Am 13.05.2015 um 19:11 schrieb Martin Pitt:
Lennart Poettering [2015-05-13 17:55 +0200]:On Wed, 13.05.15 17:01, Martin Pitt (martin.p...@ubuntu.com) wrote:So, obviously we need to fix the crash; but I was wondering what the desired behaviour should be? In the sense of "be liberal what you accept" I think the extra space(s) should just be ignored; or should that count as an error and the unit get rejected?Neither. It should be considered an error, logged about, but the line should be ignored and we should continue. This is how we usually do it so far, to ensure unit files stay relatively portable between version, but on the other hand we aren't too liberal with accepting any data.You mean ignoring this single line, but still starting the unit (with any other Exec*=)? That feels quite odd to me, TBH -- it feels more robust if a unit is either completely valid, or completely inert?
no it is not - where do you draw the lineif a unit contains options for systemd-216 and completly valid on F21 but the same src.rpm is used for F20 would you like to fail the service or just have the log noise of the unknown option?
signature.asc
Description: OpenPGP digital signature
_______________________________________________ systemd-devel mailing list systemd-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/systemd-devel
