[MPlayer-cvslog] r27759 - trunk/stream/cache2.c

[Reimar Döffinger]

On Tue, Oct 14, 2008 at 09:06:32AM -0400, compn wrote:
> On Mon, 13 Oct 2008 22:53:55 +0200, Diego Biurrun wrote:
> >On Mon, Oct 13, 2008 at 06:38:00PM +0200, Reimar Döffinger wrote:
> >> On Mon, Oct 13, 2008 at 05:51:29PM +0200, diego wrote:
> >> Hell no! You can't just remove the WINAPI without risking hell to break
> >> loose (i.e. crashes). And yes the Windows ThreadProc does have a return
> >> value, we wer just missing the return statement.
> >
> >I guess the root problem is that we are missing some header then?  Why
> >do we need the forward declaration?

Huh? The forward declaration is needed because the function is used
before the declaration.
And no it is of course not in any header, this is a custom function.
Someone went with the MSDN docu examples and called ThreadProc, but they
could just as well have called it PinkBunny.

> a quick google returns this: http://support.microsoft.com/kb/189996

A better one is this:
http://msdn.microsoft.com/en-us/library/ms686736(VS.85).aspx
Which (while I can't really imagine why) also says:
"Do not declare this callback function with a void return type and cast
the function pointer to LPTHREAD_START_ROUTINE when creating the thread.
Code that does this is common, but it can crash on 64-bit Windows."

Does not exactly affect us yet, but I hope it is now absolutely clear
that your commit was nonsense.
Now, there is on the other hand the question why we do not just use the
code for OS/2, except that manually setting the stack size to 256 kB is
questionable it should work on Windows, too.

Greetings,
Reimar Döffinger