[PD-dev] [ pure-data-Patches-3160982 ] [bugfix] allow to type UTF8 characters in symbolatoms

[SourceForge.net]

Patches item #3160982, was opened at 2011-01-18 09:16
Message generated for change (Comment added) made by millerpuckette
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=3160982&group_id=55736

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: None
>Status: Pending
>Resolution: Accepted
Priority: 5
Private: No
Submitted By: IOhannes m zmlnig (zmoelnig)
Assigned to: Miller Puckette (millerpuckette)
Summary: [bugfix] allow to type UTF8 characters in symbolatoms

Initial Comment:
since Pd-0.43 patches can handle UTF-8 characters. great!

unfortunately, it doesn't seem to work when typing in symbolboxes.

example:
 i switch to a russian(DOS) keyboard layout
 i create a message box and type the keys 'asd' which correctly shows up as 'ыфв'
 i create a symbolgatom and type the keys 'asd' which incorrectly show up as 'DK2'

note: sending a [symbol ыфв( to the symbolgatom, will correctly display it.

it seems like the keyboard input sanitation code for the symbolbox perverts my input.


----------------------------------------------------------------------

>Comment By: Miller Puckette (millerpuckette)
Date: 2011-02-01 20:45

Message:
applied to 0.43 test

----------------------------------------------------------------------

Comment By: IOhannes m zmlnig (zmoelnig)
Date: 2011-01-19 03:04

Message:
thanks.

i added your amendments to the attached patch

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2011-01-19 02:51

Message:
... of course that was meant to be:

- char utf8[4];
+ char utf8[UTF8_MAXBYTES];



----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2011-01-19 02:51

Message:
argh... beat me to it! :-)

anyhoo, you might want to also:

+ #include "s_utf8.h"
...
-             char utf8[4];
-             char utf8[UTF8_MAXBYTES];

so that things don't break if we need the full UCS-4 range someday. 
Sorry; I would have attached a patch (now that I've managed to acquire a
rudimentary working knowledge of git :-), but I don't seem to be able to
(looks like an instance of
http://sourceforge.net/apps/trac/sourceforge/ticket/16001)

----------------------------------------------------------------------

Comment By: IOhannes m zmlnig (zmoelnig)
Date: 2011-01-19 02:11

Message:
mukau, i'm IOhannes not hans :-)
i'm working on pd-vanilla:
http://pure-data.git.sourceforge.net/git/gitweb.cgi?p=pure-data/pure-data;a=summary

----------------------------------------------------------------------

Comment By: IOhannes m zmlnig (zmoelnig)
Date: 2011-01-19 02:09

Message:
anyhow, i have a fix for the problem.
instead of writing the wchar into a char, the wchar is first decoded to
utf8 using  u8_wc_toutf8() and only then written into the symbol-buffer of
the atom.

this seems to work fine.

the attached patch modifies g_text.c and applies to f5cabb ("took out
debug msg")

----------------------------------------------------------------------

Comment By: Bryan Jurish (mukau)
Date: 2011-01-19 02:08

Message:
ok; trying to look into this now; have you any idea what sources you built
from?  I'm currently using most recent snapshot from
http://pure-data.git.sourceforge.net/git/gitweb.cgi?p=pure-data/pd-extended.git;a=summary
; ... svn doesn't appear to have the utf-8 patches applied anywhere yet...

----------------------------------------------------------------------

Comment By: IOhannes m zmlnig (zmoelnig)
Date: 2011-01-19 02:01

Message:
@mukau: no, LC_* are not changed at all (btw, they are set to de_AT.UTF-8
here)
the kbd layout changes only effect the key->character mapping, way before
tk is involved.
also note, that tk can handle all quite well, as can most parts of Pd
(confirm: i can type cyrillic characters in objects, msgboxes and
comments)

it's only gatom's that do some additional magic to filter out unwanted
characters, and then write the received wchar into a char without checking.

----------------------------------------------------------------------

Comment By: Bryan Jurish (mukau)
Date: 2011-01-18 12:45

Message:
What exactly does "changing the keyboard layout" entail?  Does it e.g.
change your locale (LC_ALL, LC_CTYPE, etc.)?  That might annoy Tk's unicode
handling...

----------------------------------------------------------------------

Comment By: IOhannes m zmlnig (zmoelnig)
Date: 2011-01-18 09:31

Message:
one more thing:
in the original report i said that typing 'asd' should result in 'ыфв'
which is wrong.
it should rseult in ''фыв'
('a'='ф', 's'='ы', 'd'='в')

----------------------------------------------------------------------

Comment By: IOhannes m zmlnig (zmoelnig)
Date: 2011-01-18 09:30

Message:
btw, in normal (de) kbd-layout, when i type (and releae) 'a', pd-gui
sends:
.x99e9708 key 1 97 0;
.x99e9708 key 0 97 0;

whereas when i switch to russian, pd-gui sends (again key 'a'):
.x99e9708 key 1 1092 0;
.x99e9708 key 0 a 0;

"1092" is the unicode number for "ф" (which is what should show up, when
i press 'a')
i don't know why the keyup sends 'a' instead.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=3160982&group_id=55736