On Fri, Jun 04, 2021 at 04:43:36PM -0500, Zebediah Figura wrote:
libmpg123 is readily available in Mac OS package managers.
Brendan found a single-file MP3 decoder library which we could use in place of mpg123:
https://github.com/mackron/dr_libs/blob/master/dr_mp3.h
Alexandre said it'd be nice to explore this option so we can drop the mpg123 dependency for l3codeca and mp3dmod entirely and have one less thing to convert to PE.
Have any thoughts on that option?
Andrew
Signed-off-by: Zebediah Figura z.figura12@gmail.com
configure.ac | 2 +- dlls/l3codeca.acm/Makefile.in | 2 +- dlls/l3codeca.acm/mpegl3.c | 390 +--------------------------------- 3 files changed, 3 insertions(+), 391 deletions(-)
diff --git a/configure.ac b/configure.ac index fd266c6fb89..425f3541d94 100644 --- a/configure.ac +++ b/configure.ac @@ -1893,7 +1893,7 @@ then MPG123_LIBS="" fi]) fi -WINE_NOTICE_WITH(mpg123,[test "x$ac_cv_lib_mpg123_mpg123_feed" != xyes -a x"$ac_cv_header_CoreAudio_CoreAudio_h" != xyes], +WINE_NOTICE_WITH(mpg123,[test "x$ac_cv_lib_mpg123_mpg123_feed" != xyes], [libmpg123 ${notice_platform}development files not found (or too old), mp3 codec won't be supported.], [enable_l3codeca_acm]) test "x$ac_cv_lib_mpg123_mpg123_feed" = xyes || enable_mp3dmod=${enable_mp3dmod:-no} diff --git a/dlls/l3codeca.acm/Makefile.in b/dlls/l3codeca.acm/Makefile.in index 4f946ed6542..d46523a557a 100644 --- a/dlls/l3codeca.acm/Makefile.in +++ b/dlls/l3codeca.acm/Makefile.in @@ -1,7 +1,7 @@ MODULE = l3codeca.acm IMPORTS = winmm user32 EXTRAINCL = $(MPG123_CFLAGS) -EXTRALIBS = $(MPG123_LIBS) $(COREAUDIO_LIBS) +EXTRALIBS = $(MPG123_LIBS)
C_SRCS = \ mpegl3.c diff --git a/dlls/l3codeca.acm/mpegl3.c b/dlls/l3codeca.acm/mpegl3.c index 2e929371ae8..e625a75b7ac 100644 --- a/dlls/l3codeca.acm/mpegl3.c +++ b/dlls/l3codeca.acm/mpegl3.c @@ -3,8 +3,6 @@
- Copyright (C) 2002 Eric Pouech
- Copyright (C) 2009 CodeWeavers, Aric Stewart
- Copyright (C) 2010 Kristofer Henriksson
- This library is free software; you can redistribute it and/or
- modify it under the terms of the GNU Lesser General Public
@@ -28,17 +26,7 @@ #include <stdarg.h> #include <string.h>
-#ifdef HAVE_MPG123_H -# include <mpg123.h> -#else -# ifdef HAVE_COREAUDIO_COREAUDIO_H -# include <CoreFoundation/CoreFoundation.h> -# include <CoreAudio/CoreAudio.h> -# endif -# ifdef HAVE_AUDIOTOOLBOX_AUDIOCONVERTER_H -# include <AudioToolbox/AudioConverter.h> -# endif -#endif +#include <mpg123.h>
#include "windef.h" #include "winbase.h" @@ -123,8 +111,6 @@ static DWORD MPEG3_GetFormatIndex(LPWAVEFORMATEX wfx) return 0xFFFFFFFF; }
-#ifdef HAVE_MPG123_H
typedef struct tagAcmMpeg3Data { void (*convert)(PACMDRVSTREAMINSTANCE adsi, @@ -298,380 +284,6 @@ static LRESULT MPEG3_StreamClose(PACMDRVSTREAMINSTANCE adsi) return MMSYSERR_NOERROR; }
-#elif defined(HAVE_AUDIOTOOLBOX_AUDIOCONVERTER_H)
-static const unsigned short Mp3BitRates[2][16] = -{
- {0, 32, 40, 48, 56, 64, 80, 96, 112, 128, 160, 192, 224, 256, 320, 0},
- {0, 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160, 0}
-};
-static const unsigned short Mp3SampleRates[2][4] = -{
- {44100, 48000, 32000, 0},
- {22050, 24000, 16000, 0}
-};
-typedef struct tagAcmMpeg3Data -{
- LRESULT (*convert)(PACMDRVSTREAMINSTANCE adsi, unsigned char*,
LPDWORD, unsigned char*, LPDWORD);
- AudioConverterRef acr;
- AudioStreamBasicDescription in,out;
- AudioBufferList outBuffer;
- AudioBuffer inBuffer;
- SInt32 tagBytesLeft;
- UInt32 NumberPackets;
- AudioStreamPacketDescription *PacketDescriptions;
-} AcmMpeg3Data;
-/***********************************************************************
MPEG3_drvOpen
- */
-static LRESULT MPEG3_drvOpen(LPCSTR str) -{
- return 1;
-}
-/***********************************************************************
MPEG3_drvClose
- */
-static LRESULT MPEG3_drvClose(DWORD_PTR dwDevID) -{
- return 1;
-}
-/*
- When it asks for data, give it all we have. If we have no data, we assume
- we will in the future, so give it no packets and return an error, which
- signals that we will have more later.
