strictness rules are created
out of whole cloth, it is crucial to provide for transition, and not
just decree a new rule.  It appears that glib is very unlikely to do
anything to help with the transition, which creates major extra work.
If you could impress this upon the upstream maintainers, it would be
great.
> 
>  The first patch will probably be reverted in the lenny cycle, and the
>  second patch might be kept if upstream merges it.

If the patch to support spaces gets reverted, can we please at least
coordinate that with gnucash?  (That is, please don't just revert it
when the mood strikes; check and see whether the upstream gnucash fix is
in place in Debian unstable first.)

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: