diff options
| -rw-r--r-- | generic/tclEncoding.c | 30 | ||||
| -rw-r--r-- | tests/chanio.test | 2 |
2 files changed, 18 insertions, 14 deletions
diff --git a/generic/tclEncoding.c b/generic/tclEncoding.c index 0478519..69b7b6c 100644 --- a/generic/tclEncoding.c +++ b/generic/tclEncoding.c @@ -3413,18 +3413,22 @@ TableToUtfProc( if (prefixBytes[byte]) { src++; if (src >= srcEnd) { - /* - * TODO - this is broken. For consistency with other - * decoders, an error should be raised only if strict. - * However, doing that check cause a whole bunch of test - * failures. Need to verify if those tests are in fact - * correct. - */ - src--; - result = TCL_CONVERT_MULTIBYTE; - break; + if (!(flags & TCL_ENCODING_END)) { + src--; + result = TCL_CONVERT_MULTIBYTE; + break; + } else if (PROFILE_STRICT(flags)) { + src--; + result = TCL_CONVERT_SYNTAX; + break; + } else if (PROFILE_REPLACE(flags)) { + ch = UNICODE_REPLACE_CHAR; + } else { + ch = (Tcl_UniChar)byte; + } + } else { + ch = toUnicode[byte][*((unsigned char *)src)]; } - ch = toUnicode[byte][*((unsigned char *)src)]; } else { ch = pageZero[byte]; } @@ -3447,7 +3451,7 @@ TableToUtfProc( * Special case for 1-byte utf chars for speed. */ - if (ch && ch < 0x80) { + if ((unsigned)ch - 1 < 0x7F) { *dst++ = (char) ch; } else { dst += Tcl_UniCharToUtf(ch, dst); @@ -3648,7 +3652,7 @@ Iso88591ToUtfProc( * Special case for 1-byte utf chars for speed. */ - if (ch && ch < 0x80) { + if ((unsigned)ch - 1 < 0x7F) { *dst++ = (char) ch; } else { dst += Tcl_UniCharToUtf(ch, dst); diff --git a/tests/chanio.test b/tests/chanio.test index d2008e6..7221141 100644 --- a/tests/chanio.test +++ b/tests/chanio.test @@ -1104,7 +1104,7 @@ test chan-io-7.3 {FilterInputBytes: split up character at EOF} -setup { lappend x [chan gets $f line] $line } -cleanup { chan close $f -} -result [list 15 "1234567890123０１" 18 0 1 -1 ""] +} -result [list 16 "1234567890123０１\x82" 18 0 1 -1 ""] test chan-io-7.4 {FilterInputBytes: recover from split up character} -setup { variable x "" } -constraints {stdio fileevent} -body { |