Bug 8852 - wang keypad modularisation
Summary: wang keypad modularisation
Status: RESOLVED FIXED
Alias: None
Product: xkeyboard-config
Classification: Unclassified
Component: General (show other bugs)
Version: unspecified
Hardware: x86 (IA32) Linux (All)
: high normal
Assignee: xkb
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on: 11611
Blocks:
  Show dependency treegraph
 
Reported: 2006-11-01 06:48 UTC by Nicolas Mailhot
Modified: 2007-07-16 01:27 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
Proposed patch (10.54 KB, patch)
2006-11-01 06:49 UTC, Nicolas Mailhot
Details | Splinter Review
Proposed patch (9.51 KB, patch)
2007-01-04 12:59 UTC, Nicolas Mailhot
Details | Splinter Review
Proposed patch (9.10 KB, patch)
2007-07-15 06:12 UTC, Nicolas Mailhot
Details | Splinter Review
Proposed patch (9.01 KB, patch)
2007-07-15 08:37 UTC, Nicolas Mailhot
Details | Splinter Review

Description Nicolas Mailhot 2006-11-01 06:48:12 UTC
The attached patch modularises the wang keypad. It seems to do the right thing
inside the be(wang) layout, not so sure when the option is used directly. Not
that running rawhide at the start of a fedora cycle is a good way to detect
brokenness :)
Comment 1 Nicolas Mailhot 2006-11-01 06:49:04 UTC
Created attachment 7619 [details] [review]
Proposed patch
Comment 2 Nicolas Mailhot 2006-11-01 06:49:41 UTC
Please test this wang modularization
Comment 3 Nicolas Mailhot 2006-11-01 08:14:28 UTC
(BTW some parts of xkeyboard-config use "numpad", others "keypad", this is all
very confusing)
Comment 4 Sergey V. Udaltsov 2006-11-01 15:01:08 UTC
Nicolas, it seems be(wang) = be(basic) + keypad(legacy_wang). May be, it would
make sense to drop be(wang) altogether? Just creating the compatibility rule
mapping it to be(basic)...
Comment 5 Nicolas Mailhot 2006-11-01 15:33:22 UTC
Well actually if you look at be(wang) it also tweaks two keys. Using only
keypad(legacy_wang) would kill this tweaking

I modularized wang because I was looking at be for be(oss) and it seemed the
right thing to do (esp. since be(wang) comments points at a uk keyboard so I
suppose the wang keypad is not used only in Belgium)

But I have no wang keyboard so I haven't the faintest idea if dropping the rest
of be(wang) is ok. Probably whoever asked for be(wang) would be better placed to
comment
Comment 6 vdb128 2006-11-01 17:35:20 UTC
The last Wang azerty Belgium 724 keyboard I bought in 1996 had two 
additional engravings, a tilde and a brokenbar located in the top right 
corner of the TLDE and LSGT key respectively.  This model was 
unavailable shortly after so I guess it is uncommon. 

It would probably be fine to drop support for this layout.  On the other 
hand, the tilde and brokenbar could be added to the be default since it 
is a strict addition.  

Model number overview:
  724 qwerty US = pc101 + 8 keys (Help, F13-F16, Cancel, Erase, GL) (725-3770)
  724 azerty BE = pc102 + 8 keys + three level keypad (item 725-3771-ae)
    pre 1996: came with a coiled grey PE DIN/5 to RJ11 telephone cable
    1997: used a high quality coiled ivory synthetic rubber cable 
  724 qwerty UK = pc102 + 8 keys + three level keypad (item 725-3771-uk)

Pre 1990 keyboards came without PC-style arrow engravings on the numeric 
keypad.  It was the digit with or without the additional symbol.  
The PC upgrade kit contained three level keycaps, a tweezer, and 
instructions, all shrink-wrapped on a piece of cardboard.  
Comment 7 Sergey V. Udaltsov 2006-11-02 12:16:22 UTC
More questions:
1. Why not just make keypad(*number_wang) public? Instead of keypad(*_wang)?
2. Why THREE_LEVELS_KEYPAD is necessary? AFAICS FOUR_LEVELS_KEYPAD would be
enough, wouldn't it? I do not think that processing of NumLock + Level3 and
Shift + Level3 is critical - I would prefer not to create new types without real
need.
Comment 8 Nicolas Mailhot 2006-11-03 04:46:18 UTC
(In reply to comment #7)
> More questions:
> 1. Why not just make keypad(*number_wang) public? Instead of keypad(*_wang)?

Your call. Would be an easy decision if every layout used the modularized keypad
definition, but right now only two (fr(oss) and be(oss) do). Also while it does
not change a thing for basic wang, you'll lose the math operators for unicode_wang

> 2. Why THREE_LEVELS_KEYPAD is necessary? AFAICS FOUR_LEVELS_KEYPAD would be
> enough, wouldn't it? I do not think that processing of NumLock + Level3 and
> Shift + Level3 is critical - I would prefer not to create new types without real
> need.

Maybe I've overengineered a little :) It's really a three-level keypad though, I
like the fact the code made that clear. I can rework the patch a little if you
really want it 4-level

Comment 9 Sergey V. Udaltsov 2006-11-04 15:54:28 UTC
> Also while it does
> not change a thing for basic wang, you'll lose the math operators for unicode_wang
Well, you're right at this point. Let's leave it as it is.
 
> like the fact the code made that clear. I can rework the patch a little if you
> really want it 4-level
Yes please do so. Even though we do not use level 4 - I think it can be safely
ignored, for simplicity sake.
Comment 10 Nicolas Mailhot 2007-01-04 12:59:29 UTC
Created attachment 8294 [details] [review]
Proposed patch

Updated patch that does not need a new type
Comment 11 Nicolas Mailhot 2007-01-04 13:03:28 UTC
BTW it seems I've been modularising keypad layouts (keypad options) while people
have been modularising keypad rules (numpad options)

The problem is the two changes ignore each other : if can't both use a
modularised keypad layout option and a numpad rule. Unfortunately the numpad
changes make my head spin, but something should be done about this
Comment 12 Nicolas Mailhot 2007-01-04 13:37:29 UTC
(In reply to comment #10)

> Updated patch that does not need a new type

Thinking about it some more — maybe putting a Nosymbol at fourth level is more
correct ? IIRC fourth level inherits third if not defined
Comment 13 Nicolas Mailhot 2007-07-15 06:12:29 UTC
Created attachment 10738 [details] [review]
Proposed patch
Comment 14 Nicolas Mailhot 2007-07-15 06:14:14 UTC
Ok, new patch
Since I don't have a wang keyboard this will be the last one unless someone finds a problem.

It depends on the kpdl stuff in bug 11611
Comment 15 Nicolas Mailhot 2007-07-15 08:37:23 UTC
Created attachment 10741 [details] [review]
Proposed patch

Remove cut & paste duplicate
Comment 16 Sergey V. Udaltsov 2007-07-16 01:27:07 UTC
Again, forgot base.o_s.part ;)
Committed, thanks.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.