summaryrefslogtreecommitdiffstats
path: root/generic/tclBinary.c
diff options
context:
space:
mode:
Diffstat (limited to 'generic/tclBinary.c')
-rw-r--r--generic/tclBinary.c569
1 files changed, 403 insertions, 166 deletions
diff --git a/generic/tclBinary.c b/generic/tclBinary.c
index 444e7fa..981f174 100644
--- a/generic/tclBinary.c
+++ b/generic/tclBinary.c
@@ -87,10 +87,13 @@ static int BinaryDecodeHex(ClientData clientData,
static int BinaryEncode64(ClientData clientData,
Tcl_Interp *interp,
int objc, Tcl_Obj *const objv[]);
-static int BinaryDecodeUu(ClientData clientData,
+static int BinaryDecode64(ClientData clientData,
Tcl_Interp *interp,
int objc, Tcl_Obj *const objv[]);
-static int BinaryDecode64(ClientData clientData,
+static int BinaryEncodeUu(ClientData clientData,
+ Tcl_Interp *interp, int objc,
+ Tcl_Obj *const objv[]);
+static int BinaryDecodeUu(ClientData clientData,
Tcl_Interp *interp,
int objc, Tcl_Obj *const objv[]);
@@ -128,6 +131,30 @@ static const char B64Digits[65] = {
};
/*
+ * How to construct the ensembles.
+ */
+
+static const EnsembleImplMap binaryMap[] = {
+ { "format", BinaryFormatCmd, TclCompileBasicMin1ArgCmd, NULL, NULL, 0 },
+ { "scan", BinaryScanCmd, TclCompileBasicMin2ArgCmd, NULL, NULL, 0 },
+ { "encode", NULL, NULL, NULL, NULL, 0 },
+ { "decode", NULL, NULL, NULL, NULL, 0 },
+ { NULL, NULL, NULL, NULL, NULL, 0 }
+};
+static const EnsembleImplMap encodeMap[] = {
+ { "hex", BinaryEncodeHex, TclCompileBasic1ArgCmd, NULL, NULL, 0 },
+ { "uuencode", BinaryEncodeUu, NULL, NULL, NULL, 0 },
+ { "base64", BinaryEncode64, NULL, NULL, NULL, 0 },
+ { NULL, NULL, NULL, NULL, NULL, 0 }
+};
+static const EnsembleImplMap decodeMap[] = {
+ { "hex", BinaryDecodeHex, TclCompileBasic1Or2ArgCmd, NULL, NULL, 0 },
+ { "uuencode", BinaryDecodeUu, TclCompileBasic1Or2ArgCmd, NULL, NULL, 0 },
+ { "base64", BinaryDecode64, TclCompileBasic1Or2ArgCmd, NULL, NULL, 0 },
+ { NULL, NULL, NULL, NULL, NULL, 0 }
+};
+
+/*
* The following object type represents an array of bytes. An array of bytes
* is not equivalent to an internationalized string. Conceptually, a string is
* an array of 16-bit quantities organized as a sequence of properly formed
@@ -180,9 +207,10 @@ typedef struct ByteArray {
#define BYTEARRAY_SIZE(len) \
((unsigned) (TclOffset(ByteArray, bytes) + (len)))
#define GET_BYTEARRAY(objPtr) \
- ((ByteArray *) (objPtr)->internalRep.otherValuePtr)
+ ((ByteArray *) (objPtr)->internalRep.twoPtrValue.ptr1)
#define SET_BYTEARRAY(objPtr, baPtr) \
- (objPtr)->internalRep.otherValuePtr = (void *) (baPtr)
+ (objPtr)->internalRep.twoPtrValue.ptr1 = (void *) (baPtr)
+
/*
*----------------------------------------------------------------------
@@ -301,20 +329,18 @@ Tcl_SetByteArrayObj(
Tcl_Panic("%s called with shared object", "Tcl_SetByteArrayObj");
}
TclFreeIntRep(objPtr);
- Tcl_InvalidateStringRep(objPtr);
+ TclInvalidateStringRep(objPtr);
- length = (length < 0) ? 0 : length;
+ if (length < 0) {
+ length = 0;
+ }
byteArrayPtr = ckalloc(BYTEARRAY_SIZE(length));
byteArrayPtr->used = length;
byteArrayPtr->allocated = length;
- if (length) {
- if (bytes) {
- memcpy(byteArrayPtr->bytes, bytes, (size_t) length);
- } else {
- memset(byteArrayPtr->bytes, 0, (size_t) length);
- }
- }
+ if ((bytes != NULL) && (length > 0)) {
+ memcpy(byteArrayPtr->bytes, bytes, (size_t) length);
+ }
objPtr->typePtr = &tclByteArrayType;
SET_BYTEARRAY(objPtr, byteArrayPtr);
}
@@ -398,7 +424,7 @@ Tcl_SetByteArrayLength(
byteArrayPtr->allocated = length;
SET_BYTEARRAY(objPtr, byteArrayPtr);
}
- Tcl_InvalidateStringRep(objPtr);
+ TclInvalidateStringRep(objPtr);
byteArrayPtr->used = length;
return byteArrayPtr->bytes;
}
@@ -584,9 +610,7 @@ UpdateStringOfByteArray(
*
* This function appends an array of bytes to a byte array object. Note
* that the object *must* be unshared, and the array of bytes *must not*
- * refer to the object being appended to. Also the caller must have
- * already checked that the final length of the bytearray after the
- * append operations is complete will not overflow the int range.
+ * refer to the object being appended to.
*
* Results:
* None.
@@ -605,6 +629,7 @@ TclAppendBytesToByteArray(
int len)
{
ByteArray *byteArrayPtr;
+ int needed;
if (Tcl_IsShared(objPtr)) {
Tcl_Panic("%s called with shared object","TclAppendBytesToByteArray");
@@ -613,64 +638,57 @@ TclAppendBytesToByteArray(
Tcl_Panic("%s must be called with definite number of bytes to append",
"TclAppendBytesToByteArray");
}
+ if (len == 0) {
+ /* Append zero bytes is a no-op. */
+ return;
+ }
if (objPtr->typePtr != &tclByteArrayType) {
SetByteArrayFromAny(NULL, objPtr);
}
byteArrayPtr = GET_BYTEARRAY(objPtr);
+ if (len > INT_MAX - byteArrayPtr->used) {
+ Tcl_Panic("max size for a Tcl value (%d bytes) exceeded", INT_MAX);
+ }
+
+ needed = byteArrayPtr->used + len;
/*
* If we need to, resize the allocated space in the byte array.
*/
- if (byteArrayPtr->used + len > byteArrayPtr->allocated) {
- unsigned int attempt, used = byteArrayPtr->used;
- ByteArray *tmpByteArrayPtr = NULL;
+ if (needed > byteArrayPtr->allocated) {
+ ByteArray *ptr = NULL;
+ int attempt;
- attempt = byteArrayPtr->allocated;
- if (attempt < 1) {
- /*
- * No allocated bytes, so must be none used too. We use this
- * method to calculate how many bytes to allocate because we can
- * end up with a zero-length buffer otherwise, when doubling can
- * cause trouble. [Bug 3067036]
- */
-
- attempt = len + 1;
- } else {
- do {
- attempt *= 2;
- } while (attempt < used+len);
+ if (needed <= INT_MAX/2) {
+ /* Try to allocate double the total space that is needed. */
+ attempt = 2 * needed;
+ ptr = attemptckrealloc(byteArrayPtr, BYTEARRAY_SIZE(attempt));
}
-
- if (BYTEARRAY_SIZE(attempt) > BYTEARRAY_SIZE(used)) {
- tmpByteArrayPtr = attemptckrealloc(byteArrayPtr,
- BYTEARRAY_SIZE(attempt));
+ if (ptr == NULL) {
+ /* Try to allocate double the increment that is needed (plus). */
+ unsigned int limit = INT_MAX - needed;
+ unsigned int extra = len + TCL_MIN_GROWTH;
+ int growth = (int) ((extra > limit) ? limit : extra);
+
+ attempt = needed + growth;
+ ptr = attemptckrealloc(byteArrayPtr, BYTEARRAY_SIZE(attempt));
}
-
- if (tmpByteArrayPtr == NULL) {
- attempt = used + len;
- if (BYTEARRAY_SIZE(attempt) < BYTEARRAY_SIZE(used)) {
- Tcl_Panic("attempt to allocate a bigger buffer than we can handle");
- }
- tmpByteArrayPtr = ckrealloc(byteArrayPtr,
- BYTEARRAY_SIZE(attempt));
+ if (ptr == NULL) {
+ /* Last chance: Try to allocate exactly what is needed. */
+ attempt = needed;
+ ptr = ckrealloc(byteArrayPtr, BYTEARRAY_SIZE(attempt));
}
-
- byteArrayPtr = tmpByteArrayPtr;
+ byteArrayPtr = ptr;
byteArrayPtr->allocated = attempt;
- byteArrayPtr->used = used;
SET_BYTEARRAY(objPtr, byteArrayPtr);
}
- /*
- * Do the append if there's any point.
- */
-
- if (len > 0) {
+ if (bytes) {
memcpy(byteArrayPtr->bytes + byteArrayPtr->used, bytes, len);
- byteArrayPtr->used += len;
- Tcl_InvalidateStringRep(objPtr);
}
+ byteArrayPtr->used += len;
+ TclInvalidateStringRep(objPtr);
}
/*
@@ -690,26 +708,6 @@ TclAppendBytesToByteArray(
*----------------------------------------------------------------------
*/
-static const EnsembleImplMap binaryMap[] = {
-{ "format", BinaryFormatCmd, NULL, NULL, NULL, 0 },
-{ "scan", BinaryScanCmd, NULL, NULL, NULL, 0 },
-{ "encode", NULL, NULL, NULL, NULL, 0 },
-{ "decode", NULL, NULL, NULL, NULL, 0 },
-{ NULL, NULL, NULL, NULL, NULL, 0 }
-};
-static const EnsembleImplMap encodeMap[] = {
-{ "hex", BinaryEncodeHex, NULL, NULL, (ClientData)HexDigits, 0 },
-{ "uuencode", BinaryEncode64, NULL, NULL, (ClientData)UueDigits, 0 },
-{ "base64", BinaryEncode64, NULL, NULL, (ClientData)B64Digits, 0 },
-{ NULL, NULL, NULL, NULL, NULL, 0 }
-};
-static const EnsembleImplMap decodeMap[] = {
-{ "hex", BinaryDecodeHex, NULL, NULL, NULL, 0 },
-{ "uuencode", BinaryDecodeUu, NULL, NULL, NULL, 0 },
-{ "base64", BinaryDecode64, NULL, NULL, NULL, 0 },
-{ NULL, NULL, NULL, NULL, NULL, 0 }
-};
-
Tcl_Command
TclInitBinaryCmd(
Tcl_Interp *interp)
@@ -873,9 +871,9 @@ BinaryFormatCmd(
if (count == BINARY_ALL) {
count = listc;
} else if (count > listc) {
- Tcl_AppendResult(interp,
+ Tcl_SetObjResult(interp, Tcl_NewStringObj(
"number of elements in list does not match count",
- NULL);
+ -1));
return TCL_ERROR;
}
}
@@ -884,9 +882,8 @@ BinaryFormatCmd(
case 'x':
if (count == BINARY_ALL) {
- Tcl_AppendResult(interp,
- "cannot use \"*\" in format string with \"x\"",
- NULL);
+ Tcl_SetObjResult(interp, Tcl_NewStringObj(
+ "cannot use \"*\" in format string with \"x\"", -1));
return TCL_ERROR;
} else if (count == BINARY_NOCOUNT) {
count = 1;
@@ -1198,8 +1195,9 @@ BinaryFormatCmd(
badValue:
Tcl_ResetResult(interp);
- Tcl_AppendResult(interp, "expected ", errorString,
- " string but got \"", errorValue, "\" instead", NULL);
+ Tcl_SetObjResult(interp, Tcl_ObjPrintf(
+ "expected %s string but got \"%s\" instead",
+ errorString, errorValue));
return TCL_ERROR;
badCount:
@@ -1217,12 +1215,13 @@ BinaryFormatCmd(
Tcl_UtfToUniChar(errorString, &ch);
buf[Tcl_UniCharToUtf(ch, buf)] = '\0';
- Tcl_AppendResult(interp, "bad field specifier \"", buf, "\"", NULL);
+ Tcl_SetObjResult(interp, Tcl_ObjPrintf(
+ "bad field specifier \"%s\"", buf));
return TCL_ERROR;
}
error:
- Tcl_AppendResult(interp, errorString, NULL);
