Re: [wmii] Stacks

From: Anselm R. Garbe <>
Date: Fri, 3 Mar 2006 12:24:28 +0100

On Fri, Mar 03, 2006 at 11:55:56AM +0100, Uriel wrote:
> And I found a fundamental flaw in this idea, would be nice... except
> you actually need to keep track of pages for how all the clients are
> positioned and organized in each page, in virtual pages there would
> not be any way to maintain a layout, which is very dynamic and also
> very stupid and useless.

Yes, something like a virt-ws cache for each tag in use would be
necessary (an entry is created if the first tag has been written
for the first time and removed if no client exists with such a
tag). In general there might be only a bunch of tags, but not
many. This means, you always have layout information per tag.

> Well, I guess we could keep track of what the layout under a certain
> name(set of tags) did look like, can't be much data should grow very
> little if at all, and it can easily be garbage-collected when no
> client has any of the the tags for that 'virtualws'. This would make
> things certainly much more dynamic, but it would make it impossible to
> list
> all available ws, only way would be to make them up on the fly (ie.,
> associate key bindings to commands like I described before), and one
> could easily list all tags currently in use, so maybe it's not such a
> big deal(you could have a on-the-fly-pager which would do just that
> and allow you to click on any of the available tags)

With above idea to create an (unselected) virt-ws if a tag
occures for the first time, you can simply re-use the current
pager to see which clients relate to which tag.

> Actually, you got it backwards, there are no 'tag-groups' except the
> current set of visible tags. The references would be in the other
> direction, from the clients, to their own private list of tag names,
> when switching the list of currently visible tags, you would go
> through all windows and show the ones that match.

That is clear, but I meant the term 'tag-groups' in the same
sense, though maybe an odd term.

> And yes, I suspect this would simplify and generalize the sendto
> system, to send the selected client to the email 'virtual-ws'(again,
> this ws would not exist anywhere, would only exist once one sets the
> 'current tags' to something that contains "email")
> echo email > /wmii/client/current/tags
> To add the client to an existing virtual-ws but without making it go away:
> echo email >> /wmii/client/current/tags


> (I'm talking about virtual-ws in a metaphorical way, as I said, there
> is no such thing, except maybe as a cache of layout of windows once it
> has been made the active set of tags once)

If you give a client a tag, you can simply create a virt-ws for
it, if not existing yet - each tag should be intended to be

> Yes, you could allow even wildcards in the tag names, so a client with
> a tag of '*' would match any selection and always show up (I think
> here a special case should be made if that client is floating, it's
> layout should be made 'universal' and not keep in per-virtual-ws and
> instead associated directly with the client)

Well, that would make the above idea impossible to create a
virt-ws ad hoc on the first occurence of a tag. But maybe it
might be more of use, than having that pager... The pager
information is pretty limited in such a dynamic environment.

Though dunno, how well '*' would scale *g*.


 Anselm R. Garbe  ><><  ><><  GPG key: 0D73F361
Received on Fri Mar 03 2006 - 12:24:29 UTC

This archive was generated by hypermail 2.2.0 : Sun Jul 13 2008 - 16:00:36 UTC