Subject: RE: [E-devel] [PATCH] notify immodule to know the cursor
location
In the current edje_entry, there are some bugs related to sequence and miss
to call '_edje_entry_imf_cursor_info_set()'.
Cursor geometry information should be transferred after
ecore_imf_context_focus_in API is called.
SVN revision: 62039
The color class was removed from the list but the edje object
was not updated neither the signal was emmited.
Patch by: Flavio Ceolin <flavio.ceolin@profusion.mobi>
SVN revision: 62019
problems about it storing invalid config when the same build trees are
shared between multiple architectures and machines. :( better we
build right than build fast.
SVN revision: 61974
Subject: [E-devel] [PATCH] check NULL parameter in edje entry
This simple patch is for checking NULL parameter in the callback
functions
related to input method framework.
I think it is necessary for safety.
SVN revision: 61890
Subject: [E-devel] [PATCH] notify immodule to know the cursor location
In immodule, the cursor location is needed to move the word candidate window.
In this patch, calling ecore_imf_context_cursor_location_set lets immodule
know the cursor location.
Also fixed patch conflicts and 1 warning related to IMF.
SVN revision: 61877
To: enlightenment-devel@lists.sourceforge.net
Subject: Re: [E-devel] E SVN: raster trunk/edje/src/lib
Date: Wed, 22 Jun 2011 15:32:41 +0200 (CEST)
On Wed, 22 Jun 2011, Enlightenment SVN wrote:
[...]
i would have prefered a beautiful colored spank message with ERR
Vincent
SVN revision: 60591
to have it in %{release} also. So let's just tag the package as ours
and try to make sure it doesn't interfere with vendor releases.
SVN revision: 60407
Subject: [E-devel] Provide sensible errors for edje utilities
If you start edje_player or edje_inspector with a path to a file that doesn't
exist you get a bogus error message saying that the file doesn't contain any
groups. The attached patch uses access() to check if the program can read the
file, giving a sensible error message if not.
I have checked the other utilities there, too, and they work, with the
exception of edje_external_inspector. I'm not sure how this one works at all,
but it seems not to take a file but a list of modules, maybe someone with
greater insight can take a look at that.
Going through things installed under bin, I'll take a look if those behave
properly and create patches for those, too, if this one is okay.
SVN revision: 60338
Subject: [E-devel] Edje: using fdopen instead of fopen in edje_cc
On windows, using open() followed by fopen() does not work. Hence, in
edje_cc, where mkstemp (which uses open) is followed by fopen, edje_cc
fails.
Instead of fopen, we can use fdopen. I pasted a patch below. Can you
comment it (like, instead of keeping the filename in the function that
i modified, why not using it for the fd?
(changes - closefd) removed from data_write_scripts() as fclose()
handles that)
SVN revision: 60299
This reverts commit 60239
I guess the world is not ready for correct min height calculation, too
many things depend on the current behavior. Will re-apply sometime in
the future.
SVN revision: 60264
1. table to have min.h/v ability like box
2. to ACTUALLY implement box h/v (and well of course implement
tableh/v too)
this basically fixes this working at all and completes the feature to
table too.
SVN revision: 59960
and sewt a font for menu item.. or menu title. segv. (try it under
falgrind) the members list in the hash is garbage when its trying to
remove a member. double heck your changes and test :)
SVN revision: 59902
Read my comment in the email, and we even talked about it on irc. The
current call to "mainloop iterate" is hacky, Jihoon is already working
on an alternative solution.
SVN revision: 59734
When elm_entry_entry_set() is called, the preedit text in the entry should
be cleared. ATM, it doesn’t works so.
This patch will resolve this problem.
When ecore_imf_context_reset() is called, the preedit text in the immodule
will be committed by calling ‘ecore_imf_context_commit_event_add()’.
ecore_imf_context_commit_event_add puts the event in ecore event queue, so
it will be processed in the idle state of ecore loop.
According to this mechanism, commit text is inserted later, so I use
‘ecore_main_loop_iterate();’ for inserting the commit string instantly.
SVN revision: 59732
We don't really need or use the coordinates anywhere, we just saved them
and worked hard on maintaining them, but didn't really use them.
Now we only calculate them in the one specific case where we actually
need them. This speeds things up and also means the layouting may not be
called in some cases when there's no need to layout.
SVN revision: 59520
Ok was shared between width and height and in the error, and in the error
case we don't reset maxw and maxh; This caused maxw and maxh to be added
a couple of times even when not needed. This would cause weird sizing.
SVN revision: 59450