+ Tcl_SetObjResult(interp, Tcl_NewStringObj(errorString, -1));
return TCL_ERROR;
}
@@ -1586,12 +1585,13 @@ BinaryScanCmd(
Tcl_UtfToUniChar(errorString, &ch);
buf[Tcl_UniCharToUtf(ch, buf)] = '\0';
- Tcl_AppendResult(interp, "bad field specifier \"", buf, "\"", NULL);
+ Tcl_SetObjResult(interp, Tcl_ObjPrintf(
+ "bad field specifier \"%s\"", buf));
return TCL_ERROR;
}
error:
- Tcl_AppendResult(interp, errorString, NULL);
+ Tcl_SetObjResult(interp, Tcl_NewStringObj(errorString, -1));
return TCL_ERROR;
}
@@ -2307,7 +2307,6 @@ BinaryEncodeHex(
Tcl_Obj *resultObj = NULL;
unsigned char *data = NULL;
unsigned char *cursor = NULL;
- const char *digits = clientData;
int offset = 0, count = 0;
if (objc != 2) {
@@ -2319,8 +2318,8 @@ BinaryEncodeHex(
data = Tcl_GetByteArrayFromObj(objv[1], &count);
cursor = Tcl_SetByteArrayLength(resultObj, count * 2);
for (offset = 0; offset < count; ++offset) {
- *cursor++ = digits[((data[offset] >> 4) & 0x0f)];
- *cursor++ = digits[(data[offset] & 0x0f)];
+ *cursor++ = HexDigits[((data[offset] >> 4) & 0x0f)];
+ *cursor++ = HexDigits[(data[offset] & 0x0f)];
}
Tcl_SetObjResult(interp, resultObj);
return TCL_OK;
@@ -2357,7 +2356,7 @@ BinaryDecodeHex(
static const char *const optStrings[] = { "-strict", NULL };
if (objc < 2 || objc > 3) {
- Tcl_WrongNumArgs(interp, 1, objv, "data");
+ Tcl_WrongNumArgs(interp, 1, objv, "?options? data");
return TCL_ERROR;
}
for (i = 1; i < objc-1; ++i) {
@@ -2381,29 +2380,32 @@ BinaryDecodeHex(
while (data < dataend) {
value = 0;
for (i=0 ; i<2 ; i++) {
- if (data < dataend) {
- c = *data++;
-
- if (!isxdigit((int) c)) {
- if (strict || !isspace(c)) {
- goto badChar;
- }
- i--;
- continue;
- }
+ if (data >= dataend) {
value <<= 4;
- c -= '0';
- if (c > 9) {
- c += ('0' - 'A') + 10;
- }
- if (c > 16) {
- c += ('A' - 'a');
+ break;
+ }
+
+ c = *data++;
+ if (!isxdigit((int) c)) {
+ if (strict || !isspace(c)) {
+ goto badChar;
}
- value |= (c & 0xf);
- } else {
- value <<= 4;
- cut++;
+ i--;
+ continue;
+ }
+
+ value <<= 4;
+ c -= '0';
+ if (c > 9) {
+ c += ('0' - 'A') + 10;
}
+ if (c > 16) {
+ c += ('A' - 'a');
+ }
+ value |= (c & 0xf);
+ }
+ if (i < 2) {
+ cut++;
}
*cursor++ = UCHAR(value);
value = 0;
@@ -2431,7 +2433,7 @@ BinaryDecodeHex(
* This implements a generic 6 bit binary encoding. Input is broken into
* 6 bit chunks and a lookup table passed in via clientData is used to
* turn these values into output characters. This is used to implement
- * base64 and uuencode binary encodings.
