[39] in bug-owl
More feature requests for 1.2
daemon@ATHENA.MIT.EDU (Mitchell E Berger)
Thu Jun 27 15:00:54 2002
Message-Id: <200206271900.PAA28849@athenaphobia.mit.edu>
To: bug-owl@MIT.EDU
Date: Thu, 27 Jun 2002 15:00:33 -0400
From: Mitchell E Berger <mitchb@MIT.EDU>
- If possible, allow pexec to run interactive commands. I don't have any
idea how hard this is; I know I don't know how to do it.
- Allow [pa]?exec to run pipes, like "pexec ps -ef|grep init".
- An undo-last-delete command, default bound to M-U perhaps, that unmarks
the last messages marked for deletion. This would allow people to recover
from an accidental "T" typed when they thought they were typing a zephyr
but weren't. I'll try implementing this when a 1.2 snapshot comes out.
- It seems like $owl::fields[1] on a zephyr is only the first line of the
message, or at least that's what "perl $owl::fields[1]" shows. I need to
adapt my owl::change_curmsg patch for 1.1.3 to check if this is really a
problem or was fixed, but it may be a bug.
- Just so it gets written down if it hasn't yet, we need something like zwgc's
minttl to be able to support something like the followlast option, lest
people delete a message just as another one comes in.
- jdaniel has mentioned wanting the current pointer to point at n+1, where n
is the number of the last message in a view, if he deletes that message. I
think that's doable without interfering with followlast... it might cause
a problem with scrolling (or failing to) if we don't recalculate the top
message, but I'll check and try to get it to play nice with followlast when
a 1.2 snapshot comes out.
- The option infrastructure sort of automatically generates part of the help
message, so when a new option is implemented, no extra work has to be done
to have it be listed in the help page. I don't know if nygren's keybinding
work is going to set up a similar infrastructure for keybindings automatically
fitting in the help page, but it would be really cool if they did; this would
both allow people who rebind to get a quick recap of what's currently set,
and would prevent bugs like Vadim's from coming up.
- Accessing fields from the current message, and a perl variable for the name
of the current view, which jdaniel requested, are actually both done by the
patches I sent in to allow a perl .owlconf function to run when the current
message or view is changed, ftr.
- It would be nice if the developers could see snapshots in between releases
more often, or ideally even use cvs. Then we'd be able to try out other
patches that have been taken and work relative to them instead of waiting
for the next release or copying code out-of-band.
- Tibbetts mentioned that we could get a sipb-rt queue if we want, and there
was much enthusiasm on -c sipb earlier. We need a way to coordinate work so
we know what other people are doing and don't step on them, and this seems
like a potentially very good solution.
Mitch
