diff options
| author | jan.nijtmans <nijtmans@users.sourceforge.net> | 2020-11-18 13:51:02 (GMT) |
|---|---|---|
| committer | jan.nijtmans <nijtmans@users.sourceforge.net> | 2020-11-18 13:51:02 (GMT) |
| commit | 8e1f957a669b2f4b84dca7e8a27f2985c0625172 (patch) | |
| tree | dba1e5192d3d569c9588d12db8188ae501e1e0e9 /generic/tclUtf.c | |
| parent | c55eb93bb4ebc7ba1de131099e9c0c4a1c0ff4b2 (diff) | |
| download | tcl-8e1f957a669b2f4b84dca7e8a27f2985c0625172.zip tcl-8e1f957a669b2f4b84dca7e8a27f2985c0625172.tar.gz tcl-8e1f957a669b2f4b84dca7e8a27f2985c0625172.tar.bz2 | |
More usage of TclUtfToUCS4/TclUniCharToUCS4 in stead of it's UniChar variants: This handles surrogate pairs better.
Diffstat (limited to 'generic/tclUtf.c')
| -rw-r--r-- | generic/tclUtf.c | 15 |
1 files changed, 14 insertions, 1 deletions
diff --git a/generic/tclUtf.c b/generic/tclUtf.c index 11bde5c..525cd50 100644 --- a/generic/tclUtf.c +++ b/generic/tclUtf.c @@ -2629,12 +2629,25 @@ TclUniCharToUCS4( * by the Tcl_UniChar string. */ { if (((src[0] & 0xFC00) == 0xD800) && ((src[1] & 0xFC00) == 0xDC00)) { - *ucs4Ptr = (((src[0] & 0x3FF) << 10) | (src[01] & 0x3FF)) + 0x10000; + *ucs4Ptr = (((src[0] & 0x3FF) << 10) | (src[1] & 0x3FF)) + 0x10000; return 2; } *ucs4Ptr = src[0]; return 1; } + +const Tcl_UniChar *TclUCS4Prev(const Tcl_UniChar *src, const Tcl_UniChar *ptr) { + if (src <= ptr + 1) { + return ptr; + } + if (((src[-1] & 0xFC00) == 0xDC00) && ((src[-2] & 0xFC00) == 0xD800)) { + return src - 2; + } + return src - 1; +} + + + #endif /* |