+ * base64 binary encodings.
*
* Results:
* Interp result set to an encoded byte array object
@@ -2467,7 +2469,6 @@ BinaryEncode64(
{
Tcl_Obj *resultObj;
unsigned char *data, *cursor, *limit;
- const char *digits = clientData;
int maxlen = 0;
const char *wrapchar = "\n";
int wrapcharlen = 1;
@@ -2490,6 +2491,13 @@ BinaryEncode64(
if (Tcl_GetIntFromObj(interp, objv[i+1], &maxlen) != TCL_OK) {
return TCL_ERROR;
}
+ if (maxlen < 0) {
+ Tcl_SetObjResult(interp, Tcl_NewStringObj(
+ "line length out of range", -1));
+ Tcl_SetErrorCode(interp, "TCL", "BINARY", "ENCODE",
+ "LINE_LENGTH", NULL);
+ return TCL_ERROR;
+ }
break;
case OPT_WRAPCHAR:
wrapchar = Tcl_GetStringFromObj(objv[i+1], &wrapcharlen);
@@ -2520,17 +2528,17 @@ BinaryEncode64(
for (i = 0; i < 3 && offset+i < count; ++i) {
d[i] = data[offset + i];
}
- OUTPUT(digits[d[0] >> 2]);
- OUTPUT(digits[((d[0] & 0x03) << 4) | (d[1] >> 4)]);
+ OUTPUT(B64Digits[d[0] >> 2]);
+ OUTPUT(B64Digits[((d[0] & 0x03) << 4) | (d[1] >> 4)]);
if (offset+1 < count) {
- OUTPUT(digits[((d[1] & 0x0f) << 2) | (d[2] >> 6)]);
+ OUTPUT(B64Digits[((d[1] & 0x0f) << 2) | (d[2] >> 6)]);
} else {
- OUTPUT(digits[64]);
+ OUTPUT(B64Digits[64]);
}
if (offset+2 < count) {
- OUTPUT(digits[d[2] & 0x3f]);
+ OUTPUT(B64Digits[d[2] & 0x3f]);
} else {
- OUTPUT(digits[64]);
+ OUTPUT(B64Digits[64]);
}
}
}
@@ -2542,6 +2550,125 @@ BinaryEncode64(
/*
*----------------------------------------------------------------------
*
+ * BinaryEncodeUu --
+ *
+ * This implements the uuencode binary encoding. Input is broken into 6
+ * bit chunks and a lookup table is used to turn these values into output
+ * characters. This differs from the generic code above in that line
+ * lengths are also encoded.
+ *
+ * Results:
+ * Interp result set to an encoded byte array object
+ *
+ * Side effects:
+ * None
+ *
+ *----------------------------------------------------------------------
+ */
+
+static int
+BinaryEncodeUu(
+ ClientData clientData,
+ Tcl_Interp *interp,
+ int objc,
+ Tcl_Obj *const objv[])
+{
+ Tcl_Obj *resultObj;
+ unsigned char *data, *start, *cursor;
+ int offset, count, rawLength, n, i, j, bits, index;
+ int lineLength = 61;
+ const unsigned char SingleNewline[] = { (unsigned char) '\n' };
+ const unsigned char *wrapchar = SingleNewline;
+ int wrapcharlen = sizeof(SingleNewline);
+ enum { OPT_MAXLEN, OPT_WRAPCHAR };
+ static const char *const optStrings[] = { "-maxlen", "-wrapchar", NULL };
+
+ if (objc < 2 || objc%2 != 0) {
+ Tcl_WrongNumArgs(interp, 1, objv,
+ "?-maxlen len? ?-wrapchar char? data");
+ return TCL_ERROR;
+ }
+ for (i = 1; i < objc-1; i += 2) {
+ if (Tcl_GetIndexFromObj(interp, objv[i], optStrings, "option",
+ TCL_EXACT, &index) != TCL_OK) {
+ return TCL_ERROR;
+ }
+ switch (index) {
+ case OPT_MAXLEN:
+ if (Tcl_GetIntFromObj(interp, objv[i+1], &lineLength) != TCL_OK) {
+ return TCL_ERROR;
+ }
+ if (lineLength < 3 || lineLength > 85) {
+ Tcl_SetObjResult(interp, Tcl_NewStringObj(
+ "line length out of range", -1));
+ Tcl_SetErrorCode(interp, "TCL", "BINARY", "ENCODE",
+ "LINE_LENGTH", NULL);
+ return TCL_ERROR;
+ }
+ break;
+ case OPT_WRAPCHAR:
+ wrapchar = Tcl_GetByteArrayFromObj(objv[i+1], &wrapcharlen);
+ break;
+ }
+ }
+
+ /*
+ * Allocate the buffer. This is a little bit too long, but is "good
+ * enough".
