Re: [systemd-devel] [PATCH] path-util: Fix path_is_mount_point for files

2015-05-29 Thread Martin Pitt
Lennart Poettering [2015-05-29 11:33 +0200]: > > Exactly, that's why I called it "parent"; but I'm not fussed about the > > name, "dir" or "containing_dir" would work as well. I'd just not call > > it "filename" as that would be confusing -- this is *not* the file > > name of fd, but the directory

Re: [systemd-devel] [PATCH] path-util: Fix path_is_mount_point for files

2015-05-29 Thread Lennart Poettering
On Thu, 28.05.15 21:03, Martin Pitt (martin.p...@ubuntu.com) wrote: > Hello Lennart, > > Lennart Poettering [2015-05-28 19:44 +0200]: > > On Wed, 27.05.15 10:07, Martin Pitt (martin.p...@ubuntu.com) wrote: > > > > > -int fd_is_mount_point(int fd) { > > > +int fd_is_mount_point(int fd, const char

Re: [systemd-devel] [PATCH] path-util: Fix path_is_mount_point for files

2015-05-28 Thread Martin Pitt
Hello Lennart, Lennart Poettering [2015-05-28 19:44 +0200]: > On Wed, 27.05.15 10:07, Martin Pitt (martin.p...@ubuntu.com) wrote: > > > -int fd_is_mount_point(int fd) { > > +int fd_is_mount_point(int fd, const char *parent) { > > Hmm, now I am confused? Why "parent"? > > I really think this sho

Re: [systemd-devel] [PATCH] path-util: Fix path_is_mount_point for files

2015-05-28 Thread Lennart Poettering
On Wed, 27.05.15 10:07, Martin Pitt (martin.p...@ubuntu.com) wrote: > -int fd_is_mount_point(int fd) { > +int fd_is_mount_point(int fd, const char *parent) { Hmm, now I am confused? Why "parent"? I really think this should work as close as the usual *at() calls work. i.e. take a dir fd as first

Re: [systemd-devel] [PATCH] path-util: Fix path_is_mount_point for files

2015-05-27 Thread systemd github import bot
Patchset imported to github. Pull request: -- Generated by https://github.com/haraldh/mail2git ___ systemd-devel mailing list systemd-devel@

[systemd-devel] [PATCH] path-util: Fix path_is_mount_point for files

2015-05-27 Thread Martin Pitt
Martin Pitt [2015-05-25 21:06 +0200]: > in 220, path_is_mount_point() now always fails with "-20 ENOTDIR" when > calling it on files. This is problematic as it's perfectly valid to > have bind-mounted files; in fact, systemd's machine_id_setup() itself > creates a /run/machine-id → /etc/machine-id