libX11: Changes to 'master'

[Pander]

On 2012-04-09 18:21, Julien Cristau wrote:
> On Wed, Apr  4, 2012 at 13:28:17 +0200, Pander wrote:
> 
>> On 2012-03-31 13:05, Julien Cristau wrote:
>>> On Thu, Mar 22, 2012 at 16:36:27 +0100, Pander wrote:
>>>
>>>> James, Julien, did you receive enough compelling arguments to resolve
>>>> this conflict so that <o> <apostrophe> results in ó. Then we can finally
>>>> close the book on gtk+/GNOME conflicts.
>>>>
>>> OK.  Can you send a patch?
>>
>> Sure, see attachtment. Now the apostrophe series is consistent and no
>> more gtk+/GNOME conflicts exist, see:
>>
> That's fine with me.  

Could you please apply that one already so it is of the list and we
discuss the rest below? THanks.

I would still argue for reverting some of the
> new sequences added in cf040016d455bc37f7665d6714337c5eafd8ea94 though,
> e.g. for ¥, §, ¤, ² and so on (I've only looked at the top of the diff).

From:

http://cgit.freedesktop.org/xorg/lib/libX11/commit/nls/en_US.UTF-8/Compose.pre?id=cf040016d455bc37f7665d6714337c5eafd8ea94

We already threw out several redundant gtk+/GNOME compose sequences for
¤, ¥ and § that were a bit over the top, there mainly 'o' was replaced
by a '0' (zero) in additional sequences..

¤ does indeed not really need XO, OX, oX, Xo, Ox and xO variations in
the added compose sequences. These new six sequences are for me not
important but are available in gtk+/GNOME. I'm OK with removing:
  XO
  OX
  oX
  Xo
  Ox
  xO

For ¥, I would argue to keep the added upper case Y combinations that
have been added because the shape of glyph looks like an upper case Y.
And also the added lower case sequences to support = and -. For §, a
similar reasoning.

For ², ³, µ, etc. keep at least the lower case sequences, also the
reverse sequences. On the upper case 2S, S2, 3S, S3, U/ and /U I have
the same argument as on ¤ above. Doubtful but available in gtk+/GNOME.
I'm OK with removing:
  2S
  S2
  3S
  S3
  U/
  /U

> Also you added some sequences like '<Multi_key> <$letter> <minus>'
> sometimes for '$letter with tilde', and sometimes for '$letter with
> macron', or '$letter with stroke'.  For keysyms that already had
> existing sequences, and where the new ones don't seem to add much
> benefit as far as I can tell.

You mean the ã, Ã, ñ, Ñ, õ and Õ created from letter + minus. These are
also not important to me. If you all insist I am OK with replacing:
  a-
  -a
  A-
  -A
  o-
  -o
  O-
  -O
with sequences resulting in āĀōŌ conform the current series and removing:
  n-
  -n
  N-
  -N

Regards,

Pander

> Cheers,
> Julien