+ */
+
+ resultObj = Tcl_NewObj();
+ offset = 0;
+ data = Tcl_GetByteArrayFromObj(objv[objc-1], &count);
+ rawLength = (lineLength - 1) * 3 / 4;
+ start = cursor = Tcl_SetByteArrayLength(resultObj,
+ (lineLength + wrapcharlen) *
+ ((count + (rawLength - 1)) / rawLength));
+ n = bits = 0;
+
+ /*
+ * Encode the data. Each output line first has the length of raw data
+ * encoded by the output line described in it by one encoded byte, then
+ * the encoded data follows (encoding each 6 bits as one character).
+ * Encoded lines are always terminated by a newline.
+ */
+
+ while (offset < count) {
+ int lineLen = count - offset;
+
+ if (lineLen > rawLength) {
+ lineLen = rawLength;
+ }
+ *cursor++ = UueDigits[lineLen];
+ for (i=0 ; i<lineLen ; i++) {
+ n <<= 8;
+ n |= data[offset++];
+ for (bits += 8; bits > 6 ; bits -= 6) {
+ *cursor++ = UueDigits[(n >> (bits-6)) & 0x3f];
+ }
+ }
+ if (bits > 0) {
+ n <<= 8;
+ *cursor++ = UueDigits[(n >> (bits + 2)) & 0x3f];
+ bits = 0;
+ }
+ for (j=0 ; j<wrapcharlen ; ++j) {
+ *cursor++ = wrapchar[j];
+ }
+ }
+
+ /*
+ * Fix the length of the output bytearray.
+ */
+
+ Tcl_SetByteArrayLength(resultObj, cursor-start);
+ Tcl_SetObjResult(interp, resultObj);
+ return TCL_OK;
+}
+
+/*
+ *----------------------------------------------------------------------
+ *
* BinaryDecodeUu --
*
* Decode a uuencoded string.
@@ -2565,13 +2692,13 @@ BinaryDecodeUu(
Tcl_Obj *resultObj = NULL;
unsigned char *data, *datastart, *dataend;
unsigned char *begin, *cursor;
- int i, index, size, count = 0, cut = 0, strict = 0;
- char c;
+ int i, index, size, count = 0, strict = 0, lineLen;
+ unsigned char c;
enum {OPT_STRICT };
static const char *const optStrings[] = { "-strict", NULL };
if (objc < 2 || objc > 3) {
- Tcl_WrongNumArgs(interp, 1, objv, "data");
+ Tcl_WrongNumArgs(interp, 1, objv, "?options? data");
return TCL_ERROR;
}
for (i = 1; i < objc-1; ++i) {
@@ -2592,44 +2719,112 @@ BinaryDecodeUu(
dataend = data + count;
size = ((count + 3) & ~3) * 3 / 4;
begin = cursor = Tcl_SetByteArrayLength(resultObj, size);
+ lineLen = -1;
+
+ /*
+ * The decoding loop. First, we get the length of line (strictly, the
+ * number of data bytes we expect to generate from the line) we're
+ * processing this time round if it is not already known (i.e., when the
+ * lineLen variable is set to the magic value, -1).
