Hi. On Jul 08 2015, jean-marc montanier wrote: > Just a quick message to report that the nasty workaround is anyway > efficient. It's working on my Jessie, thanks.
The basic idea behind using at is not that bad at all. I have investigated (a very tiny bit) of the behavior of fuse mounts and the thing is that the fuse filesystems stay with their respective fuse programs running until the filesystems are unmounted, which, apparently, leads udev to think that they are stuck (which is not necessarily the case). With the current workaround with at, we "offload" the task of a long running process to at and udev then sees the whole procedure being "completed" in a short time, and does not complain or kill the subprocess. What I believe is a clener solution is that, instead of using at, we can simply use a program that "daemonizes" the processes that we want. Perhaps a very thin wrapper written in C around the daemon(2) function would be sufficient for our needs. Or, perhaps, we can stick with at, since it is a proven tool that stood the test of the time (and the usbmount package would still continue to be arch: all). > If udisks2 is covering the functionalities of usbmount and has an active > development, may be the usbmount package should be removed from the future > versions of debian ? I *think* that udisks2 *may* cover the functionalities of usbmount, but I had a very bad time telling udisks2 that I want my NTFS filesystems mounted with the options that I want (in particular, noatime, nodiratime, and big_writes), but, apparently from what I read on blogs and other sites, those *seem* to be hardcoded behavior, which is very unpleasant. Nevertheless, I am thinking of moving the development of usbmount to my github account and accepting all the pull requests there. It is much easier (read: I can react much more quickly) to me to accept pull requests there than receiving patches via e-mail and applying them (of course, I will still be willing to receive patches by e-mail, but that's not my preferred option). I will keep the git repository in Debian mirrored with the github repo, so that if github goes away (which I sincerely doubt), then we don't lose much. How does that sound? BTW, one very important question: are the problems with usbmount only apparent when using fuse filesystems or are they also apparent when using filesystems supported by the kernel (e.g., vfat or ext4)? Regards, -- Rogério Brito : rbrito@{ime.usp.br,gmail.com} : GPG key 4096R/BCFCAAAA http://cynic.cc/blog/ : github.com/rbrito : profiles.google.com/rbrito DebianQA: http://qa.debian.org/developer.php?login=rbrito%40ime.usp.br -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org