summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
Diffstat
-rw-r--r--generic/regc_lex.c10
-rw-r--r--generic/tclEncoding.c11
-rw-r--r--generic/tclInt.h2
-rw-r--r--generic/tclParse.c14
-rw-r--r--tests/encoding.test15
-rw-r--r--tests/reg.test1
6 files changed, 27 insertions, 26 deletions
diff --git a/generic/regc_lex.c b/generic/regc_lex.c
index 1b00b71..2208c0e 100644
--- a/generic/regc_lex.c
+++ b/generic/regc_lex.c
@@ -843,12 +843,18 @@ lexescape(
if (ISERR()) {
FAILW(REG_EESCAPE);
}
- if (i > 0xFFFF) {
+#if CHRBITS > 16
+ if ((unsigned)i > 0x10FFFF) {
+ i = 0xFFFD;
+ }
+#else
+ if ((unsigned)i & ~0xFFFF) {
/* TODO: output a Surrogate pair
*/
i = 0xFFFD;
}
- RETV(PLAIN, (uchr) i);
+#endif
+ RETV(PLAIN, (uchr)i);
break;
case CHR('v'):
RETV(PLAIN, CHR('\v'));
diff --git a/generic/tclEncoding.c b/generic/tclEncoding.c
index e012570..6ab0510 100644
--- a/generic/tclEncoding.c
+++ b/generic/tclEncoding.c
@@ -2557,11 +2557,6 @@ UtfToUnicodeProc(
}
src += TclUtfToUniChar(src, chPtr);
- /*
- * Need to handle this in a way that won't cause misalignment by
- * casting dst to a Tcl_UniChar. [Bug 1122671]
- */
-
if (clientData) {
#if TCL_UTF_MAX > 4
if (*chPtr <= 0xFFFF) {
@@ -2793,11 +2788,7 @@ TableFromUtfProc(
len = TclUtfToUniChar(src, &ch);
#if TCL_UTF_MAX > 4
- /*
- * This prevents a crash condition. More evaluation is required for
- * full support of int Tcl_UniChar. [Bug 1004065]
- */
-
+ /* Unicode chars > +U0FFFF cannot be represented in any table encoding */
if (ch & 0xFFFF0000) {
word = 0;
} else
diff --git a/generic/tclInt.h b/generic/tclInt.h
index 780ea30..5df9aac 100644
--- a/generic/tclInt.h
+++ b/generic/tclInt.h
@@ -3065,8 +3065,6 @@ MODULE_SCOPE int TclObjUnsetVar2(Tcl_Interp *interp,
Tcl_Obj *part1Ptr, Tcl_Obj *part2Ptr, int flags);
MODULE_SCOPE int TclParseBackslash(const char *src,
int numBytes, int *readPtr, char *dst);
-MODULE_SCOPE int TclParseHex(const char *src, int numBytes,
- int *resultPtr);
MODULE_SCOPE int TclParseNumber(Tcl_Interp *interp, Tcl_Obj *objPtr,
const char *expected, const char *bytes,
int numBytes, const char **endPtrPtr, int flags);
diff --git a/generic/tclParse.c b/generic/tclParse.c
index 4d7e6b8..f834480 100644
--- a/generic/tclParse.c
+++ b/generic/tclParse.c
@@ -167,6 +167,8 @@ static int ParseTokens(const char *src, int numBytes, int mask,
int flags, Tcl_Parse *parsePtr);
static int ParseWhiteSpace(const char *src, int numBytes,
int *incompletePtr, char *typePtr);
+static int ParseHex(const char *src, int numBytes,
+ int *resultPtr);
/*
*----------------------------------------------------------------------
@@ -754,7 +756,7 @@ TclParseAllWhiteSpace(
/*
*----------------------------------------------------------------------
*
- * TclParseHex --
+ * ParseHex --
*
* Scans a hexadecimal number as a Tcl_UniChar value (e.g., for parsing
* \x and \u escape sequences). At most numBytes bytes are scanned.
@@ -774,7 +776,7 @@ TclParseAllWhiteSpace(
*/
int
-TclParseHex(
+ParseHex(
const char *src, /* First character to parse. */
int numBytes, /* Max number of byes to scan */
int *resultPtr) /* Points to storage provided by caller where
@@ -899,7 +901,7 @@ TclParseBackslash(
result = 0xB;
break;
case 'x':
- count += TclParseHex(p+1, (numBytes > 3) ? 2 : numBytes-2, &result);
+ count += ParseHex(p+1, (numBytes > 3) ? 2 : numBytes-2, &result);
if (count == 2) {
/*
* No hexadigits -> This is just "x".
