Comment 4 for bug 9364

Revision history for this message
Daniel Stone (daniels) wrote :

As discussed on IRC:
14:43 < daniels> seb128: there are two bugs
14:43 < daniels> seb128: one is on X, the other GTK.
14:43 < seb128> daniels: there is not even a gtk app mentionned in this bug
6774
:43 < daniels> seb128: i added ?Dij (U0133) and IJ (U0132) to the X compse
map, but that doesn't matter
14:44 < Keybuk> GTK+'s compose map needs an overhaul
14:44 < daniels> so now if I right-click in gedit, hit Input Methods, and select
X Input Method, I can compose ij and IJ
                 with multi-key
14:44 < daniels> seb128: but, stupidly, GTK's default mapping is *not* this, and
has its own compose handling
14:44 < seb128> ok, so that's good
14:44 < Treenaks> you mean ? and ? ;)
14:44 < daniels> seb128: the table is in gtk/gtkimcontextsimple.c
14:44 < daniels> Treenaks: (yeah, can't be arsed restarting gnome-terminal, and
my utf8-fu is broken for irc anyway)
14:45 < daniels> seb128: so { GDK_Multi_key, GDK_i, GDK_j, 0, 0, 0x0133 }, needs
to be added to the table, presumably
14:45 < seb128> ok
14:45 < daniels> but that didn't work when I was trying something else, so I
dunno if it will be any good
14:45 < daniels> so I've fixed my bug, now you get your half :)
14:45 < seb128> you could provide some details in the bugs
14:45 < Keybuk> there's a huge patch for gtk/gtkimcontextsimple.c in Bugzilla
that adds a few hundred new mappings
14:45 < Keybuk> (GNOME Bugzilla, that is)
14:46 < seb128> ie: giving the xorg fix so I can apply it to test gtk :p

The X.Org fix is:
<Multi_key> <i> <j> : "ij" U0133 # LATIN SMALL LIGATURE IJ
<Multi_key> <I> <J> : "IJ" U0132 # LATIN CAPITAL LIGATURE IJ
in the Compose files of the relevant locale (e.g. en_US.UTF-8).