Elaborate TIP #481 implementation: Make clear that Tcl_GetUnicodeFromObj and Tcl_GetStringFromObj panic if lengthPtr points to an int and length > INT_MAX.

Also if sizeof(int) == sizeof(size_t), prefer the size_t variant of the functions

4 files changed, 22 insertions, 11 deletions

diff --git a/doc/StringObj.3 b/doc/StringObj.3
index c9bdd4a..2526a01 100644
--- a/doc/StringObj.3
+++ b/doc/StringObj.3
@@ -188,7 +188,9 @@ Even in the limited situations where writing to this pointer is
 acceptable, one should take care to respect the copy-on-write
 semantics required by \fBTcl_Obj\fR's, with appropriate calls
 to \fBTcl_IsShared\fR and \fBTcl_DuplicateObj\fR prior to any
-in-place modification of the string representation.
+in-place modification of the string representation. If \fIlengthPtr\fR
+points to an \fBint\fR variable, and the string has more than 2^31 bytes,
+a panic will result.
 The procedure \fBTcl_GetString\fR is used in the common case
 where the caller does not need the length of the string
 representation.
@@ -200,7 +202,8 @@ value as a Unicode string.  This is given by the returned pointer and
 byte pointer is owned by the value manager and should not be modified by
 the caller.  The procedure \fBTcl_GetUnicode\fR is used in the common case
 where the caller does not need the length of the unicode string
-representation.
+representation. If \fIlengthPtr\fR points to an \fBint\fR variable,
+and the string has more than 2^31 unicode characters, a panic will result. 
 .PP
 \fBTcl_GetUniChar\fR returns the \fIindex\fR'th character in the
 value's Unicode representation. If the index is out of range or
diff --git a/generic/tclDecls.h b/generic/tclDecls.h
index 0995678..02c365e 100644
--- a/generic/tclDecls.h
+++ b/generic/tclDecls.h
@@ -3916,24 +3916,24 @@ extern const TclStubs *tclStubsPtr;
 #undef Tcl_GetBytesFromObj
 #if defined(USE_TCL_STUBS)
 #define Tcl_GetStringFromObj(objPtr, sizePtr) \
-	(sizeof(*(sizePtr)) <= sizeof(int) ? tclStubsPtr->tclGetStringFromObj(objPtr, (int *)(sizePtr)) : tclStubsPtr->tcl_GetStringFromObj(objPtr, (size_t *)(sizePtr)))
+	((sizeof(*(sizePtr)) <= sizeof(int) && sizeof(int) != sizeof(size_t)) ? tclStubsPtr->tclGetStringFromObj(objPtr, (int *)(sizePtr)) : tclStubsPtr->tcl_GetStringFromObj(objPtr, (size_t *)(sizePtr)))
 #define Tcl_GetBytesFromObj(interp, objPtr, sizePtr) \
-	(sizeof(*(sizePtr)) <= sizeof(int) ? tclStubsPtr->tclGetBytesFromObj(interp, objPtr, (int *)(sizePtr)) : tclStubsPtr->tcl_GetBytesFromObj(interp, objPtr, (size_t *)(sizePtr)))
+	((sizeof(*(sizePtr)) <= sizeof(int)) ? tclStubsPtr->tclGetBytesFromObj(interp, objPtr, (int *)(sizePtr)) : tclStubsPtr->tcl_GetBytesFromObj(interp, objPtr, (size_t *)(sizePtr)))
 #define Tcl_GetByteArrayFromObj(objPtr, sizePtr) \
-	(sizeof(*(sizePtr)) <= sizeof(int) ? tclStubsPtr->tclGetBytesFromObj(NULL, objPtr, (int *)(sizePtr)) : tclStubsPtr->tcl_GetBytesFromObj(NULL, objPtr, (size_t *)(sizePtr)))
+	((sizeof(*(sizePtr)) <= sizeof(int)) ? tclStubsPtr->tclGetBytesFromObj(NULL, objPtr, (int *)(sizePtr)) : tclStubsPtr->tcl_GetBytesFromObj(NULL, objPtr, (size_t *)(sizePtr)))
 #define Tcl_GetUnicodeFromObj(objPtr, sizePtr) \
