diff options
Diffstat (limited to 'generic/tclEncoding.c')
| -rw-r--r-- | generic/tclEncoding.c | 19 |
1 files changed, 13 insertions, 6 deletions
diff --git a/generic/tclEncoding.c b/generic/tclEncoding.c index 350a13c..a326856 100644 --- a/generic/tclEncoding.c +++ b/generic/tclEncoding.c @@ -2409,17 +2409,24 @@ UtfToUtfProc( dst += Tcl_UniCharToUtf(*chPtr, dst); } else { src += TclUtfToUniChar(src, chPtr); - if ((*chPtr | 0x3FF) == 0xDBFF) { - /* A high surrogate character is detected, handle especially */ + if ((*chPtr | 0x7FF) == 0xDFFF) { + /* A surrogate character is detected, handle especially */ Tcl_UniChar low = *chPtr; - if (src <= srcEnd-3) { - Tcl_UtfToUniChar(src, &low); - } - if ((low | 0x3FF) != 0xDFFF) { + size_t len = (src <= srcEnd-3) ? Tcl_UtfToUniChar(src, &low) : 0; + if (((low | 0x3FF) != 0xDFFF) || !(*chPtr & 0x800)) { *dst++ = (char) (((*chPtr >> 12) | 0xE0) & 0xEF); *dst++ = (char) (((*chPtr >> 6) | 0x80) & 0xBF); *dst++ = (char) ((*chPtr | 0x80) & 0xBF); continue; + } else if (TCL_UTF_MAX <= 4) { + int full = (((*chPtr & 0x3FF) << 10) | (low & 0x3FF)) + 0x10000; + *dst++ = (char) (((full >> 18) | 0xF0) & 0xF7); + *dst++ = (char) (((full >> 12) | 0x80) & 0xBF); + *dst++ = (char) (((full >> 6) | 0x80) & 0xBF); + *dst++ = (char) ((full | 0x80) & 0xBF); + *chPtr = 0; + src += len; + continue; } } dst += Tcl_UniCharToUtf(*chPtr, dst); |
