This reverts commit c7ca547b55.
cd710f5 fixed the issue with embedding into a parent that has alpha and
partially reverted c7ca547
this commit fully reverts c7ca547 , as these changes aren't needed for
embedding into an alpha-parent.
Before all the predated commands where kept in an array and their
indexes were used in bindings. This meant that users couldn't add their
own functions from the config file. Now key/mouse bindings have been
changed to to store the function ptr (wrapped in a cmd_t struct to also
store the mode) directly instead.
General cleanup done in this commit:
Defined `MODE_ALL` instead of using magic number.
For example, suppose one had bindings like:
{ 0, XK_q, g_quit, None },
{ ShitMask, XK_q, {quit_err}, None }
{ ControlMask, XK_q, {quit_err, .mode=MODE_IMAGE}, None }
The existing binding `q` has been left unchanged and is defined the same
way. However, the new hypothetical binding `Shift-q` can be used to call
the custom function quit_err in any mode (default). `Ctrl-q` on the
other hand will be called only on image mode.
Closes#50
Previously, the value of imgcursor was determined by where a pointer
binding was set to a ci_cursor_navigate. If it was then the pointer
would change to left/right arrows depending on the position relative to
the window. Now the user has full control of over it which also allows
them to preserve the behavior in case they wrap the function.
* Fix regression introduced in c7ca547 which made nsxiv not start in
non-TrueColor X server.
* Introduce a new fix for embedding into tabbed-alpha.
* Fixes a visual glitch from original sxiv when drawing transparent images in 8
bit depth. In 8 bit PseudoColor, `.pixel` is just an index into the 256
defined colors and thus trying to extract rgb bits from it would result in
visual glitch. The values `.color.red` on the other hand and so on are always
integers between 0 and 0xFFFF representing the color as expected.
* Use XColor for win_bg/fg and mrk_fg
Co-authored-by: NRK <nrk@disroot.org>
* remove duplicate comment
* remove empty tabs and blank lines
* move macros and globals ontop
* comment to seprate function implementation
* fix alignment
* switch to *argv[] similar to other suckless code
* kill all empty last lines
* append comment to endif
* reuse existing ARRLEN macro
* comment fall through
* use while (true) everywhere
Co-authored-by: NRK <nrk@disroot.org>
libXft and libfontconfig are now optional dependencies which can be
disabled via `HAVE_LIBFONTS=0`. Disabling them means disabling the
statusbar. This also does not search for freetype2 header if disabled.
Co-authored-by: NRK <nrk@disroot.org>
Currently when running the key-handler the statusbar shows a
"Running key-handler..." message, but there's no indication of the prefix key
being pressed.
There's a slight functional benefit of this patch in the sense
that users can visually tell if the key-handler is listening on input or if the
key-handler has been aborted or not.
* Remove non-POSIX extensions and commands
* Drop autodetection in favor of OPT_DEP_DEFAULT
* Use += for LDLIBS as some BSD distros need to add extra flags
* Change DOCPREFIX -> EGPREFIX
* Use ?= for MANPREFIX and EGPREFIX
* Update docs
With this, we should have a stable build system. No further significant
changes should be needed.
ten_ms needed to be a global but after the following commit
3724d3fc17 this no longer holds true.
it can simply be local to run, as it's not used anywhere else.
this does not need to be a fatal error.
if im is NULL we're going to load it with imlib2 anyways.
one other problem this solves is that before, due to the fatal error,
the tmpfile opened under /tmp wouldn't get cleaned up.
Closes: https://github.com/nsxiv/nsxiv/issues/69
currently, the key-handler will not receive the `S-` modifier if there's
a capital equivalent of that KEY.
if https://github.com/nsxiv/nsxiv/pull/78 is to be merged, then this
behaviour may change.
however as it currently stands, we should fix the comment. we can update
it later if needed.
