* added a number of new e_comp functions and macros
* options for disabling effects on objects: this option does not currently have any effect
* all modules which used gadcon popups have been adjusted
* all modules which used input windows to detect close events for gadcon popups have been adjusted to use new popup autoclose functionality
* shelves are now always drawn on the compositor canvas, meaning objects will never get clipped by the shelf (ticket #1810)
* shelves no longer have an event object
SVN revision: 84060
This reverts commit r79842 for the exact reason stated in the commit
log. First I thought it would not affect the usability, but it indeed
did. And it's a bit worse than having to wait 5 seconds to see the state
changes: you never see them, except if you close and open the popup
again. This could be "fixed" by setting up the timer in
_econnman_popup_update() inside "if (inst->popup_locked)", but we would
still not see the transitions, only the final state.
A proper solution would be to first separate the service state changes
from the list updates, then do something like what was proposed in
r79842. Revert it while we don't have a solution that doesn't break the
more common cases of clicking in a service and see it connecting.
SVN revision: 81407
If time is not good enough, can be fine tunned
very easily (just change the define).
I couldn't test this commit, so if tasn
or other could, it would be great.
I believe it won't be a perfect solution
yet, since this way it will take some time
to confirm visually the network states
(association / configuration / connect)
in the popup. We need to handle this
case properly (if I'm correct).
SVN revision: 79842
module build chnages because its fundamentally broken. it DOES NOT
PRODUCE .SO FILES. just .la and .a files. the only reason u dont
notice is.. you ALREADY had .so's installed. i just got in from the
airport... synced and updated.. rebuilt and was met with all modules
not loading... literally - no .so's are installed int he module dirs.
try rm -rf the instaleld module tree.
regardless... this has to be reverted be3cause it's a major break. the
idea is right/nice. the implementation is causing... problems.
SVN revision: 79015
Get rid of warnings like below with recent automake (1.11 and newer):
src/bin/Makefile.am:6: warning: 'INCLUDES' is the old name for 'AM_CPPFLAGS' (or '*_CPPFLAGS')
src/bin/Makefile.am: installing './depcomp'
src/modules/access/Makefile.am:11: warning: 'INCLUDES' is the old name for 'AM_CPPFLAGS' (or '*_CPPFLAGS')
SVN revision: 78720
Checking two times the same message does not make any sense. The only logical
thing seems to be checking for the properties message here. Leaving the wifi
properties message out as not every setup has wifi. At least on my desktop
here I don't have /net/connman/technology/wifi so we should not fail without
it.
Dermarchi or Bdilly please cry or fix if this is not as intended.
(Brought to you by double_condition_check.cocci)
SVN revision: 77440
We need to call the function to udpate the current instance when adding
a new gadget, otherwise it will not represent the current state.
SVN revision: 77017
there is no need to configure it as it will use the EConnMan desktop
file, whatever is available (user can create a new desktop with such
Name= and it will work, this can be done by a gui in e17 itself).
the module will warn people that do not have the application installed.
SVN revision: 76212
With that it should already be possible to connect to many wifi networks:
wep, wpa, hidden, wpa enterprise... using wps.
Actually I didn't test anything.
SVN revision: 76068
It compiles, but it's untested. I didn't allocate Connman_Field because
I'm not sure it will be useful afterwards or if we only have to use it
to create the widget. If it will be used, it would be good to use
einastringshare for its values.
SVN revision: 76065