diff options
| author | jan.nijtmans <nijtmans@users.sourceforge.net> | 2022-12-15 23:59:48 (GMT) |
|---|---|---|
| committer | jan.nijtmans <nijtmans@users.sourceforge.net> | 2022-12-15 23:59:48 (GMT) |
| commit | 9faf91402bd760daffdd9a305bd63065e9c9dc11 (patch) | |
| tree | a2eb3f7e2e38c24e28ca0236c7c9723c56f28f75 | |
| parent | 7b36aa11b6880b342a177a864601d97d839b7279 (diff) | |
| download | tcl-9faf91402bd760daffdd9a305bd63065e9c9dc11.zip tcl-9faf91402bd760daffdd9a305bd63065e9c9dc11.tar.gz tcl-9faf91402bd760daffdd9a305bd63065e9c9dc11.tar.bz2 | |
Make -strict work the same for UTF-16 as for UTF-8
| -rw-r--r-- | generic/tclEncoding.c | 13 |
1 files changed, 9 insertions, 4 deletions
diff --git a/generic/tclEncoding.c b/generic/tclEncoding.c index 169e975..78b0b9d 100644 --- a/generic/tclEncoding.c +++ b/generic/tclEncoding.c @@ -2565,6 +2565,13 @@ Utf32ToUtfProc( } else { ch = (src[0] & 0xFF) << 24 | (src[1] & 0xFF) << 16 | (src[2] & 0xFF) << 8 | (src[3] & 0xFF); } + if (ch >= 0x10FFFF || (((flags & TCL_ENCODING_STRICT) == TCL_ENCODING_STRICT) + && !Tcl_UniCharIsUnicode(ch))) { + if (STOPONERROR) { + result = TCL_CONVERT_SYNTAX; + break; + } + } /* * Special case for 1-byte utf chars for speed. Make sure we work with @@ -2655,12 +2662,11 @@ UtfToUtf32Proc( break; } len = TclUtfToUCS4(src, &ch); - if (!Tcl_UniCharIsUnicode(ch)) { + if (!Tcl_UniCharIsUnicode(ch) && (((ch & ~0x7FF) == 0xD800) || ((flags & TCL_ENCODING_STRICT) == TCL_ENCODING_STRICT))) { if (STOPONERROR) { result = TCL_CONVERT_UNKNOWN; break; } - ch = 0xFFFD; } src += len; if (flags & TCL_ENCODING_LE) { @@ -2858,12 +2864,11 @@ UtfToUtf16Proc( break; } len = TclUtfToUCS4(src, &ch); - if (!Tcl_UniCharIsUnicode(ch)) { + if (!Tcl_UniCharIsUnicode(ch) && (((ch & ~0x7FF) == 0xD800) || ((flags & TCL_ENCODING_STRICT) == TCL_ENCODING_STRICT))) { if (STOPONERROR) { result = TCL_CONVERT_UNKNOWN; break; } - ch = 0xFFFD; } src += len; if (flags & TCL_ENCODING_LE) { |
