On Thu, 21 Mar 2013 04:03:32 -0400 Daniel Dickinson <[email protected]> wrote:
> > I poked a little around in the bash and cat source code, and this is > > what they are doing here: > > > > - bash writes the text in the heredoc to a temporary file, unlinks > > that file (so that nobody can mess around it) and sets standard > > input for cat to a file descriptor pointing to that file. > > > > - cat runs fstat(2) on that file descriptor; it does so for both > > input and output in order to detect whether they are the same. > > However, fstat() fails with ENOENT and cat errors out. > > > > Why would fstat() ever return ENOENT? This makes no sense to me. > > My current thinking is that it's got something to do with the > disappearing emacs23 directory, based on a similar error with a > directory removal while a program was running (the directory is still > present for some contexts but not others, until all file descriptors > are released). > The other possibility I am considering is 9p specific...The issue could be the permissions on the underlying os due to the way qemu-kvm creates them are such that when the directories/files get created that the vm doesn't have sufficient access for the required operations. I'll try to find time to verify this theory. Regards, Daniel -- <erno> hm. I've lost a machine.. literally _lost_. it responds to ping, it works completely, I just can't figure out where in my apartment it is. -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
