process to stop e from ever "locking up". this unfortunately breaks the other
source stuff and thus breaks the app editor config dialogs and anything using
fm2 for .desktop stuff from the e applications/all repo. i am doing this as i
expect this to go away with efreet anyway so not too much of an issue. not
all fm operatiosn work currently - note. so beware. it can view files though :)
SVN revision: 28528
all group names prefaced by e/
its entirely possible i missed a few small places in this, so there may be some growing pains in the next day or two
also, spectrum content -> e.swallow.content
SVN revision: 25057
finish up the color selector widgets (plug some leaks)
TODO:
* color_dialog needs a way of specifying initial color
* color_well needs to use this when showing its color dialog
* the grad dialog should remember the last selected colors
* the color dialog whould probably keep a list of favorite / recent colors
Also, it looks like there's a leak (according to valgrind at least) in e_editable.c when you set the text and it reallocs.
SVN revision: 24953
basically if moduels wish to put objects on a desktop - they now will have to
DIY and look after themselves. until we have gadcons with "free" layout that
do similar stuff - but this simply means module authors need to d a bit more
work. gadman code is unmaintained and a burden - thus removed.
also borders and e_win's have their own pointers now - this should nuke the
stuck pointer bug... :)
SVN revision: 24925
dialog api's - added name and class params so u can set the name and class -
so users can later use things like rememebr to rememebr where a dialog will
go for usability and convenience.
SVN revision: 24691
pass in a color struct to update
(use the change hook to know when changes have occured)
use a short timer instead of an idler to update the spectrum image data
-- ideally, we would add a pre/post render hooks for smart objects. then we could redraw int he pre-render callback.
SVN revision: 24676
Change the last test in e_test.c to #if 1 to play.
Issues:
Redrawing the H,S and V spectra is a bit slow. Needs to be optimized.
Spacing / Sizing of the various pieces of the csel widget needs to be cleaned up a bit.
I haven't actually themed anything yet. Just some fairly basic place holders to get it working. We should wrap each piece in shadowed cutout graphic, and make fancier cursors.
A color well widget
- show the currently selected color
- allows dnd of colors (this probably requires some reworking of ecore selection code)
For anyone wanting to play with this, a single E_Color struct is used by all parts of the selector, preventing the need to copy values around. This includes RGB *and* HSV values since the two spaces aren't 1:1.
SVN revision: 24669
fm2 now is much better at handling massive dirs of files - my test case is
5400 files or so in a dir (all jpg's). it seems rather happy dealing with it
and doesnt lock up or hang or seem to have issues. it just plods along.
SVN revision: 24290
i need to add some form of indicator that it is still scanning the dir and
building the file list as its a little hard to tel that it's busy (but still
usable)
SVN revision: 24289
fixed :) all hail a better sorting algorithm - incriemntal insertion quick sort
:) (or whatever you want to call it) and a little better "scheduling" in the
code :)
SVN revision: 24246
2. lots of code on efm tomake it work well for file selection
3. a file selector widget with all the bits and pieces
4. first move - make wallpaper bg import tool use the new file selector -
it's a start.
5. pants on!
SVN revision: 24139
is pretty much almost perfectly working. i have fixed up some e_thumb stuff
and allowed e_thumb to be more responsive and skip items that are known to be
"generated" and bring them ahead in the list of things to thumb - so kind of
a priority skiplist - process what it KNOWS will be already done first
quickly and leave the slower stuff until later. efm is fairly well refined
now - as above. the test selector works nicely. also added an almost-sha1
generator - use sha1 sums of the path for thumbs - less likelihood of
collisions. the prolbme is given the small size of the input data... it's
hard to do well - but anyway :)
SVN revision: 24128