On Fri, Jan 04, 2008 at 04:30:42PM +0100, Josselin Mouette wrote: > reassign 452162 gnucash > thanks > > Whoops, sorry for replying without reading this additional information. > > Le dimanche 25 novembre 2007 à 12:14 -0800, Andrew Sackville-West a > écrit : > > I went back to step zero. This problem showed up after a segfault in > > gnucash. This segfault occurred while printing checks (uses > > gtkprint). Subsequent instances of gnucash could no longer print > > checks. Since I had two machines fully-up-to-date. I decided to try a > > different user on the broken machine. lo and behold, no problem. :( or > > :) depending on your perspective. Anyway, moving ~/.gconf out of the > > way solved the problem. > > > > Clearly, the segfault corrupted something in ~/.gconf/* resulting in > > this behavior. There is probably no way to track down the segfault as > > it happened in an instance of gnucash that may have been up for many > > days and may have been up across a variety of upgrades... > > > > I have a copy of the problematic ~/.gconf available if you would > > like. > > This definitely looks like a bug in gnucash, and the copy of the > problematic gconf keys will certainly be useful to the gnucash > maintainer.
I have that available. Well, I have the whole stinking tree available ;) if anyone wants it. The more I think about it, though, and the more I work on gnucash, I suspect this is a fleeting, one-time thing. I strongly suspect that instance of gnucash had been live across at least a couple significant updates in sid, which brought it to its knees. I don't think any app could be expected to gracefully recover from that kind of systemic change. It was an unfortunate side-effect of the crash that corrupted some gconf keys. This should probably be closed as unreproducible, or needinfo, or whatever. Oh and thanks for finally getting to this. I know it was cryptic at best. A
signature.asc
Description: Digital signature