@@ -914,7 +916,7 @@ TclParseBackslash(
}
break;
case 'u':
- count += TclParseHex(p+1, (numBytes > 5) ? 4 : numBytes-2, &result);
+ count += ParseHex(p+1, (numBytes > 5) ? 4 : numBytes-2, &result);
if (count == 2) {
/*
* No hexadigits -> This is just "u".
@@ -926,7 +928,7 @@ TclParseBackslash(
/* If high surrogate is immediately followed by a low surrogate
* escape, combine them into one character. */
int low;
- int count2 = TclParseHex(p+7, 4, &low);
+ int count2 = ParseHex(p+7, 4, &low);
if ((count2 == 4) && ((low & 0xDC00) == 0xDC00)) {
result = ((result & 0x3FF)<<10 | (low & 0x3FF)) + 0x10000;
count += count2 + 2;
@@ -935,7 +937,7 @@ TclParseBackslash(
}
break;
case 'U':
- count += TclParseHex(p+1, (numBytes > 9) ? 8 : numBytes-2, &result);
+ count += ParseHex(p+1, (numBytes > 9) ? 8 : numBytes-2, &result);
if (count == 2) {
/*
* No hexadigits -> This is just "U".
diff --git a/tests/encoding.test b/tests/encoding.test
index 6fef748..a8ce162 100644
--- a/tests/encoding.test
+++ b/tests/encoding.test
@@ -36,7 +36,7 @@ proc runtests {} {
testConstraint testencoding [llength [info commands testencoding]]
testConstraint testbytestring [llength [info commands testbytestring]]
testConstraint teststringbytes [llength [info commands teststringbytes]]
-testConstraint fullutf [expr {[format %c 0x010000] != "\ufffd"}]
+testConstraint fullutf [expr {[format %c 0x010000] ne "\ufffd"}]
testConstraint exec [llength [info commands exec]]
testConstraint testgetdefenc [llength [info commands testgetdefenc]]
@@ -284,16 +284,16 @@ test encoding-11.6 {LoadEncodingFile: invalid file} -constraints {testencoding}
# OpenEncodingFile is fully tested by the rest of the tests in this file.
test encoding-12.1 {LoadTableEncoding: normal encoding} {
- set x [encoding convertto iso8859-3 \u120]
- append x [encoding convertto iso8859-3 \ud5]
- append x [encoding convertfrom iso8859-3 \xd5]
+ set x [encoding convertto iso8859-3 \u0120]
+ append x [encoding convertto iso8859-3 \xD5]
+ append x [encoding convertfrom iso8859-3 \xD5]
} "\xd5?\u120"
test encoding-12.2 {LoadTableEncoding: single-byte encoding} {
set x [encoding convertto iso8859-3 ab\u0120g]
- append x [encoding convertfrom iso8859-3 ab\xd5g]
+ append x [encoding convertfrom iso8859-3 ab\xD5g]
} "ab\xd5gab\u120g"
test encoding-12.3 {LoadTableEncoding: multi-byte encoding} {
- set x [encoding convertto shiftjis ab\u4e4eg]
+ set x [encoding convertto shiftjis ab\u4E4Eg]
append x [encoding convertfrom shiftjis ab\x8c\xc1g]
} "ab\x8c\xc1gab\u4e4eg"
test encoding-12.4 {LoadTableEncoding: double-byte encoding} {
@@ -305,6 +305,9 @@ test encoding-12.5 {LoadTableEncoding: symbol encoding} {
append x [encoding convertto symbol \u67]
append x [encoding convertfrom symbol \x67]
} "\x67\x67\u3b3"
+test encoding-12.6 {LoadTableEncoding: overflow in char value} fullutf {
+ encoding convertto iso8859-3 \U010000
+} "?"
test encoding-13.1 {LoadEscapeTable} {
viewable [set x [encoding convertto iso2022 ab\u4e4e\u68d9g]]
diff --git a/tests/reg.test b/tests/reg.test
index d040632..a95d1e2 100644
--- a/tests/reg.test
+++ b/tests/reg.test
@@ -626,6 +626,7 @@ expectMatch 13.14 P "a\\rb" "a\rb" "a\rb"
expectMatch 13.15 P "a\\tb" "a\tb" "a\tb"
expectMatch 13.16 P "a\\u0008x" "a\bx" "a\bx"
expectMatch 13.17 P {a\u008x} "a\bx" "a\bx"
+expectError 13.17.1 - {a\ux} EESCAPE
expectMatch 13.18 P "a\\u00088x" "a\b8x" "a\b8x"
expectMatch 13.19 P "a\\U00000008x" "a\bx" "a\bx"
expectMatch 13.20 P {a\U0000008x} "a\bx" "a\bx"
generated by cgit v0.12 at 2025-09-05 10:01:48 (GMT)