MINGW32 + performencs ( was Re: [MPlayer-dev-eng] [PATCH] directx, typo, fs )

[Raindel Shachar]

On Wed, 2 Apr 2003, Joey Parrish wrote:

> Hello,
>
> Attached below is a patch for vo_directx.c.It's a patch I've just
> added to my cygwin packages, and it seemed clean enough to submit for
> approval.It does the following:
>
> 1) moves the fullscreen toggling code into a function called toggle_fs
> 2) fixes a small typo in the keystroke handling code
> 3) when directx window hears ALT+ENTER, toggles fullscreen status
>
> I could have done this by pushing 'f' to mplayer's input buffer, but
> since a user may have other keybindings for 'f', this seemed bad.I
> tried to figure out how to add support for ALT+whatever in keycodes.h,
> but the only such keys I found to have unique codes are ALT+0 through
> ALT+9 (0x140-0x149, IIRC).All others generated two codes in
> succession, such as ALT+k producing 'k',0x1b for any letter like 'k'.
> Because of this, I just decided that having ALT+ENTER for fullscreen
> could probably be okay as a directx-specific feature.(Directx, unlike
> linux console, gives a unique way to identify this.)
>
> It's a common keybinding for windows video players IMHO, so I'm going
> to include it in my packages anyway for the sake of those who've never
> used mplayer on linux and wouldn't know to hit 'f'.I hope you find
> this useful enough to commit, and if not I'll still have it in my
> package's patch set, no hurt feelings.
>
> -----
> Off-topic, but related to cygwin packages:
>
> I haven't been running the main branch on cygwin.Does anyone notice if
> there is a performance boost when using osdep/timer-win over
> linux/timer-lx undercygwin?Would it be worth patching my 0_90 sources
> to backport it?
>
> And I've also noticed that a few ifdef MINGW lines have popped up in
> main branch.Is mplayer getting closer to being able to compile with
> only mingw?I haven't tried it in months.
>
> Thanks,
> --Joey
>

About MINGW32 - I have been able, after about half of day of work, to run
Mplayer under MINGW32, but the input handeling, the caching layer,
streming and the bmovl vf are broken (as well as all of the stuff that is
not available under cygwin). I hope to fix this in this weekend, and send
patch which will allow compiling it under MINGW32.

About the performences issue: I have noticed that muting Mplayer's output
(or redirecting it to a file instead of a cygwin console) has significally
improved mplayer's performences on my system (I use win 98 for testing
cygwin/MINGW32 builds) when the console window was in the background.

Regards
--Shachar