On Thu, Mar 18, 2010 at 8:32 AM, Yves-Alexis Perez <cor...@debian.org> wrote: > On mer., 2010-03-17 at 17:26 +0100, Goswin von Brederlow wrote: >> On the other hand including free space information will be difficult to >> handle properly and possibly expose information the user considers >> private. > > But places where packages are supposed to write aren't that private. > Usually /home (but if people mount one fs per user there might be > privacy issues), /tmp (not really private), /var (either). Packages > install/upgrades will write in /usr and /etc and few other folders > (/boot..) which aren't really private either. > > Real issues will appear for stuff using “named” mount point in /srv > (or /home as said above). >> >> >> But if you do go ahead with this consider this: >> >> The bugreport gives the following argument: "I suspect that a lot of my >> packages would fail in mysterious ways if the root partition was full." >> How will his software behave with the root partition being mounted >> read-only? Listing the free space on root or /usr would be completly >> meaningless if they are read-only and writing to them would be a >> critical bug anyway. > > One thing which could be done too is to run the df, detect if there are > full fs, and if there are, ask for the user if she wants to include the > result.
Add also quota information if available -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/195c7a901003180212p56c3616aj72306e8d62750...@mail.gmail.com
