Re: A whoops in the sequences fixups?

Since I think we (you) are discussing something based
on the diff I sent to exmh-users I'll try to comment.

>   | Should that be 'seqShow {} {Show Sequences}'?
> 
> No argument on that one.

Agreed. My diff was a quick and dirty one
(should have pointed that out more explicitly,
 to invite people more explicitly to 'shoot at it', 
 sorry about that)
I really overlooked the default setting (cur).
I do remember having to get rid of 'cur' myself without
understanding (nor looking) where it came from.

> 
> | (Also, 'cur' being listed in seqShow seems to over-ride it's being listed
> | in seqNeverShow - I currently have seqNeverShow as 'pseq cur', and have
> | an unseen and cur in the seqwin.  A misordered logic check someplace?
> 
> But here I think the way it works is correct -

agreed - that works as I intended, I guess.

> I suspect it is the use of the word "Never" that confuses things.
> 
> If it worked the way you're thinking it should, there'd be no (easy) way to
> list the sequences that are to be shown, instead of the ones that aren't.
> 
> Now one simply uses '*' as the "Never show", which removes all of the
> sequences, then lists the sequences to display in "Show" - if the negative
> list overrode the positive list, that wouldn't be possible (and since the
> default is "show", having a positive list to augment that before the
> negative list to subtract sequences adds nothing useful).

After posting my initial diff I have been wondering whether it
makes sense to allow the user to choose the override order
(which overides what, a bit like the the possibilty to specify
 the 'order' as in 'order allow,disallow' that I think I
 remember having seen in web server (apache?) config files
 when access rights are specified)

> Perhaps just changing "Never" (in the menu label) to "Don't" would be
> enough to make it a little clearer what happens?

Probably a lot can be cleaned up/cleared up by better wording
of the text in the preferences window and the help texts, agreed
(again, my diff was quick and dirty, sorry)


Alternatively/additionally, if a sequence matches both with the
positive list(s) and the negative list(s?), we could try to see if
one match is more 'precise' (maybe there is a better word)
than the other one, and give precedence to the most precise match,
where I mean by precise: less globbing needed to match
(so c*r is a more precise match for cur than *; cur is most precise)
However, this does not solve the case where a sequence appears
both in positive and negative lists.

On a related note: I like the concept (words) postive and negative
because I think it would be useful to have three list of sequences:
 - one negative (don't show is indeed maybe a better name)
 - one positive (show when non-empty)
 - another positive (show 'always')
where, when it comes to deciding whether or not a sequence
is a canditate for showing, both positive lists have the same
effect, i.e. a sequence needs only to be in one of the three
lists, and it is not necessary to have (e.g.) unseen in
both the 'show-when-non-empty' and the 'show always'.

Axel.