Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!cs.utexas.edu!ut-emx!walt.cc.utexas.edu!erlkonig
From: erlkonig@walt.cc.utexas.edu (Christopher North-Keys)
Newsgroups: comp.windows.x
Subject: Re: International Keyboards and Modifier Combinations
Keywords: lock, keyboards, shift, X
Message-ID: <28037@ut-emx.UUCP>
Date: 11 Apr 90 21:56:43 GMT
References: <5567@scolex.sco.COM>
Sender: news@ut-emx.UUCP
Reply-To: erlkonig@walt.cc.utexas.edu (Christopher North-Keys)
Organization: Packaging/Interconnect, M.C.C.
Lines: 49

/----[mikep@sco.COM (Mike Patnode) The Santa Cruz Operation, Inc.]
| I'm also interested in dealing with Lock keys other than Shift/Caps
| Lock.  Particularly, the NumLock key on a PC/AT keyboard as well as
| other locking modifier keys.  Caps Lock seems to be sort of hacked into
| X without any provisions for other "locking" keys.

The solution looks clear to me --- code a "soft-lock" or "polylock" key the
applies to whichever set of modifiers is depressed when the softlock is
struck.  You could probably do this with a small X program;  I don't think
an extension would be necessary, but I not that deep in the source yet.

More specifically, if:

	{M==meta, S==shift, L==lock, [xX]==any char, ^==up, v==down},

then the sequence:

	XvX^ Sv LvL^ Sv XvX^ XvX^ Sv^ XvX^

should produce the output:

	xXXx

Also:

	Sv Lv^ S^ ...                shift-lock (caps) set
	Sv Lv^ S^ ... Sv^ ...        caps set, then unset
	Sv Mv Lv^ S^ M^ ... Mv^      caps and meta locked, then meta unlocked
	Lv^                          all soft-locks cleared

The implementation would need to track the shifted states for (say)
Meta, Control, Upper(case/numeric), Super, Hyper, Greek, etc.  Such a list
could validly be a dozen or more long, and should indicate:

state 0>  unset, awaiting shift-down
state 1>  set,   shift-down has been seen, but no shift-up yet.
state 2>  lock+, lock was seen while shift set, ignore next shift-up.
state 3>  lock,  lock and shift up seen, next shift-down triggers state 1.

The implementation could use a simple state transition table, and might be
best coded externally from the kernel.

Just thinking...


------------------------------------/\----------------------------------------
Seo:  Harp[@Mcc.Com]               /  \/\ ^*^           Christopher North-Keys
Tha mi gu trang a'cluich.         /    \ \         Assoc. Systems Analyst, MCC
--------------------------------(disclaimer)----------------------------------