On Wed, Jul 18, 2001 at 04:37:23PM -0400, Glenn F. Maynard wrote:
> > It is possible to edit queue with feeder - just scan feeder chain to the
> > end and first feeder contains saved_buf of the queue before any
> > feeder was added. cmd_queue should also be modified to append commands
> > to the saved_buf if the queue has a feeder.
>
> Er, took me a little while to parse this. Taste of my own medicine, I
> think. :)
>
> I think I understand what you're saying to do, but could you give me a way
> to test this? I'm not sure how feeders ever get stacked, so I don't know how
> to set up tests for this. I can get a feeder onto the queue (so I see the
> fed commands in 'jobs') with 'queue source file', but I can't get a queued
> feeder with manually queued jobs ahead of it. "queue sle forever ; queue
> echo command; queue source file" just queues a literal "source file" until
> it actually gets there.
>
> Alright, part two would fix the "queuing commands interactively while a feeder
> is queued", right? (Just stared at the saved_buf stuff for a while.) I
> suppose just implementing that would give me a way to test #1. Except, that'd
> seem to be a completely different code path just for regular queuing, which is
> a bit ugly; there's no Feed* commands for saved_buf, so that'd all have to
> be copied over, and we can't know the cwd/lcwd, so that failsafe would be
> lost.
Ok, nevermind. I don't think it has to be fixed. Most people don't do
`queue source file'.
> Tangent:
>
> I wish there was a specific feeder for queues--keeping a simple list of
> commands, with associated cwd and lcwds for each command, so cds and lcds
> could be fed on the fly, queues could be reordered without worrying about
> that, and queue editing would be reordering a list, not string manip. The
> fact that queues are a string is why I did 'queue edit' instead of 'delete
> queue job', 'move queue job', and so on. (I'm going to keep this idea in
> mind; it may be unfeasible to implement, but I may attempt it just to see how
> well it works. I'd wouldn't expect you to integrate that large a change,
> of course, but if I implement it, I'll let you know how it turns out.)
>
> Again, updated patch with the above, other than the feeder stuff.
Thanks! I have applied it with some FIXME notes.
--
Alexander.