-	(sizeof(*(sizePtr)) <= sizeof(int) ? tclStubsPtr->tclGetUnicodeFromObj(objPtr, (int *)(sizePtr)) : tclStubsPtr->tcl_GetUnicodeFromObj(objPtr, (size_t *)(sizePtr)))
+	((sizeof(*(sizePtr)) <= sizeof(int)) ? tclStubsPtr->tclGetUnicodeFromObj(objPtr, (int *)(sizePtr)) : tclStubsPtr->tcl_GetUnicodeFromObj(objPtr, (size_t *)(sizePtr)))
 #define Tcl_GetIndexFromObjStruct(interp, objPtr, tablePtr, offset, msg, flags, indexPtr) \
 	(tclStubsPtr->tcl_GetIndexFromObjStruct((interp), (objPtr), (tablePtr), (offset), (msg), (flags)|(int)(sizeof(*(indexPtr))<<1), (indexPtr)))
 #else
 #define Tcl_GetStringFromObj(objPtr, sizePtr) \
-	(sizeof(*(sizePtr)) <= sizeof(int) ? (TclGetStringFromObj)(objPtr, (int *)(sizePtr)) : (Tcl_GetStringFromObj)(objPtr, (size_t *)(sizePtr)))
+	((sizeof(*(sizePtr)) <= sizeof(int) && sizeof(int) != sizeof(size_t)) ? (TclGetStringFromObj)(objPtr, (int *)(sizePtr)) : (Tcl_GetStringFromObj)(objPtr, (size_t *)(sizePtr)))
 #define Tcl_GetBytesFromObj(interp, objPtr, sizePtr) \
-	(sizeof(*(sizePtr)) <= sizeof(int) ? (TclGetBytesFromObj)(interp, objPtr, (int *)(sizePtr)) : (Tcl_GetBytesFromObj)(interp, objPtr, (size_t *)(sizePtr)))
+	((sizeof(*(sizePtr)) <= sizeof(int) && sizeof(int) != sizeof(size_t)) ? (TclGetBytesFromObj)(interp, objPtr, (int *)(sizePtr)) : (Tcl_GetBytesFromObj)(interp, objPtr, (size_t *)(sizePtr)))
 #define Tcl_GetByteArrayFromObj(objPtr, sizePtr) \
-	(sizeof(*(sizePtr)) <= sizeof(int) ? (TclGetBytesFromObj)(NULL, objPtr, (int *)(sizePtr)) : (Tcl_GetBytesFromObj)(NULL, objPtr, (size_t *)(sizePtr)))
+	((sizeof(*(sizePtr)) <= sizeof(int) && sizeof(int) != sizeof(size_t)) ? (TclGetBytesFromObj)(NULL, objPtr, (int *)(sizePtr)) : (Tcl_GetBytesFromObj)(NULL, objPtr, (size_t *)(sizePtr)))
 #define Tcl_GetUnicodeFromObj(objPtr, sizePtr) \
-	(sizeof(*(sizePtr)) <= sizeof(int) ? (TclGetUnicodeFromObj)(objPtr, (int *)(sizePtr)) : Tcl_GetUnicodeFromObj(objPtr, (size_t *)(sizePtr)))
+	((sizeof(*(sizePtr)) <= sizeof(int) && sizeof(int) != sizeof(size_t)) ? (TclGetUnicodeFromObj)(objPtr, (int *)(sizePtr)) : Tcl_GetUnicodeFromObj(objPtr, (size_t *)(sizePtr)))
 #define Tcl_GetIndexFromObjStruct(interp, objPtr, tablePtr, offset, msg, flags, indexPtr) \
 	((Tcl_GetIndexFromObjStruct)((interp), (objPtr), (tablePtr), (offset), (msg), (flags)|(int)(sizeof(*(indexPtr))<<1), (indexPtr)))
 #endif
diff --git a/generic/tclObj.c b/generic/tclObj.c
index b4d6332..300e408 100644
--- a/generic/tclObj.c
+++ b/generic/tclObj.c
@@ -1675,7 +1675,11 @@ TclGetStringFromObj(
 	}
     }
     if (lengthPtr != NULL) {
-	*lengthPtr = (objPtr->length < INT_MAX)? objPtr->length: INT_MAX;
+	if (objPtr->length > INT_MAX) {
+	    Tcl_Panic("Tcl_GetStringFromObj with 'int' lengthPtr"
+		    "cannot handle such long strings. Please use 'size_t'");
+	}
+	*lengthPtr = (int)objPtr->length;
     }
     return objPtr->bytes;
 }
diff --git a/generic/tclStringObj.c b/generic/tclStringObj.c
index a3ee1c3..d0bac17 100644
--- a/generic/tclStringObj.c
+++ b/generic/tclStringObj.c
@@ -686,6 +686,10 @@ TclGetUnicodeFromObj(
     }
 
     if (lengthPtr != NULL) {
+	if (stringPtr->numChars > INT_MAX) {
+	    Tcl_Panic("Tcl_GetUnicodeFromObj with 'int' lengthPtr"
+		    "cannot handle such long strings. Please use 'size_t'");
+	}
 	*lengthPtr = (int)stringPtr->numChars;
     }
     return stringPtr->unicode;