On Wed, Mar 15, 2006 at 11:53:15PM +0100, Jordi Mallach wrote: > On Wed, Mar 15, 2006 at 04:57:54PM -0500, Justin Pryzby wrote: > > Have all of Thomas's concerns been addressed by this upload? > > Just the segfault, which was the main topic of the bug. I'm sorry I > didn't ask thomas to file bugs as necessary for the different issues; > that'll be easier to track. > > 1) I claimed to be fixed. > 2) isn't, AFAIK, but now its a lot less noticable, of course. > 3) is pretty closely related to 2), isn't it? gpg has a similar problem report, for both of 2,3 I guess. If the PID is stored in the lockfile, and the pid isn't running (or is otherwise not a process which [cs]ould have created the lockfile, or isn't holding it, etc.), then the lock should be silently dropped, I think. This is (3); the robust and convenient thing to do. 2) is just a preventative, whereas (3) would even deal with a user creating the lockfile accidentally, somehow :)
> 4) is a nice wishlist > 5) is totally correct This is my understanding also; I don't know if 4) is really necessary; it seems like it would be much more useful to have an explanatory message "The lockfile \`$s' exists, so $0 will not run ...". Justin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