- */
-static OSStatus Mp3AudioConverterComplexInputDataProc(
- AudioConverterRef inAudioConverter,
- UInt32 *ioNumberDataPackets,
- AudioBufferList *ioData,
- AudioStreamPacketDescription **outDataPacketDescription,
- void *inUserData
-) -{
- AcmMpeg3Data *amd = (AcmMpeg3Data*)inUserData;
- if (amd->inBuffer.mDataByteSize > 0)
- {
*ioNumberDataPackets = amd->NumberPackets;
ioData->mNumberBuffers = 1;
ioData->mBuffers[0] = amd->inBuffer;
amd->inBuffer.mDataByteSize = 0;
if (outDataPacketDescription)
*outDataPacketDescription = amd->PacketDescriptions;
return noErr;
- }
- else
- {
*ioNumberDataPackets = 0;
return -74;
- }
-}
-/*
- Get the length of the current frame. We need to be at the start of a
- frame now. The buffer must have at least the four bytes for the header.
- */
-static SInt32 Mp3GetPacketLength(const unsigned char* src) -{
- unsigned char mpegv;
- unsigned short brate, srate;
- unsigned int size;
- /*
Check that our position looks like an MP3 header and see which type
of MP3 file we have.
*/
- if (src[0] == 0xff && src[1] >> 1 == 0x7d) mpegv = 0; /* MPEG-1 File */
- else if (src[0] == 0xff && src[1] >> 1 == 0x79) mpegv = 1; /* MPEG-2 File */
- else return -1;
- /* Fill in bit rate and sample rate. */
- brate = Mp3BitRates[mpegv][(src[2] & 0xf0) >> 4];
- srate = Mp3SampleRates[mpegv][(src[2] & 0xc) >> 2];
- /* Certain values for bit rate and sample rate are invalid. */
- if (brate == 0 || srate == 0) return -1;
- /* Compute frame size, round down */
- size = 72 * (2 - mpegv) * brate * 1000 / srate;
- /* If there is padding, add one byte */
- if (src[2] & 0x2) return size + 1;
- else return size;
-}
-/*
- Apple's AudioFileStream does weird things so we deal with parsing the
- file ourselves. It was also designed for a different use case, so this
- is not unexpected. We expect to have MP3 data as input (i.e. we can only
- deal with MPEG-1 or MPEG-2 Layer III), which simplifies parsing a bit. We
- understand the ID3v2 header and skip over it. Whenever we have data we
- want to skip at the beginning of the input, we do this by setting *ndst=0
- and *nsrc to the length of the unwanted data and return no error.
- */
-static LRESULT mp3_leopard_horse(PACMDRVSTREAMINSTANCE adsi,
unsigned char* src, LPDWORD nsrc,
unsigned char* dst, LPDWORD ndst)
-{
- OSStatus err;
- UInt32 size, aspdi, synci, syncSkip;
- short framelen[4];
- const unsigned char* psrc;
- AcmMpeg3Data* amd = (AcmMpeg3Data*)adsi->dwDriver;
- TRACE("ndst %u %p <- %u %p\n", *ndst, dst, *nsrc, src);
- TRACE("First 16 bytes to input: %s\n", wine_dbgstr_an((const char *)src, 16));
- /* Parse ID3 tag */
- if (!memcmp(src, "ID3", 3) && amd->tagBytesLeft == -1)
- {
amd->tagBytesLeft = (src[6] << 21) + (src[7] << 14) + (src[8] << 7) + src[9];
if (src[5] & 0x10) amd->tagBytesLeft += 20; /* There is a footer */
else amd->tagBytesLeft += 10;
- }
- /* Consume the tag */
- if (amd->tagBytesLeft >= (SInt32)*nsrc)
- {
*ndst = 0;
amd->tagBytesLeft -= *nsrc;
TRACE("All %d bytes of source data is ID3 tag\n", *nsrc);
return MMSYSERR_NOERROR;
- }
- else if (amd->tagBytesLeft > 0)
- {
src += amd->tagBytesLeft;
*nsrc -= amd->tagBytesLeft;
TRACE("Skipping %ld for ID3 tag\n", amd->tagBytesLeft);
- }
- /*
Sync to initial MP3 frame. The largest possible MP3 frame is 1440.
