Media Player Classic - BE Win32/x64

[chros]

Originally Posted by v0lt

MPC-BE 1.5.4 (билд 4969) stable

Thank you guys for the new release!
I have a small feature request (since there's no possibility to run external commands from the player):
can you add different exit codes for all the events in "After playback" menu? (exit, standby, logoff, etc)

I want to run commands in a batch script *only* when exit or standby was selected in the player. Consider the following script:

REM Script to run mpc-be and commands after it exited
cd d:\Progs\MPC-BE\

REM Start mpc-be
mpc-be64.exe

REM Turn off TV when mpc-be exited via After Playback Exit (exit code: 11)
IF %ERRORLEVEL% EQU 11 (
    aiopylgtvcommand 192.168.1.78 power_off
)

Thanks!

[Klaus1189]

Updated german translation based on r4971

Copy image to clipboard

[Liisachan]

Devs: Thank you for the new official release!
4971 is simply defining IDS_AG_COPY_IMAGE, and the actual command is not yet implemented?

[v0lt]

Liisachan
Copying works for me. I checked in IrfanView, Paint, Paint.NET and LibreOffice Writer.
MPC-BE 1.5.5.4972 beta x64.

[Liisachan]

@v0lt: confirmed
It's been added in r4971, where changed files are listed in 3 pages. I didn't notice that and only saw the first page, so I was confused. Besides, the newest nightly was 4970 when I posted previously.

For those who'd like to test this new feature: the command is currently not as a menu item under "File". Instead, it's in Options | Key, between Save Image and Save thumbnail. You can assign a shortcut you like (e.g. Ctrl+Alt+I) to it, then it works with that short cut key! In my test not only paused, but you can copy the frame to Clipboard while playing the video (it necessary).

So far it works very well. I'm so happy! Thanks very much!!!


PS: How this is convenient (just a few examples):
- You can directly paste the frame image to your image editor, when you want to trim it, when you want to insert text, or when you want to save it as JPEG with various options (Quality, Progress or not, Enable or disable chroma subsampling), etc. etc.

PS: MainFrm.cpp random observation
1) // Lock the handle and copy the text to the buffer <-- not “the text”?
2) WCHAR buff[256] has 512 bytes; len = _countof(buff) is 256; memset(buff, 0, len) doesn't zero-init the whole buffer?

[wushantao]

https://send.firefox.com/download/b6...0DLspeMKzwByEg

simplified chinese is coming

btw, some minor fixes

[v0lt]

Thanks to all.

@Klaus1189
Updated in r4972.

@wushantao
Updated in r4974.

@ryrynz
Typo "Dran'n'drop" corrected in r4976.

@Liisachan
Typo "text" corrected in r4975.
I did not find the incorrect "memset (buff, 0, len)" in MainFrm.cpp. What is the line number you're talking about?

[Manni]

Thank you for the new release and for all the work done on this project.

[ryrynz]

Originally Posted by v0lt

@ryrynz
Typo "Dran'n'drop" corrected in r4976.

Cheers, could you fix up the first post in the thread with the typo as well?

[v0lt]

ryrynz
Thanks. Fixed.

[Klaus1189]

Talking about cosmetics, I think it would be nice, if the screenshot of MPC-BE in the first post in this thread would show the actual stable release version, for now 1.5.4.

[Liisachan]

@v0lt
Sorry I was not clear. I was looking at IsRendererCompatibleWithSaveImage in
https://sourceforge.net/p/mpcbe/code...rc/MainFrm.cpp
and thought that this might be potentially bad:

		WCHAR buff[256];
		ULONG len = _countof(buff);
		memset(buff, 0, len);

[cmhrky]

Latest Turkish translation.

https://www.upload.ee/files/10907411...tr.rc.txt.html

[nevcairiel]

Originally Posted by Liisachan

@v0lt
Sorry I was not clear. I was looking at IsRendererCompatibleWithSaveImage in
https://sourceforge.net/p/mpcbe/code...rc/MainFrm.cpp
and thought that this might be potentially bad:

		WCHAR buff[256];
		ULONG len = _countof(buff);
		memset(buff, 0, len);

While correct that it only inits half the buffer, the entire memset could be removed, its not necessary here. CRegKey::QueryStringValue automatically inserts a null-terminator into any retrieved strings.

[Liisachan]

Originally Posted by nevcairiel

While correct that it only inits half the buffer, the entire memset could be removed, its not necessary here. CRegKey::QueryStringValue automatically inserts a null-terminator into any retrieved strings.

If it's just a wrapper of RegQueryValueExW, the terminating NULL is not guaranteed (according to the docs), meaning that even if you zero-init the whole buffer and ERROR_SUCCESS is returned, buff might be still not NULL-terminated (in an unlikely situation where the stored string is not NULL-terminated and has exactly 256 characters). In reality, the version string should be much shorter, though.

[Aleksoid1978]

Originally Posted by Liisachan

If it's just a wrapper of RegQueryValueExW, the terminating NULL is not guaranteed (according to the docs), meaning that even if you zero-init the whole buffer and ERROR_SUCCESS is returned, buff might be still not NULL-terminated (in an unlikely situation where the stored string is not NULL-terminated and has exactly 256 characters). In reality, the version string should be much shorter, though.

I don't know about docs - but call CRegKey::QueryStringValue() always return NULL terminated data.

[v0lt]

MPC Video Renderer v0.3.8.1060
Description.

Changes:
Improved compatibility with third-party players.
Fixed switching supported formats for DXVA2 and D3D11 video processors during playback.
Faster statistics rendering. A pre-prepared texture with symbols is used.

[JNW]

What's happened to this link? It's the one I always used
https://yadi.sk/d/r7JYQ2NjHvE4lA

[Liisachan]

@Aleksoid1978
That must be the case (nevcairiel said so too). I wouldn't doubt any of you guys For the record, the docs:

(1) RegQueryValueExW: “the string may not have been stored with the proper terminating null characters. [...] the application should ensure that the string is properly terminated” (inconvenient)
(2) RegGetValueW: “If the data is a string, the function checks for a terminating null character. If one is not found, the string is stored with a null terminator if the buffer is large enough” (nicer)

REG_SZ data without a terminator exists and in old days programmers were often responsible for dealing with that, as in (1). Perhaps that's why the original author of that function put memset there just in case. In this specific case, the value to be read seems to be the path to madVR.ax, and by definition a path ends with a terminator, so there's no problem at all anyway. If anything, the buffer length 260 (MAX_PATH) might be slightly more reasonable than 256.

[ryrynz]

Originally Posted by JNW

What's happened to this link? It's the one I always used
https://yadi.sk/d/r7JYQ2NjHvE4lA

Gone forever. Use this https://yadi.sk/d/AjAXDDHtHRIELg