Co-authored-by: Berke Kocaoğlu <berke.kocaoglu@metu.edu.tr>
currently the code-base doesn't make use of variable length array
despite being -std=c99. it was irresponsible of me to introduce VLA in
here.
since this function will be called quite often, i did not want to make
calls to malloc and free as they have some overhead.
512 should be sufficient enough and probably is far bigger than any
window title bar can display anyways.
currently the README only mentions what deps are optional but has no
info on what they do. we had an issue where a user was confused about
what libexif is used for : https://github.com/nsxiv/nsxiv/issues/58
this makes it clear what each of the optional deps do so that users can
make more informed decision on weather they want something or not.
* add -0 for outputting null-terminated list
this doesn't add much, if any, additional complexity to the codebase and
can be quite handy for scripting purposes.
Closes: https://github.com/nsxiv/nsxiv/issues/67
* Fix typo
Co-authored-by: Berke Kocaoğlu <berke.kocaoglu@metu.edu.tr>
Previous the code only allowed ShiftMask,ControlMask or Mod1Mask to be
used in keybindings and the presence of any others modifiers would be
ignored. Most problems generally allow certain modifiers to be be
ignored but not most and certainly don't allow Super-A to be treated
like A.
Now users can use any modifiers they want in keybindings and can also
ignore any modifiers they want. By default only ModMask2 (commonly
numlock is ignored)
Co-authored-by: N-R-K <79544946+N-R-K@users.noreply.github.com>
* Fix font memory leak.
This memory leak has always been present in sxiv.
The font opened on window.c:58 was never closed, so I closed it, fixing a 2kB memory leak.
* document changes
Co-authored-by: NRK <nrk@disroot.org>
If the user closed our window, the program won't automatically be die.
It may look dead as there would be no graphical indication that it was
running, but it still would be using/wasting the same resources.
Now the program will abruptly exit when its window is killed.
* Added simple configure script
Added simple script to autodetect if optional dependencies are installed
and enable/disable them as needed. Note this solution uses the compiler
directly instead of an external program like pkgconfig so it doesn't
require any extra dependencies. It is intended to work with any
arbitrary compiler; it has been tested with tcc and gcc.
There are some "breaking" changes hidden here
- HAVE_GIFLIB was renamed to HAVE_LIBGIF to match HAVE_LIBEXIF
- Simply typing `make` will no longer try to build with optional
dependencies
* use implicit RM variable
* General clean-up in configure
- Use printf instead of echo
- Format style
* honor env PREFIX, use LDLIBS
* Revert "General clean-up in configure"
This reverts commit 8683c179dbf273a330f9a224a4d481a7bea42c5f.
* honor env LDFLAGS if set
* Don't set OPTIONAL_LIBS in configure
* make OBJ all caps
* follow suckless style build system
- remove configure script.
- HAVE_LIBGIF and HAVE_LIBEXIF defaults back to 1
- unload several varibales onto config.mk
- make version all-caps
- add -O2 optimization
- use CPPFLAGS for includes and defines
* Revert "follow suckless style build system"
This reverts commit 8bf75b1f68d72df349edba8d998d4659dd956dd8.
* Generate config.mk from make
* Inlined configure in Makefile
* update docs
* cleanups
- changes to config.mk should trigger a rebuild
- remove potentially confusing variables form Makefile
* Use install instead of mkdir/cp/chmod when sensible
* fixup! Inlined configure in Makefile
* Don't generate config.mk on rm -f *.o sxiv
* update docs and cleanups
- make config.mk silent
- mention editing config.mk in README
* fallback to 0 if user edits config.mk in unexpected way
* add comment on config.mk
* remove invalid comment
configure script is removed
* slight restructure
- make version all caps
- restructure variables that users may want to edit to top
- use CPPFLAGS for defines
- add some comments
- remove needless echos since we have verbose output now
* add echos back
Co-authored-by: NRK <nrk@disroot.org>
Co-authored-by: Berke Kocaoğlu <berke.kocaoglu@metu.edu.tr>
