Basic idea is that
(1) interactive commands which deal with 'changes' (i.e. record, amend, revert,
unrevert) should implement the 'v' command
(2) pressing 'h' should not display the current patch (as it may be very long),
but remind the user to press 'v' to see the current patch again
17:31:20 < fasta> When I do darcs record; <press ?>, the help message is shown
_above_ my patch, making the help message invisible. How is this caused? Do I
have a bad terminal or is this a darcs bug?
17:33:26 < fasta> 1.0.9rc1 (release candidate 1), btw
17:33:27 < kowey> fasta: i think the problem is that if we did it the other way
around, you'd have to scroll past the help message to see what your patch is
17:33:50 < kowey> which might also be confusing
17:34:26 < fasta> kowey: well, I would say that pressing ? means the user wants
to see what all the options mean. The user can later press another key to see
the patch again.
17:34:51 < kowey> hmm... maybe record should have a 'v' option
17:34:59 < kowey> (in the interactive dialog)
17:35:03 < fasta> right
17:35:50 < kowey> sounds like the kind of thing you'd want to discuss on the
-users list
17:36:25 < kowey> arguments you might expect to respond to are what i suggested,
and "oh adding a v switch just for that purpose might introduce clutter"
17:37:09 < kowey> or maybe...
17:37:20 < kowey> the help screen, and only the help screen could say "hit v to
see your patch again"
17:38:02 < fasta> kowey: yes, I guess "designing" an application to make
everyone instantly happy is quite difficult. :)
17:38:45 < kowey> yeah... i wish we had some people on board that had insight
into this kind of stuff
17:48:09 < kowey> another idea is that help should 'block' the interactive
dialog - and you'd have to hit Esc or something to exit help
17:49:35 < fasta> kowey: if the message fits in the whole screen, that is
better, yes.
17:49:58 < fasta> kowey: hitting any key might be better
17:50:29 < kowey> i'm not sure about that - i wouldn't want the user to hit a
key (thinking they were saying yes to a patch)
17:50:35 < kowey> and not have it taken into account
17:51:17 < kowey> i also don't want the user to "worry" about accidentally
sending commands to the interactive dialog
17:51:24 < kowey> like "what if i accidentally hit y twice"
17:52:41 < fasta> kowey: you can ignore keys pressed in rapid succession.
17:52:49 < fasta> kowey: e.g. .5 seconds.
17:53:00 < kowey> well, i'm more thinking of the user model of darcs
17:53:15 < kowey> in a sense, having a priviledged 'get-out-of-help' key
17:53:22 < fasta> kowey: although, maybe Esc is indeed better.
17:53:30 < kowey> guarantees to the user that trying to get out of help won't
make anything else happen
17:53:50 < kowey> it is a trade-off though... because now if the help screen
blocks the whole dialog, we lose a little immediacy
17:54:13 < kowey> for example, it's not longer "what was that command to get out
of darcs? <h> .... ah, it's q! <q>"
17:54:33 < kowey> but "<h>... ah! it's q <q>... wait nothing happened... oh!
<Esc> <q>"
17:54:52 < fasta> kowey: I think it would be best to just display the output of
the help directly, and not above the patch.
17:55:04 < fasta> kowey: then just do v to view the patch. User model did not
change.
17:56:08 < kowey> yeah, that's probably the conservative way to go about it...
other commands have 'v' anyway
17:57:53 < kowey> i think you should submit something to bugs@darcs.net,
proposing these two changes together (1) help should only display help (2) all
commands should have a 'v' option, including record
17:59:02 < kowey> well, you'd also have to think about how help works in other
commands, maybe (1) only applies to record
17:59:42 < kowey> or more precisely, to commands that deal with individual
changes (hunks, etc), as opposed to darcs patches
17:59:54 < kowey> (revert, unrevert, amend, record)
|