+ */
+
while (data < dataend) {
char d[4] = {0, 0, 0, 0};
+ if (lineLen < 0) {
+ c = *data++;
+ if (c < 32 || c > 96) {
+ if (strict || !isspace(c)) {
+ goto badUu;
+ }
+ i--;
+ continue;
+ }
+ lineLen = (c - 32) & 0x3f;
+ }
+
+ /*
+ * Now we read a four-character grouping.
+ */
+
for (i=0 ; i<4 ; i++) {
if (data < dataend) {
d[i] = c = *data++;
- if (c < 33 || c > 96) {
- if (strict || !isspace(UCHAR(c))) {
- goto badUu;
+ if (c < 32 || c > 96) {
+ if (strict) {
+ if (!isspace(c)) {
+ goto badUu;
+ } else if (c == '\n') {
+ goto shortUu;
+ }
}
i--;
continue;
}
- } else {
- cut++;
}
}
- if (cut > 3) {
- cut = 3;
+
+ /*
+ * Translate that grouping into (up to) three binary bytes output.
+ */
+
+ if (lineLen > 0) {
+ *cursor++ = (((d[0] - 0x20) & 0x3f) << 2)
+ | (((d[1] - 0x20) & 0x3f) >> 4);
+ if (--lineLen > 0) {
+ *cursor++ = (((d[1] - 0x20) & 0x3f) << 4)
+ | (((d[2] - 0x20) & 0x3f) >> 2);
+ if (--lineLen > 0) {
+ *cursor++ = (((d[2] - 0x20) & 0x3f) << 6)
+ | (((d[3] - 0x20) & 0x3f));
+ lineLen--;
+ }
+ }
+ }
+
+ /*
+ * If we've reached the end of the line, skip until we process a
+ * newline.
+ */
+
+ if (lineLen == 0 && data < dataend) {
+ lineLen = -1;
+ do {
+ c = *data++;
+ if (c == '\n') {
+ break;
+ } else if (c >= 32 && c <= 96) {
+ data--;
+ break;
+ } else if (strict || !isspace(c)) {
+ goto badUu;
+ }
+ } while (data < dataend);
}
- *cursor++ = (((d[0] - 0x20) & 0x3f) << 2)
- | (((d[1] - 0x20) & 0x3f) >> 4);
- *cursor++ = (((d[1] - 0x20) & 0x3f) << 4)
- | (((d[2] - 0x20) & 0x3f) >> 2);
- *cursor++ = (((d[2] - 0x20) & 0x3f) << 6)
- | (((d[3] - 0x20) & 0x3f));
}
- if (cut > size) {
- cut = size;
+
+ /*
+ * Sanity check, clean up and finish.
+ */
+
+ if (lineLen > 0 && strict) {
+ goto shortUu;
}
- Tcl_SetByteArrayLength(resultObj, cursor - begin - cut);
+ Tcl_SetByteArrayLength(resultObj, cursor - begin);
Tcl_SetObjResult(interp, resultObj);
return TCL_OK;
+ shortUu:
+ Tcl_SetObjResult(interp, Tcl_ObjPrintf("short uuencode data"));
+ Tcl_SetErrorCode(interp, "TCL", "BINARY", "DECODE", "SHORT", NULL);
+ TclDecrRefCount(resultObj);
+ return TCL_ERROR;
+
badUu:
Tcl_SetObjResult(interp, Tcl_ObjPrintf(
"invalid uuencode character \"%c\" at position %d",
c, (int) (data - datastart - 1)));
+ Tcl_SetErrorCode(interp, "TCL", "BINARY", "DECODE", "INVALID", NULL);
TclDecrRefCount(resultObj);
return TCL_ERROR;
}
@@ -2658,16 +2853,16 @@ BinaryDecode64(
Tcl_Obj *const objv[])
{
Tcl_Obj *resultObj = NULL;
- unsigned char *data, *datastart, *dataend, c;
+ unsigned char *data, *datastart, *dataend, c = '\0';
unsigned char *begin = NULL;
unsigned char *cursor = NULL;
int strict = 0;
int i, index, size, cut = 0, count = 0;
- enum {OPT_STRICT };
+ enum { OPT_STRICT };
static const char *const optStrings[] = { "-strict", NULL };
if (objc < 2 || objc > 3) {
- Tcl_WrongNumArgs(interp, 1, objv, "data");
+ Tcl_WrongNumArgs(interp, 1, objv, "?options? data");
return TCL_ERROR;
}
for (i = 1; i < objc-1; ++i) {
@@ -2691,43 +2886,85 @@ BinaryDecode64(
while (data < dataend) {
unsigned long value = 0;
- for (i=0 ; i<4 ; i++) {
+ /*
+ * Decode the current block. Each base64 block consists of four input
+ * characters A-Z, a-z, 0-9, +, or /. Each character supplies six bits
+ * of output data, so each block's output is 24 bits (three bytes) in
+ * length. The final block can be shorter by one or two bytes, denoted
+ * by the input ending with one or two ='s, respectively.
