On 7-Aug-2004 05:28 Derek Martin wrote:
| On Fri, Aug 06, 2004 at 07:42:51PM +0200, Adeodato Simó wrote:
| > * Derek Martin [Fri, 06 Aug 2004 15:04:18 +0900]:
| >
| > > (b) currently behaves sensibly, turning on both options, regardless of
| > > what options are set.
| >
| > (b) is currently a toggle too...
|
| My mistake. I thought I tested that, but I guess I didn't.
Well, it's sort of a toggle. If only one of "sign"/"encrypt"
is set, "both" turns them both on, but if both "sign"/"encrypt"
are set, then it acts as a toggle. I can speak authoritatively
on the subject since I authored the patch which seems to be
causing Derek so much consternation.
On toggles and menu changes:
My motivations for making the menu toggles was ease of use and in
a former version (unpublished I believe) of the patch it was made
clearer they were toggles. Given that this is the development
branch, changing old stable behaviour is not necessarily a bad
thing (if it's a perceived overall improvement). It seems hardly
suitable to assume that one would blindly use this menu without
understanding its (new) behaviour. Unfortunately, there is no
existing place in the documentation where this particular menu
is discussed (as far as I know); otherwise, I would have updated
the relevant documentation too.
If "inline" is not a toggle, then another option would be required
to get back to "pgp/mime" (although it is possible to get back to
"pgp/mime" in a roundabout way). When "inline" is a toggle it just
made more sense to me that the other options also be toggles. That
is very intuitive to me (even if it is not so for Derek); to say that
the current menu is "unintuitive" or "completely counterintuitive"
smacks of hyperbole.
Whether there's consensus on the issue (and I certainly agree there
is one) I can't say, but I will say that the current menu suits me
just fine. It's extremely easy to suggest textual changes to the
menu and change the code, but there is much more work required to
update all the existing translations to suitable choices for any
code changes (as I did for example with the (i)nline option).
It would be prudent to wait for suitable translations of any menu
changes (e.g. "forget it" to "clear", or adding info about the
inline option being a toggle).
Rather than decide on a single behaviour for every user of mutt,
I would submit two additional ideas (not necessarily mutally
exclusive):
1. Create a mutt option pgp_menu_toggle (default is off = old
mutt behaviour) to allow the use of the pgp menu options as
toggles. The only issue is with the inline option I think.
2. Use case differences to distinguish between behaviour.
So "b" with would turn on (b)oth, but "B" would them off.
This would keep the old behaviour (lowercase selections
forcing a selection "on") while allowing single key changes
to turn any options off. This can still be used with the
pgp_menu_toggle setting to enforce some behaviour when the
upper case choices are used (or make them identical to
the lower case versions when pgp_menu_toggle is on).
I'm happy to produce a patch that implements either (or both)
of these behaviours, assuming there's consensus. Simply changing
the code back the way it was doesn't seem fruitful at this point
though.
Sorry for the long-winded post.
--
-dale
Attachment:
pgpFCyz5J6sOt.pgp
Description: PGP signature