On Fri, 2006-12-29 at 00:49 +0100, Josselin Mouette wrote: > Le jeudi 28 décembre 2006 à 14:47 -0800, Thomas Bushnell BSG a écrit : > > Package: glib2.0 > > Version: 2.12.5-3 > > > > This version of glib (both 2.12.5-3 and 2.12.6-1) causes an important > > regression in gnucash, and therefore should not go into testing. > > > > See http://bugs.debian.org/404585. > > What if you told us which characters are needed by gnucash instead of > whining and asking important bug fixes to be blocked outside etch? I don't think I asked for important bug fixes to be blocked outside etch. On the other hand, the upload of 2.12.5 did not fix any bugs, according to the changelog and the BTS. I don't know what the additional characters are or syntax which is needed. I could investigate, except: 1) The release team has asked us not to upload changes which are not destined for etch, and making gnucash work with the glib in unstable is therefore a low priority; 2) I shouldn't need to spend energy asking you to please work on getting etch out, instead of making uploads that won't get in etch anyway; 3) The change altered the syntax of the file by adding restrictions. It is therefore a non-backwards-compatible change to the ABI, and therefore it needs an so-name bump. No amount of adding this or that character will change that, unless you simply restore the previous behavior. Number (3) is the most serious issue, as I understand the case. Thomas
Attachment:
signature.asc
Description: This is a digitally signed message part