+ */
+
+ for (i = 0; i < 4; i++) {
+ /*
+ * Get the next input character. At end of input, pad with at most
+ * two ='s. If more than two ='s would be needed, instead discard
+ * the block read thus far.
+ */
+
if (data < dataend) {
c = *data++;
+ } else if (i > 1) {
+ c = '=';
+ } else {
+ cut += 3;
+ break;
+ }
- if (c >= 'A' && c <= 'Z') {
- value = (value << 6) | ((c - 'A') & 0x3f);
- } else if (c >= 'a' && c <= 'z') {
- value = (value << 6) | ((c - 'a' + 26) & 0x3f);
- } else if (c >= '0' && c <= '9') {
- value = (value << 6) | ((c - '0' + 52) & 0x3f);
- } else if (c == '+') {
- value = (value << 6) | 0x3e;
- } else if (c == '/') {
- value = (value << 6) | 0x3f;
- } else if (c == '=') {
- value <<= 6;
- if (cut < 2) {
- cut++;
- }
+ /*
+ * Load the character into the block value. Handle ='s specially
+ * because they're only valid as the last character or two of the
+ * final block of input. Unless strict mode is enabled, skip any
+ * input whitespace characters.
+ */
+
+ if (cut) {
+ if (c == '=' && i > 1) {
+ value <<= 6;
+ cut++;
+ } else if (!strict && isspace(c)) {
+ i--;
} else {
- if (strict || !isspace(c)) {
- goto bad64;
- }
- i--;
- continue;
+ goto bad64;
}
- } else {
+ } else if (c >= 'A' && c <= 'Z') {
+ value = (value << 6) | ((c - 'A') & 0x3f);
+ } else if (c >= 'a' && c <= 'z') {
+ value = (value << 6) | ((c - 'a' + 26) & 0x3f);
+ } else if (c >= '0' && c <= '9') {
+ value = (value << 6) | ((c - '0' + 52) & 0x3f);
+ } else if (c == '+') {
+ value = (value << 6) | 0x3e;
+ } else if (c == '/') {
+ value = (value << 6) | 0x3f;
+ } else if (c == '=') {
value <<= 6;
cut++;
+ } else if (strict || !isspace(c)) {
+ goto bad64;
+ } else {
+ i--;
}
}
*cursor++ = UCHAR((value >> 16) & 0xff);
*cursor++ = UCHAR((value >> 8) & 0xff);
*cursor++ = UCHAR(value & 0xff);
- }
- if (cut > size) {
- cut = size;
+
+ /*
+ * Since = is only valid within the final block, if it was encountered
+ * but there are still more input characters, confirm that strict mode
+ * is off and all subsequent characters are whitespace.
+ */
+
+ if (cut && data < dataend) {
+ if (strict) {
+ goto bad64;
+ }
+ for (; data < dataend; data++) {
+ if (!isspace(*data)) {
+ goto bad64;
+ }
+ }
+ }
}
Tcl_SetByteArrayLength(resultObj, cursor - begin - cut);
Tcl_SetObjResult(interp, resultObj);
generated by cgit v0.12 at 2025-10-31 20:32:25 (GMT)