- help (-fstype help) also availabible
- support state BELOW (someone may want to use it...) and by -fstype none forcing of
not changing window layer (user request)
- drop icelayer option, it can be set by -fstype layer=<number>
- simplify vo_x11_fullscreen
- fs change code cleanup
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9318 b3059339-0415-0410-9bf9-f77b7e298cf2
output and -wid (>0) parameter is this:
Mplayer by default creates a colormap using DirectColor visual. If the
window given to mplayer uses TrueColor visual there will be an error
when mplayer sets the colormap for the window. This patch
modifies mplayer to use TrueColor visual if the window given to mplayer
uses TrueColor. Another solution is to make sure that the window given to
mplayer is created using DirectColor visual if it is supported by the
display.
Jouni Tulkki <jitulkki@cc.hut.fi>
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9279 b3059339-0415-0410-9bf9-f77b7e298cf2
I don't understand very well why my fix works.
Jouni Tulkki <jitulkki@cc.hut.fi>
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9116 b3059339-0415-0410-9bf9-f77b7e298cf2
instead of |\ it should be /|
patch by {asm} <tf198@hszk.bme.hu>
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9110 b3059339-0415-0410-9bf9-f77b7e298cf2
The problem was that the new drawing method assumes a linear
framebuffer, which is not available in those modes. This can be worked
around by using the old drawing method, which is what this patch does.
The old method can be forced, by using the "old" driver option.
This patch also enables linear addressing, since it improves write speed
to video memory considerably. The mentioned problem:
"it is not compatable with vga_draw* for some cards"
Is a bug in svgalib, which I think should be fixed in recent svgalib
versions. If someone sees this problem, please report to svgalib
maintainer (that's me).
patch by Matan Ziv-Av. matan@svgalib.org
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9047 b3059339-0415-0410-9bf9-f77b7e298cf2
* Get layers by name rather that fixed ID. Only works for DirectFB >=
0.9.16 so fall back to ID based stuff for older releases.
* New option "noinput" (useful in X)
* Add some NULL pointer checks to uninit()
* Improve error messages and use mp_msg()
* Direct rendering fix (at least IPB was broken)
* Use 0xff for alpha component in all colors
* Blit from temp surface to screen was only done in draw_osd() which
doesn't always get called. Do the blit in flip_page() in this case.
* Add GUI support to control()
* Change the driver name to include G450/G550 ;)
patch by Ville Syrjälä <syrjala@sci.fi>
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@9022 b3059339-0415-0410-9bf9-f77b7e298cf2
Works on icemw and fvwm, but not on windowmaker (bug in wmaker...).
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8867 b3059339-0415-0410-9bf9-f77b7e298cf2
- moved global init/uninit code to preinit/uninit
- -fixed-vo support, still buggy (leak some memory and after a while it
permanently kills xv, only X restart helps... or just nvidia sucks again)
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8841 b3059339-0415-0410-9bf9-f77b7e298cf2
Fixed sawfish bahavior in gmplayer: use ABOVE, FULLSCREEN is last choice.
Now mplayer window is visible over the video window.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8836 b3059339-0415-0410-9bf9-f77b7e298cf2
It makes fonts keep scale when they are set to be proportional to width.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8820 b3059339-0415-0410-9bf9-f77b7e298cf2
Removed varaibles which last patch made unused (forgot to do it before).
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8784 b3059339-0415-0410-9bf9-f77b7e298cf2
Remove (added by me before) checking wem type at each fullscreen mode
change. It doesn't help at all in case of window manager change.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8743 b3059339-0415-0410-9bf9-f77b7e298cf2
adds 2 options: -keepaspect (default) and -nokeepaspect
tested with icewm and enlightenment
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8693 b3059339-0415-0410-9bf9-f77b7e298cf2
support for resizing. To maximize window we need to use
_NET_WM_STATE_FULLSCREEN, but it doesn't tell use that it supports it.
New IceWM also doesn't support layers in fullscreen mode, changing it
doesn't change window position (again, "-icelayer" doesn't fit well).
Also while fixing fvwm metacity got broken again, so there is a new
hack or it, and small fix in code for openbox.
Now it should work on almost everything.
Filip Kalinski <filon@pld.org.pl>
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8677 b3059339-0415-0410-9bf9-f77b7e298cf2
It includes ugly hack for OpenBox, swaps checking for netwm
with layers (and makes fvwm work -> they have also broken
_NET_WM_STATE_FULLSCREEN but also supports layers) and changes
back fullscreen layer to 10, afterstep doesn't like 12 :-)
Tested with: sawfish, metacity, kwin, blackbox, openbox, waimea,
windowmaker, afterstep, icewm, mwm, fvwm
Filip Kalinski <filon@pld.org.pl>
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8670 b3059339-0415-0410-9bf9-f77b7e298cf2
generally two types of layer support for window managers:
- NetWM states (FULLSCREEN, STAYS_ON_TOP, ABOVE)
sawfish, metacity, kwin
- _WIN_LAYER
IceWM, WindowMaker
So we don't need any other window manager detection functions, we need
only to check for these two.
Code tested on lots of windowmanagers.
patch by Filip Kalinski <filon@pld.org.pl>
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8661 b3059339-0415-0410-9bf9-f77b7e298cf2
based on patch by Dmitry Baryshkov <lumag@qnc.ru>
- default height changed from 1 to 4, seems to be more optimal and works
fast on both g400max & gf4ti4200
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8655 b3059339-0415-0410-9bf9-f77b7e298cf2
for "stupid BIOSes", but just the first buffer. So when using -double
option I got flicker with such a stupid BIOS...
patch by Aurelien JACOBS <aurel@gnuage.org>
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8653 b3059339-0415-0410-9bf9-f77b7e298cf2
After applying fullscreen (x11, xv +/- gui) works with at least:
- sawfish 1.2
- metacity
- kwin
- wmaker
- balckbox
- waimea
- icewm,
so I think it doesn't break anything.
Filip Kalinski <filon@pld.org.pl>
NOTE: we can still reverse it :))) --A'rpi
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8646 b3059339-0415-0410-9bf9-f77b7e298cf2
together happily. if an oldstyle subtitle (font.desc) is found, it will
be used. otherwise mplayer will choose subfont.ttf, if freetype was
detected during ./configure.
(also you can use -font /path/to/font.desc and -font /path/to/foobar.ttf too)
patch-set by Wojtek Kaniewski <wojtekka@bydg.pdi.net>
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8636 b3059339-0415-0410-9bf9-f77b7e298cf2
-sub-bkg-color n
-sub-bkg-alpha n
They control the color and alpha value used to initialize the subtitles and OSD BBOX.
With this you can have subtitles inside a traslucent rectangle.
This is useful when a movie already have "hardcoded" subtitles and you
want to overwrite them with rendered subtitles avoiding too much confusion.
patch by Salvador Eduardo Tropea <salvador@inti.gov.ar>
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8602 b3059339-0415-0410-9bf9-f77b7e298cf2
control. Possible values are top, center, and bottom,
with bottom being the default. Alignment is relevant when
it comes to positioning subtitles with one line (or fewer
lines) of text relative to multi-line subtitles.
It is implemented as a new command (sub_alignment) that
without an argument cycles the alignment (between top,
center, and bottom), or with an argument sets the
alignment (0 for top, 1 for center, 2 for bottom).
The key 'i' is bound to this command.
patch by Oskar Liljeblad (oskar@osk.mine.nu)
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@8535 b3059339-0415-0410-9bf9-f77b7e298cf2