On 3/14/07, Scott Horowitz <[EMAIL PROTECTED]> wrote: > On 3/14/07, bardo <[EMAIL PROTECTED]> wrote: > > 2007/3/14, Alessio 'mOLOk' Bolognino <[EMAIL PROTECTED]>: > > > This happened to me too days ago when I waited much time before answering > > > the same question. > > > > I experienced it two times and I have a gut feeling... do you use alunn? > > Indeed, it happens if you wait a while before pressing entering. I ran > it with debug, but there isn't much information. Here is the last of > the debug output, I can paste the whole thing if you want too... > > > debug: check_freespace: total pkg size: 0, disk space: 4996853760 > > Targets: glib2-2.12.11-1 atk-1.18.0-2 git-1.5.0.3-1 > gstreamer0.10-0.10.12-1 pango-1.16.1-2 gtk2-2.10.11-2 > kernel26-2.6.20.2-1 libtasn1-0.3.9-1 > > Total Package Size: 32.47 MB > > Proceed with installation? [Y/n] > :: Retrieving packages from current... > debug: using 'glib2' for download progresslooking up mirror.cs.vt.edu > connecting to mirror.cs.vt.edu:21 > [EMAIL PROTECTED] ~]$
Gotcha. Thanks for this. It seems the problem is related to libdownload. Waiting at the prompt gives the socket enough time to timeout. It then goes through re-opening the connection, BUT, actually writes to the old connection with some closing stuff. This fires off a SIGPIPE as the socket itself is actually broken at this point. So it's an uncaught SIGPIPE which is causing this. I will make sure that this doesn't happen. _______________________________________________ arch mailing list arch@archlinux.org http://www.archlinux.org/mailman/listinfo/arch