Thus, in the first 1440 bytes we must find the beginning of 3 valid
frames in a row unless we reach the end of the file first.
*/
- syncSkip = 0;
- for (psrc = src; psrc <= src + *nsrc - 4 && psrc < src + 1440; psrc++)
- {
framelen[0] = 0;
for (synci = 1;
synci < 4 && psrc + framelen[synci-1] < src + *nsrc - 4;
synci++)
{
framelen[synci] = Mp3GetPacketLength(psrc + framelen[synci-1]);
if (framelen[synci] == -1)
{
synci = 0;
break;
}
framelen[synci] += framelen[synci-1];
}
if (synci > 0) /* We synced successfully */
{
if (psrc - src > 0)
{
syncSkip = psrc - src;
src += syncSkip;
*nsrc -= syncSkip;
TRACE("Skipping %ld for frame sync\n", syncSkip);
}
break;
}
- }
- if (Mp3GetPacketLength(src) == -1)
- {
*ndst = *nsrc = 0;
ERR("Frame sync failed. Cannot play file.\n");
return MMSYSERR_ERROR;
- }
- /*
Fill in frame descriptions for all frames. We use an extra pointer
to keep track of our position in the input.
*/
- amd->NumberPackets = 25; /* This is the initial array capacity */
- amd->PacketDescriptions = HeapAlloc(GetProcessHeap(), 0, amd->NumberPackets * sizeof(AudioStreamPacketDescription));
- if (amd->PacketDescriptions == 0) return MMSYSERR_NOMEM;
- for (aspdi = 0, psrc = src;
psrc <= src + *nsrc - 4;
psrc += amd->PacketDescriptions[aspdi].mDataByteSize, aspdi++)
- {
/* Return an error if we can't read the frame header */
if (Mp3GetPacketLength(psrc) == -1)
{
*ndst = *nsrc = 0;
ERR("Invalid header at %p.\n", psrc);
HeapFree(GetProcessHeap(), 0, amd->PacketDescriptions);
return MMSYSERR_ERROR;
}
/* If we run out of space, double size and reallocate */
if (aspdi >= amd->NumberPackets)
{
amd->NumberPackets *= 2;
amd->PacketDescriptions = HeapReAlloc(GetProcessHeap(), 0, amd->PacketDescriptions, amd->NumberPackets * sizeof(AudioStreamPacketDescription));
if (amd->PacketDescriptions == 0) return MMSYSERR_NOMEM;
}
/* Fill in packet data */
amd->PacketDescriptions[aspdi].mStartOffset = psrc - src;
amd->PacketDescriptions[aspdi].mVariableFramesInPacket = 0;
amd->PacketDescriptions[aspdi].mDataByteSize = Mp3GetPacketLength(psrc);
/* If this brings us past the end, the last one doesn't count */
if (psrc + amd->PacketDescriptions[aspdi].mDataByteSize > src + *nsrc) break;
- }
- /* Fill in correct number of frames */
- amd->NumberPackets = aspdi;
- /* Adjust nsrc to only include full frames */
- *nsrc = psrc - src;
- amd->inBuffer.mDataByteSize = *nsrc;
- amd->inBuffer.mData = src;
- amd->inBuffer.mNumberChannels = amd->in.mChannelsPerFrame;
- amd->outBuffer.mNumberBuffers = 1;
- amd->outBuffer.mBuffers[0].mDataByteSize = *ndst;
- amd->outBuffer.mBuffers[0].mData = dst;
- amd->outBuffer.mBuffers[0].mNumberChannels = amd->out.mChannelsPerFrame;
- /* Convert the data */
- size = amd->outBuffer.mBuffers[0].mDataByteSize / amd->out.mBytesPerPacket;
- err = AudioConverterFillComplexBuffer(amd->acr, Mp3AudioConverterComplexInputDataProc, amd, &size, &amd->outBuffer, 0);
