On 02/22/2011 11:08 PM, Phillip Lougher wrote: > Where is the output of mksquashfs going? To a pipe, and thence to a file. But I suspect that I'm flushing it frequently, which may contribute greatly to the problem. I'll have to look into that. > The -noappend option changing things is interesting but distinctly > perplexing. If the Transparent Archivist deleted the output Squashfs > file before running Mksquashfs again then -noappend shouldn't make any > difference! Well, having taken an admittedly cursory glance at ta.cc, it appears to me that TA does delete the squashfs file before generating the next iteration of it (via the checkedUnlink() function). So, I agree with you, it would be perplexing if -noappend fixed the problem. I still don't know whether it does or not.
Alas, I have only one machine with enough disk space (available at a reasonable speed -- USB 2.0 disks simply aren't fast enough for this) to make these experiments. I set up another machine with 4.1, but its disks are too congested -- with unsaved archives! -- to do the experiment. "There's a hole in the bucket, dear Charlie, dear Charlie..." > This behaviour is entirely defensive and arguably the correct > behaviour - the alternative behaviour is to have Mksquashfs always > delete the output file if it exists (i.e. -noappend is default). > Unfortunately such an default action leads to all sorts of user > mistakes, where they specify the wrong file, and get their important > archived data deleted ... > I can't argue with your logic, here. It makes perfect sense to me. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/722168 Title: mksquashfs hangs -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
