If we tried to create the new rw file, we must check if it was created
fine before continue, lots of errors are possible there, but mine was
creating the file inside an nonexistent directory.
SVN revision: 34928
The type "char" is not defined to have a sign or not, for performance
issue some platforms (ie: ARM) define it as unsigned. If we need to
rely on the sign, must say so.
SVN revision: 34924
It also add EET_G_VAR_ARRAY and EET_G_ARRAY thanks to
turran's (jorgeluis.zapata@gmail.com) last year patch.
Of course we have test code for it and eet_data.c code
is covered around 90.1%. We need to test with garbage
data for increasing it more.
WARNING: API of eet_data_descriptor_element_add is not
clean and we will need to break it, if we want a clean
API. So this is the plan for version 2.0.0.
SVN revision: 34919
EAPI int ecore_file_dir_is_empty(const char*dir)
Should we need to upgrade the ecore version? so apps that need this function can check for the right version?
Thanks
Dave
SVN revision: 34889
Currently debian/rules requires these dependancies to else
the following errors occur:
When automake is missing:
./autogen.sh
[...]
Running aclocal...
./autogen.sh: line 8: aclocal: command not found
make: *** [clean] Error 1
./autogen.sh
[...]
configure.in:20: error: possibly undefined macro: AC_LIBTOOL_WIN32_DLL
If this token and others are legitimate, please use m4_pattern_allow.
See the Autoconf documentation.
configure.in:23: error: possibly undefined macro: AC_PROG_LIBTOOL
make: *** [clean] Error 1
When pkg-config is missing:
./autogen.sh
[...]
configure: creating libtool
./configure: line 14491: PKG_PROG_PKG_CONFIG: command not found
./configure: line 14499: syntax error near unexpected token `EVIL,'
./configure: line 14499: ` PKG_CHECK_MODULES(EVIL, evil)'
make: *** [clean] Error 2
SVN revision: 34790
* src/lib/dlfcn/dlfcn.h:
* src/lib/mman/sys/mman.h:
define EFL_EVIL_MMAN_BUILD and EFL_EVIL_DLFCN_BUILD so
that functions in Evil.h are not declared as import
* src/lib/evil.c: (setenv), (unsetenv), (realpath),
(evil_sockets_init), (evil_sockets_shutdown), (evil_pipe),
(evil_tmpdir_get), (evil_homedir_get):
rename pipe() to evil_pipe().
set tmpdir to "Temp" and homedir to "My Documents" on WinCE
* src/lib/Evil.h:
define pipe() as evil_pipe(). This allow to use our pipe()
implementation to be usedwith cegcc, as it is broken on that
platform.
move the defines at the end of the file
SVN revision: 34787
add strdup define
* src/lib/evil.c: (evil_tmpdir_get), (evil_homedir_get):
include stdlib.h to define MB_CUR_MAX
On Windows CE, tmpdir and homedir are defined as "\windows"
* src/lib/mman/mman.c: (mmap), (munmap):
compile the big fat hack of mmap only on Windows CE
SVN revision: 34784
Valgrind just noticed that "writes_pending" was not being initialized,
probably because some old code used calloc() (based on the comment
at the allocation step).
SVN revision: 34777
* src/lib/evil.c: (evil_last_error_get):
add evil_last_error_get() function to get
useful error string creation.
* src/lib/dlfcn/dlfcn.c: (get_last_error), (dlopen), (dlsym),
(dladdr):
use UNICODE check instead of compiler checks when needed.
use evil_last_error_get()
* src/lib/mman/mman.c: (mmap), (munmap):
file mapping does not work on Windows CE < 5.0.
if it is the case, just read the file and return
the data.
use evil_last_error_get()
* src/lib/mman/Makefile.am:
* src/lib/Makefile.am:
add libevil dependancy
SVN revision: 34775
Needed for a use of ecore_wince without ecore_evas
* That stupid OS does not send a specific message when you show a window.
Force the SHOWWINDOW message when showing / hiding. It works with
CEgcc, but I don't know if it will work with vc++ for WinCE
SVN revision: 34740
This will check both function and data before removing the callback,
this is useful when you have lots of children monitoring parent, when
one child want to remove its monitoring function, others will remain.
Name is quite difficult to choose, I opted for "_full", but could be
"_with_data" or similar.
SVN revision: 34731
Image_Entry flag structure. This fix a bug with 16 bpp software engine.
* Change image loader module API to take any Image_Entry. Same goes
for evas_common_image_premul and evas_common_image_set_alpha_sparse.
* Use new eet API: eet_data_image_read_to_surface.
SVN revision: 34728
This functions provide a way to support tiles with eet. They also give the
possibility to directly decompress eet image inside an engine surface.
SVN revision: 34725
memory leak with cegcc
* src/lib/dlfcn/dlfcn.h:
include limits.h for MinGW users
* src/lib/mman/sys/mman.h:
add MAP_FAILED error value
SVN revision: 34717
totally done and may hand control back to a non rendering codepath - or
before u do any fp ops u are unsure of the cpu state beforehand. see the big
fat comment.
i was right. it was a missing emms.
SVN revision: 34707
* in ecore_x_pointer_xy_get, fill x & y with '-1', if the pointer is not in the same screen as the given window
* add 'same_screen' and 'root_win' information for the XEvents dealing with keyboard and mouse events
e:
* make sure e_manager_current_get returns the correct manager for the current screen
* fix various modules to use the correct window, when dealing with events
SVN revision: 34705