- HeapFree(GetProcessHeap(), 0, amd->PacketDescriptions);
- /* Add skipped bytes back into *nsrc */
- if (amd->tagBytesLeft > 0)
- {
*nsrc += amd->tagBytesLeft;
amd->tagBytesLeft = 0;
- }
- *nsrc += syncSkip;
- if (err != noErr && err != -74)
- {
*ndst = *nsrc = 0;
ERR("Feed Error: %ld\n", err);
return MMSYSERR_ERROR;
- }
- *ndst = amd->outBuffer.mBuffers[0].mDataByteSize;
- TRACE("convert %d -> %d\n", *nsrc, *ndst);
- return MMSYSERR_NOERROR;
-}
-/***********************************************************************
MPEG3_Reset
- */
-static void MPEG3_Reset(PACMDRVSTREAMINSTANCE adsi, AcmMpeg3Data* aad) -{
- AudioConverterReset(aad->acr);
-}
-/***********************************************************************
MPEG3_StreamOpen
- */
-static LRESULT MPEG3_StreamOpen(PACMDRVSTREAMINSTANCE adsi) -{
- AcmMpeg3Data* aad;
- assert(!(adsi->fdwOpen & ACM_STREAMOPENF_ASYNC));
- if (MPEG3_GetFormatIndex(adsi->pwfxSrc) == 0xFFFFFFFF ||
MPEG3_GetFormatIndex(adsi->pwfxDst) == 0xFFFFFFFF)
return ACMERR_NOTPOSSIBLE;
- aad = HeapAlloc(GetProcessHeap(), 0, sizeof(AcmMpeg3Data));
- if (aad == 0) return MMSYSERR_NOMEM;
- adsi->dwDriver = (DWORD_PTR)aad;
- if ((adsi->pwfxSrc->wFormatTag == WAVE_FORMAT_MPEGLAYER3 ||
adsi->pwfxSrc->wFormatTag == WAVE_FORMAT_MPEG) &&
adsi->pwfxDst->wFormatTag == WAVE_FORMAT_PCM)
- {
OSStatus err;
aad->in.mSampleRate = adsi->pwfxSrc->nSamplesPerSec;
aad->out.mSampleRate = adsi->pwfxDst->nSamplesPerSec;
aad->in.mBitsPerChannel = adsi->pwfxSrc->wBitsPerSample;
aad->out.mBitsPerChannel = adsi->pwfxDst->wBitsPerSample;
aad->in.mFormatID = kAudioFormatMPEGLayer3;
aad->out.mFormatID = kAudioFormatLinearPCM;
aad->in.mChannelsPerFrame = adsi->pwfxSrc->nChannels;
aad->out.mChannelsPerFrame = adsi->pwfxDst->nChannels;
aad->in.mFormatFlags = 0;
aad->out.mFormatFlags = kLinearPCMFormatFlagIsSignedInteger;
aad->in.mBytesPerFrame = 0;
aad->out.mBytesPerFrame = (aad->out.mBitsPerChannel * aad->out.mChannelsPerFrame) / 8;
aad->in.mBytesPerPacket = 0;
aad->out.mBytesPerPacket = aad->out.mBytesPerFrame;
aad->in.mFramesPerPacket = 0;
aad->out.mFramesPerPacket = 1;
aad->in.mReserved = aad->out.mReserved = 0;
aad->tagBytesLeft = -1;
aad->convert = mp3_leopard_horse;
err = AudioConverterNew(&aad->in, &aad->out, &aad->acr);
if (err != noErr)
{
ERR("Create failed: %ld\n", err);
}
else
{
MPEG3_Reset(adsi, aad);
return MMSYSERR_NOERROR;
}
- }
- HeapFree(GetProcessHeap(), 0, aad);
- adsi->dwDriver = 0;
- return MMSYSERR_NOTSUPPORTED;
-}
-/***********************************************************************
MPEG3_StreamClose
- */
-static LRESULT MPEG3_StreamClose(PACMDRVSTREAMINSTANCE adsi) -{
- AcmMpeg3Data* amd = (AcmMpeg3Data*)adsi->dwDriver;
- AudioConverterDispose(amd->acr);
- HeapFree(GetProcessHeap(), 0, amd);
- adsi->dwDriver = 0;
- return MMSYSERR_NOERROR;
-}
-#endif
/***********************************************************************
MPEG3_DriverDetails
-- 2.30.2