> Unfortunately my stated intention to look at imon in "2-4 weeks time"
> isn't looking good at present :-(
Well, it can't be any worse than my stated intention to look at imon
"in the next day or two." (I did start fiddling with the 2.3.42 patch
in a 2.3.99-pre3 tree, and promptly broke it.)
> Am I right in my understanding that this facility tells interested
> userland applications two main things:
>
> 1: That an executable is currently being executed and
>
> 2: That an inode has been altered?
>
> If so, I'm wondering how much value number 1 provides. For desktop file
> browsers, number 2 is more important?
Yeah, the start/stop exec stuff isn't nearly as useful. You get notified
when the number of running instances of a program changes from 0 to 1 or
back again, or when at least one instance is running when you start
monitoring it. I think all it's used for in IRIX is changing desktop
icons to indicate whether a program is running, which is about all you
*can* use it for, because you don't know who's running the program, or
with what arguments; you don't receive notification when additional
instances start up, etc.
> If I'm not missing something, have you considered splitting off feature
> 1 and submitting only the inode-mod feature? If this reduces the
> intrusiveness of the kernel patch this may ease its path into the
> mainstream release.
That's a good idea, and no, I hadn't considered it. The start/stop-exec
notification certainly isn't essential, and I don't think there's a way
to contain it in the imon module, which probably does make it an
impediment to the eventual acceptance of the rest of the imon stuff.
--Rusty
--
Source code, list archive, and docs: http://oss.sgi.com/projects/fam/
To unsubscribe: echo unsubscribe fam | mail majordomo@xxxxxxxxxxx
|