Quoting Tim Dijkstra (tdykstra) ([EMAIL PROTECTED]): > Package: geneweb > Version: 4.10-6 > Severity: grave > > > The prerm of geneweb finds files that it thinks are geneweb databases > located everywhere on the system and starts altering them (updating, > moving, etc).
updating: no moving : no (unless the script is buggy) This script tries to find *.gwb files and export them to *.gw files. This is mostly because, in the past, pre-exporting database files was the only way to preserve databases when upgrades changed the database structure. Moreover, the files had to be exported with the *old* gwu utility, hence the tentative to find all databases around. > All this unasked for. A maintainer script has no business messing around > with peoples data! It does not. It *adds* a xxx.update.gw file along with the found xxx.gwb file. It tries to give this file the same owner/group than the former file, and chmod it to preserve privacy. > > I can see the point in updating the databases in the dir owned by > geneweb (/var/lib/geneweb), but messing around random files on the > filesystem is not something a maintainer script should do. > I was stupid enough to have a drive mounted r/w with backups on them, which > now are also nicely updated... geneweb does not change *.gwb files So, I understand you may object to this piece of code because, yes, it deals with users files and is likely to write on places where user data are. So, for that reason, I may consider commenting the relevant code because it may be easily interpreted as non policy compliant. (and indeed the bug severity should then be "serious"...the package does not lose data unless you prove me wrong) But I would do so only for that reason....not because geneweb changes user data which it does not. Of course, tagging the bug grave while we are so close of a release is a way to force me taking the only measure I can safely take here : comment out the code which EXPORTS user data. Because this is the only safe solution I have, indeed.
