* Remove vim modelines:
find . -name '*.[chx]' -exec sed -i '/\/\*$/ {N;N;/ \* vim:ts/d}' \{\} \;
find . -name '*.[chx]' -exec sed -i '/\/[\*\/] *vim:/d' \{\} \;
* Remove leading blank lines:
find . -name '*.[cxh]' -exec sed -i '/./,$!d'
If you use vim, use this in your .vimrc:
set ts=8 sw=3 sts=8 expandtab cino=>5n-3f0^-2{2(0W1st0
SVN revision: 50816
if we have a compositor. just "mark" as unmapped and leave there.
pretend we are unmapped. disable events and stuff on the border to get
the behaviro right too so stacking or positioning is never affected.
SVN revision: 50761
that can trivially get objects inside the compositor canvas as well as
get the compositor canvas and events when it comes and goes, changes,
when client sources are added, removed etc. etc.
SVN revision: 50727
fullscreen windows).
I modified this patch slightly. Removed some unused handlers. limits.h
is already included in e.h, so no need to reinclude it. Fixed formatting.
SVN revision: 50356
The notnull.cocci script from Coccinelle finds places where you check if a
variable is NULL, but it's known not to be NULL. The check can be safely
removed. For example, this code would be caught by notnull:
if (!var) return;
if (var && var->fld) { ... }
It's needless to check again if var is not NULL because if it's in fact NULL,
it would have returned on the previous "if". This commit removes all the
trivial places where this pattern happens. Another patch will be generated for
the more complex cases.
SVN revision: 50241
int geometry_auto_move; // GUI
int geometry_auto_resize_limit; // GUI
This does influence the configure_request algorithm in the way that the user could disable the window
geometry move and resize feature that is useful for e.g. Gnome fileselector or gnome-terminal tabs.
There was a use case on the list where this was not nice so I created an option to disable it. If one thinks
a special e.src shouldn't activate this feature by default, please change it.
SVN revision: 50083
If a window likes to open with bigger
initial coordinates bigger than the screen
(minus ignore zones like not overlap allowed shelfes)
the window should be sized to the maximum allowed size.
SVN revision: 50028
If a window likes to maximize while it stays below a ignore zone (like not overlap allowed shelfes) it should be maximized only in the overlap allowed zone. At returning to normal state it should be placed in the original position.
SVN revision: 49942
If a window resizes itself without border interaction it's moved to the next available coordinates which aren't covered by ignore zones like not overlap allowed shelfes.
SVN revision: 49910
If a window likes to open with initial offscreen coordinates E should move it to the nearest onscreen coordinates.
req_window_ignore_zone:
If a window likes to open with initial coordinates in a window ignore zone (like not overlap allowed shelfes) if should be moved to the next allowed coordinates.
SVN revision: 49909
while working with it I got some strange window placements. I need to analyze these situations before I release it to the public...
SVN revision: 49673
Allow one to send window to corners. up-left, up-right, down-left and
down-right are the new possible values.
HINT: for those that don't know, the "Push" has the nice behavior to
do the push in the given direction but stops in the first
window/obstacle in that direction, thus avoiding overlapping windows.
SVN revision: 49433
Now push to edges or move to position will account for shelves and not
move things below/ontop of them.
Move to help was also changed to say "+" instead of " " as it is more
clear. The code just checks for '-', so anything else is accounted as
positive anyway.
SVN revision: 49432
The new trend these days seems to be "place a shelf on top of the
screen", in that case your desktop icons will get under that shelf and
looks annoying, people laugh at E users that use that :-D (fortunately
I'm the one that usually laugh...)
Fixed, but possible other places to be fixed as well. If one wants to
avoid the shelves, use the e_zone_useful_geometry_get().
e_zone now emits E_EVENT_ZONE_MOVE_RESIZE when the useful geometry
becomes dirty. Maybe it is worth to add a new signal, but I'm reusing
it until someone complains.
SVN revision: 49431
NB: To be used in the acpi bindings config dialog. This allows us to freeze
all existing events so that if the user presses the power button
(or something else) when adding a new binding, their machine does not
power off.
SVN revision: 49416
before it was not possible to set an icon by name and not by path.
also one has now direct feedback if an icon can be found in theme
SVN revision: 49398
now we test for hal presence through introspection call and set up handlers only if hal is present. otherwise set a poller to call introspect on hal every 30 seconds and then set up handlers if hal is later detected
note: this commit message is excessively long.
SVN revision: 49373
Add acpi bindings to config.
Add handlers in e_acpi for common/default acpi events.
Add actions for some missing ACPI events.
- NB: Dim/Undim screen do nothing currently except a printf.
SVN revision: 49309
I guess there is nothing more annoying than having a fullscreen window
to watch a quick movie, read some text or webpage or whatever and have
the DPMS to show, or the screensaver to be activated or your screen to
be locked.
This patch uses the new border signal and counts, if there is at least
one fullscreen border, then avoid doing these annoyances.
I'm not accounting for any border parameter/property, like if it is in
a visible screen or desktop. I hope this is fine, otherwise let me know.
I also asked and common sense at IRC (at least to me and englebass) is
that there is no need to configure such behavior, so there is no
configure option for it.
I'm also thinking about maximized windows to have a configurable
behavior to match this, but I'm not sure.
SVN revision: 49262
This makes the first level of the menu simpler. If you want the "Lost
windows" on the main menu back, make sure you remove the "if (dat)"
branch inside _e_int_menus_clients_pre_cb() otherwise you get duplicated menus.
SVN revision: 49199
Create a device hash to match standard device names to types.
Cleanup E_EVENT_ACPI (add missing ones, remove not needed ones).
Actually parse out the ACPI data which we receive.
You can now actually use this ACPI code for event listening, but the
Status field of the event structure needs sanatizing yet (turn into
something meaningful).
SVN revision: 48928
enables eeze_udev support where available, currently only illume2, as a device backend to replace ehal.
unless --enable-device-udev is explicitly specified, e will continue to use ehal as the device backend.
upon selection of eeze_udev, a brief warning will be shown during config summary.
components that do not support eeze_udev will still fall back to hal even if eeze_udev is selected
SVN revision: 48867
This bug was exposed by e_widget_deskpreview, that uses native E
backgrounds as edje group. As backgrounds likely do not need mouse
events, or in the case of backgrounds converted from JPEG/PNG images,
the edje would never get an event, so as e_icon, then those that rely
on events on icon will never get called as no children get events.
SVN revision: 48515
That is why stringshare returns "const char*".
ZMike, pay attention to these cases. I know this one was particularly
nasty to spot since you receive the "const char *path", but later on
you request strrchr(path..) and receive "char *p", the cast is done
implicitly by libC to make all the case works (C++ version fixed this
recently, but in C there is no way).
The crash that lead me to investigate it was spot by Lucas De Marchi
at advanced wallpaper dialog, but actually other places that uses efm
should be equaly broken.
SVN revision: 48508