static pANTLR3_INPUT_STREAM
antlr3CreateLineBufferedStream(std::istream& in)
{
// Pointer to the input stream we are going to create
//
pANTLR3_INPUT_STREAM input;
if (!in)
{
return NULL;
}
// Allocate memory for the input stream structure
//
input = (pANTLR3_INPUT_STREAM)
ANTLR3_CALLOC(1, sizeof(ANTLR3_LINE_BUFFERED_INPUT_STREAM));
if (input == NULL)
{
return NULL;
}
// Structure was allocated correctly, now we can install the pointer
//
input->data = malloc(1024);
input->isAllocated = ANTLR3_FALSE;
((pANTLR3_LINE_BUFFERED_INPUT_STREAM)input)->in = ∈
// Call the common 8 bit input stream handler
// initialization.
//
#ifdef CVC4_ANTLR3_OLD_INPUT_STREAM
antlr3AsciiSetupStream(input, ANTLR3_CHARSTREAM);
#else /* CVC4_ANTLR3_OLD_INPUT_STREAM */
antlr38BitSetupStream(input);
// In some libantlr3c 3.4-beta versions, this call is not included in the above.
// This is probably an erroneously-deleted line in the libantlr3c source since 3.2.
antlr3GenericSetupStream(input);
#endif /* CVC4_ANTLR3_OLD_INPUT_STREAM */
return input;
}
pANTLR3_INPUT_STREAM MemoryMappedInputBufferNew(const std::string& filename) {
// Pointer to the input stream we are going to create
//
pANTLR3_INPUT_STREAM input;
ANTLR3_UINT32 status;
// Allocate memory for the input stream structure
//
input = (pANTLR3_INPUT_STREAM) ANTLR3_CALLOC(1, sizeof(ANTLR3_INPUT_STREAM));
if(input == NULL) {
return NULL;
}
// Structure was allocated correctly, now we can read the file.
//
status = MemoryMapFile(input, filename);
// Call the common 8 bit ASCII input stream handler
// Initializer type thingy doobry function.
//
#ifdef CVC4_ANTLR3_OLD_INPUT_STREAM
antlr3AsciiSetupStream(input, ANTLR3_CHARSTREAM);
#else /* CVC4_ANTLR3_OLD_INPUT_STREAM */
antlr38BitSetupStream(input);
#endif /* CVC4_ANTLR3_OLD_INPUT_STREAM */
// Now we can set up the file name
//
input->istream->streamName
= input->strFactory->newStr(input->strFactory,
(uint8_t*) filename.c_str());
input->fileName = input->istream->streamName;
input->free = UnmapFile;
if(status != ANTLR3_SUCCESS) {
input->close(input);
return NULL;
}
return input;
}
static void
setupInputStream(pANTLR3_INPUT_STREAM input)
{
#if 0
ANTLR3_BOOLEAN isBigEndian;
// Used to determine the endianness of the machine we are currently
// running on.
//
ANTLR3_UINT16 bomTest = 0xFEFF;
// What endianess is the machine we are running on? If the incoming
// encoding endianess is the same as this machine's natural byte order
// then we can use more efficient API calls.
//
if (*((pANTLR3_UINT8)(&bomTest)) == 0xFE)
{
isBigEndian = ANTLR3_TRUE;
}
else
{
isBigEndian = ANTLR3_FALSE;
}
// What encoding did the user tell us {s}he thought it was? I am going
// to get sick of the questions on antlr-interest, I know I am.
//
switch (input->encoding)
{
case ANTLR3_ENC_UTF8:
// See if there is a BOM at the start of this UTF-8 sequence
// and just eat it if there is. Windows .TXT files have this for instance
// as it identifies UTF-8 even though it is of no consequence for byte order
// as UTF-8 does not have a byte order.
//
if ( (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar)) == 0xEF
&& (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar+1)) == 0xBB
&& (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar+2)) == 0xBF
)
{
// The UTF8 BOM is present so skip it
//
input->nextChar = (void *)((pANTLR3_UINT8)input->nextChar + 3);
}
// Install the UTF8 input routines
//
antlr3UTF8SetupStream(input);
break;
case ANTLR3_ENC_UTF16:
// See if there is a BOM at the start of the input. If not then
// we assume that the byte order is the natural order of this
// machine (or it is really UCS2). If there is a BOM we determine if the encoding
// is the same as the natural order of this machine.
//
if ( (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar)) == 0xFE
&& (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar+1)) == 0xFF
)
{
// BOM Present, indicates Big Endian
//
input->nextChar = (void *)((pANTLR3_UINT8)input->nextChar + 2);
antlr3UTF16SetupStream(input, isBigEndian, ANTLR3_TRUE);
}
else if ( (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar)) == 0xFF
&& (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar+1)) == 0xFE
)
{
// BOM present, indicates Little Endian
//
input->nextChar = (void *)((pANTLR3_UINT8)input->nextChar + 2);
antlr3UTF16SetupStream(input, isBigEndian, ANTLR3_FALSE);
}
else
{
// No BOM present, assume local computer byte order
//
antlr3UTF16SetupStream(input, isBigEndian, isBigEndian);
}
break;
case ANTLR3_ENC_UTF32:
// See if there is a BOM at the start of the input. If not then
// we assume that the byte order is the natural order of this
// machine. If there is we determine if the encoding
// is the same as the natural order of this machine.
//
if ( (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar)) == 0x00
&& (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar+1)) == 0x00
&& (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar+2)) == 0xFE
&& (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar+3)) == 0xFF
)
{
// BOM Present, indicates Big Endian
//
input->nextChar = (void *)((pANTLR3_UINT8)input->nextChar + 4);
antlr3UTF32SetupStream(input, isBigEndian, ANTLR3_TRUE);
}
else if ( (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar)) == 0xFF
&& (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar+1)) == 0xFE
&& (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar+1)) == 0x00
&& (ANTLR3_UINT8)(*((pANTLR3_UINT8)input->nextChar+1)) == 0x00
)
{
// BOM present, indicates Little Endian
//
input->nextChar = (void *)((pANTLR3_UINT8)input->nextChar + 4);
antlr3UTF32SetupStream(input, isBigEndian, ANTLR3_FALSE);
}
else
{
// No BOM present, assume local computer byte order
//
antlr3UTF32SetupStream(input, isBigEndian, isBigEndian);
}
break;
case ANTLR3_ENC_UTF16BE:
// Encoding is definately Big Endian with no BOM
//
antlr3UTF16SetupStream(input, isBigEndian, ANTLR3_TRUE);
break;
case ANTLR3_ENC_UTF16LE:
// Encoding is definately Little Endian with no BOM
//
antlr3UTF16SetupStream(input, isBigEndian, ANTLR3_FALSE);
break;
case ANTLR3_ENC_UTF32BE:
// Encoding is definately Big Endian with no BOM
//
antlr3UTF32SetupStream(input, isBigEndian, ANTLR3_TRUE);
break;
case ANTLR3_ENC_UTF32LE:
// Encoding is definately Little Endian with no BOM
//
antlr3UTF32SetupStream(input, isBigEndian, ANTLR3_FALSE);
break;
case ANTLR3_ENC_EBCDIC:
// EBCDIC is basically the same as ASCII but with an on the
// fly translation to ASCII
//
antlr3EBCDICSetupStream(input);
break;
case ANTLR3_ENC_8BIT:
default:
// Standard 8bit/ASCII
//
antlr38BitSetupStream(input);
break;
}
#endif /* 0 */
}
