ReceiveTempSMS
/
Sweden-Number
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

mciwave: Implement Cue as no-op. 

Tests show that it must not close the opened wave file.
This commit is contained in:
Jörg Höhle 2009-10-19 18:31:57 +02:00 committed by Alexandre Julliard
parent 2dfd3e901b
commit 096b774db2
2 changed files with 29 additions and 33 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

58
dlls/mciwave/mciwave.c
View File

@ -562,46 +562,38 @@ static LRESULT WAVE_mciOpen(MCIDEVICEID wDevID, DWORD dwFlags, LPMCI_WAVE_OPEN_P
/**************************************************************************
* WAVE_mciCue [internal]
*/
static DWORD WAVE_mciCue(MCIDEVICEID wDevID, LPARAM dwParam, LPMCI_GENERIC_PARMS lpParms)
static DWORD WAVE_mciCue(MCIDEVICEID wDevID, DWORD dwFlags, LPMCI_GENERIC_PARMS lpParms)
{
/*
FIXME
This routine is far from complete. At the moment only a check is done on the
MCI_WAVE_INPUT flag. No explicit check on MCI_WAVE_OUTPUT is done since that
is the default.
The flags MCI_NOTIFY (and the callback parameter in lpParms) and MCI_WAIT
are ignored
*/
DWORD dwRet;
WINE_MCIWAVE* wmw = WAVE_mciGetOpenDev(wDevID);
FIXME("(%u, %08lX, %p); likely to fail\n", wDevID, dwParam, lpParms);
TRACE("(%u, %08X, %p);\n", wDevID, dwFlags, lpParms);
if (wmw == NULL) return MCIERR_INVALID_DEVICE_ID;
/* Tests on systems without sound drivers show that Cue, like
* Record and Play, opens winmm, returning MCIERR_WAVE_xyPUTSUNSUITABLE.
* The first Cue Notify does not immediately return the
* notification, as if a player or recorder thread is started.
* PAUSE mode is reported when successful, but this mode is
* different from the normal Pause, because a) Pause then returns
* NONAPPLICABLE_FUNCTION instead of 0 and b) Set Channels etc. is
* still accepted, returning the original notification as ABORTED.
* I.e. Cue allows subsequent format changes, unlike Record or
* Open file, closes winmm if the format changes and stops this
* thread.
* Wine creates one player or recorder thread per async. Play or
* Record command. Notification behaviour suggests that MS-W*
* reuses a single thread to improve response times. Having Cue
* start this thread early helps to improve Play/Record's initial
* response time. In effect, Cue is a performance hint, which
* justifies our almost no-op implementation.
*/
/* FIXME */
/* always close elements ? */
if (wmw->hFile != 0) {
mmioClose(wmw->hFile, 0);
wmw->hFile = 0;
}
if (wmw == NULL) return MCIERR_INVALID_DEVICE_ID;
if (wmw->dwStatus != MCI_MODE_STOP) return MCIERR_NONAPPLICABLE_FUNCTION;
dwRet = MMSYSERR_NOERROR; /* assume success */
if ((dwFlags & MCI_NOTIFY) && lpParms)
WAVE_mciNotify(lpParms->dwCallback,wmw,MCI_NOTIFY_SUCCESSFUL);
if ((dwParam & MCI_WAVE_INPUT) && !wmw->fInput) {
dwRet = waveOutClose(wmw->hWave);
if (dwRet != MMSYSERR_NOERROR) return MCIERR_INTERNAL;
wmw->fInput = TRUE;
} else if (wmw->fInput) {
dwRet = waveInClose(wmw->hWave);
if (dwRet != MMSYSERR_NOERROR) return MCIERR_INTERNAL;
wmw->fInput = FALSE;
}
wmw->hWave = 0;
return (dwRet == MMSYSERR_NOERROR) ? 0 : MCIERR_INTERNAL;
return MMSYSERR_NOERROR;
}
/**************************************************************************

4
dlls/winmm/tests/mci.c
View File

@ -478,6 +478,10 @@ static void test_asyncWAVE(HWND hwnd)
test_notification(hwnd,"pause aborted play #1 notification",MCI_NOTIFY_SUPERSEDED);
test_notification(hwnd,"impossible pause notification",0);
err = mciSendString("cue mysound output notify", NULL, 0, hwnd);
ok(err==MCIERR_NONAPPLICABLE_FUNCTION,"mci cue output while paused returned error: %d\n", err);
test_notification(hwnd,"cue output notify #2",0);
/* Seek or even Stop used to hang Wine on MacOS. */
err = mciSendString("seek mysound to 0 wait", NULL, 0, NULL);
ok(!err,"mci seek to start returned error: %d\n", err);