Carsten Pfeiffer wrote: > On Saturday 07 October 2006 22:41, you wrote: > > >>> I want to make sure that Carsten's problem will be solved with the >>> latest CVS code. It would be great if you could check the latest package. >>> >> I sent him the link, but I see no reason why it wouldn't, since he's the >> one who wrote the patch, and it is included in the CVS release. >> > > Yes, the patch I wrote definitely fixes the problem for me. > I just tested the tarball from dlgeek.net (I'm on i386 so cannot test > the .deb). It works fine, although it seems to be a little bit more prone to > filename encoding problems: > > #0 ucs2len (string=0x0) at ucs2char.c:71 > #1 0x080591a5 in ucs2dup (src=0x0) at ucs2char.c:190 > #2 0x08053e53 in h10db_set_title (h10db=0x8064310, index=56, value=0x0) at > h10db.c:856 > #3 0x0805a98c in gettag_set_info (h10db=0x8064310, index=56, > info=0xb45e4e08) > at getmediainfo.c:136 > #4 0x0804df70 in easyh10_database (path_to_root=0xbfa047e4, > path_to_db=0xbfa007e0, path_to_music=0xbf9fe7de, model_filename=0xbf9fa7da, > info_extractor=0x0, flags=67174400, instance=0xbfa097fc, > progress_proc=0x804a9b0 <database_progress>, error_proc=0x804adc0 > <easyh10_error>) > at ../common/easyh10_database.c:737 > #5 0x0804bfab in main (argc=6, argv=0xbfa098b4) at main.c:1108 > > The filename was kinda broken, at least it wasn't latin1, as I specified > with --encoding=ISO-8859-1 > > I don't know why, but the previous version didn't crash on this. > > Cheers, > Carsten > The tarball is just a checkout from upstream CVS. It's fairly trivial to build a source package, but I'll upload an x86 package as soon as my pbuilder environment updates. I don't know why the regression would have happened, talk to Nyaochi.
Nyaochi, Geoff: How did the remote debug go? Benjamin
signature.asc
Description: OpenPGP digital signature
