diff options
| -rw-r--r-- | generic/tcl.decls | 2 | ||||
| -rw-r--r-- | generic/tclUtf.c | 6 |
2 files changed, 7 insertions, 1 deletions
diff --git a/generic/tcl.decls b/generic/tcl.decls index 61a71bd..6a34a53 100644 --- a/generic/tcl.decls +++ b/generic/tcl.decls @@ -2544,7 +2544,7 @@ export { export { void Tcl_InitSubsystems(void) } -export { +export { int TclZipfs_AppHook(int *argc, char ***argv) } diff --git a/generic/tclUtf.c b/generic/tclUtf.c index fde30c5..f56abd8 100644 --- a/generic/tclUtf.c +++ b/generic/tclUtf.c @@ -92,6 +92,8 @@ static const unsigned char complete[256] = { #if TCL_UTF_MAX > 3 4,4,4,4,4, #else + /* Tcl_UtfToUniChar() accesses src[1] and src[2] to check whether + * the UTF-8 sequence is valid, so we cannot use 1 here. */ 3,3,3,3,3, #endif 1,1,1,1,1,1,1,1,1,1,1 @@ -971,6 +973,10 @@ Tcl_UtfNext( const char *next; if (((*src) & 0xC0) == 0x80) { + /* Continuation byte, so we start 'inside' a (possible valid) UTF-8 + * sequence. Since we are not allowed to access src[-1], we cannot + * check if the sequence is actually valid, the best we can do is + * just assume it is valid and locate the end. */ if ((((*++src) & 0xC0) == 0x80) && (((*++src) & 0xC0) == 0x80)) { ++src; } |