Merge branch 'gawk-4.1-stable'

11 files changed, 117 insertions, 17 deletions

diff --git a/ChangeLog b/ChangeLog
index ff048b2f..f80edb7f 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,21 @@
+2015-08-25         Arnold D. Robbins     <arnold@skeeve.com>
+
+	* node.c (str2wstr): Upon finding an invalid character, if
+	using UTF-8, use the replacement character instead of skipping
+	it. Helps match() and other functions work better in the face
+	of unexpected data.  Make the lint warning an unconditional
+	warning.
+
+	Unrelated:
+
+	* awk.h: Add explanatory comment on the flags related to
+	types and values.
+	* mpfr.c (mpg_force_number): If setting NUMBER, clear STRING also
+	when clearing MAYBE_NUM.
+	(set_PREC): Check STRCUR instead of STRING.
+	* node.c (r_force_number): If setting NUMBER, clear STRING also
+	when clearing MAYBE_NUM.
+
 2015-08-15         Arnold D. Robbins     <arnold@skeeve.com>
 
 	* dfa.c (dfamust): Restore c90 compat by moving some
diff --git a/awk.h b/awk.h
index 53af7f1b..8bf78fcd 100644
--- a/awk.h
+++ b/awk.h
@@ -404,6 +404,37 @@ typedef struct exp_node {
 #		define	MALLOC	0x0001       /* can be free'd */
 
 /* type = Node_val */
+	/*
+	 * STRING and NUMBER are mutually exclusive. They represent the
+	 * type of a value as assigned.
+	 *
+	 * STRCUR and NUMCUR are not mutually exclusive. They represent that
+	 * the particular type of value is up to date.  For example,
+	 *
+	 * 	a = 5		# NUMBER | NUMCUR
+	 * 	b = a ""	# Adds STRCUR to a, since a string value
+	 * 			# is now available. But the type hasn't changed!
+	 *
+	 * 	a = "42"	# STRING | STRCUR
+	 * 	b = a + 0	# Adds NUMCUR to a, since numeric value
+	 * 			# is now available. But the type hasn't changed!
+	 *
+	 * MAYBE_NUM is the joker.  It means "this is string data, but
+	 * the user may have really wanted it to be a number. If we have
+	 * to guess, like in a comparison, turn it into a number."
+	 * For example,    gawk -v a=42 ....
+	 * Here, `a' gets STRING|STRCUR|MAYBE_NUM and then when used where
+	 * a number is needed, it gets turned into a NUMBER and STRING
+	 * is cleared.
+	 *
+	 * WSTRCUR is for efficiency. If in a multibyte locale, and we
+	 * need to do something character based (substr, length, etc.)
+	 * we create the corresponding wide character string and store it,
+	 * and add WSTRCUR to the flags so that we don't have to do the
+	 * conversion more than once.
+	 *
+	 * We hope that the rest of the flags are self-explanatory. :-)
+	 */
 #		define	STRING	0x0002       /* assigned as string */
 #		define	STRCUR	0x0004       /* string value is current */
 #		define	NUMCUR	0x0008       /* numeric value is current */
diff --git a/mpfr.c b/mpfr.c
index 080ed7fa..b7526cfb 100644
--- a/mpfr.c
+++ b/mpfr.c
@@ -347,7 +347,7 @@ mpg_force_number(NODE *n)
 		return n;
 
 	if ((n->flags & MAYBE_NUM) != 0) {
-		n->flags &= ~MAYBE_NUM;
+		n->flags &= ~(MAYBE_NUM|STRING);
 		newflags = NUMBER;
 	}
 
@@ -525,7 +525,7 @@ set_PREC()
 	if ((val->flags & MAYBE_NUM) != 0)
 		force_number(val);
 
-	if ((val->flags & (STRING|NUMBER)) == STRING) {
+	if ((val->flags & STRCUR) != 0) {
 		int i, j;
 
 		/* emulate IEEE-754 binary format */
diff --git a/node.c b/node.c
index b15159f0..9227cf2d 100644
--- a/node.c
+++ b/node.c
@@ -76,7 +76,7 @@ r_force_number(NODE *n)
 			return n;
 		} else if (n->stlen == 4 && is_ieee_magic_val(n->stptr)) {
 			if ((n->flags & MAYBE_NUM) != 0)
-				n->flags &= ~MAYBE_NUM;
+				n->flags &= ~(MAYBE_NUM|STRING);
 			n->flags |= NUMBER|NUMCUR;
 			n->numbr = get_ieee_magic_val(n->stptr);
 
@@ -103,7 +103,7 @@ r_force_number(NODE *n)
 
 	if ((n->flags & MAYBE_NUM) != 0) {
 		newflags = NUMBER;
-		n->flags &= ~MAYBE_NUM;
+		n->flags &= ~(MAYBE_NUM|STRING);
 	} else
 		newflags = 0;
 
@@ -716,22 +716,37 @@ str2wstr(NODE *n, size_t **ptr)
 		case (size_t) -2:
 		case (size_t) -1:
 			/*
-			 * Just skip the bad byte and keep going, so that
-			 * we get a more-or-less full string, instead of
-			 * stopping early. This is particularly important
-			 * for match() where we need to build the indices.
-			 */
-			sp++;
-			src_count--;
-			/*
 			 * mbrtowc(3) says the state of mbs becomes undefined
 			 * after a bad character, so reset it.
 			 */
 			memset(& mbs, 0, sizeof(mbs));
-			/* And warn the user something's wrong */
-			if (do_lint && ! warned) {
+
+			/* Warn the user something's wrong */
+			if (! warned) {
 				warned = true;
-				lintwarn(_("Invalid multibyte data detected. There may be a mismatch between your data and your locale."));
+				warning(_("Invalid multibyte data detected. There may be a mismatch between your data and your locale."));
+			}
+
+			/*
+			 * 8/2015: If we're using UTF, then instead of just
+			 * skipping the character, plug in the Unicode
+			 * replacement character. In most cases this gives
+			 * us "better" results, in that character counts
+			 * and string lengths tend to make more sense.
+			 *
+			 * Otherwise, just skip the bad byte and keep going,
+			 * so that we get a more-or-less full string, instead of
+			 * stopping early. This is particularly important
+			 * for match() where we need to build the indices.
+			 */
+			if (using_utf8()) {
+				count = 1;
+				wc = 0xFFFD;	/* unicode replacement character */
+				goto set_wc;
+			} else {
+				/* skip it and keep going */
+				sp++;
+				src_count--;
 			}
 			break;
 
@@ -739,6 +754,7 @@ str2wstr(NODE *n, size_t **ptr)
 			count = 1;
 			/* fall through */
 		default:
+		set_wc:
 			*wsp++ = wc;
 			src_count -= count;
 			while (count--)  {
diff --git a/test/ChangeLog b/test/ChangeLog
index 8cc7915e..736860db 100644
--- a/test/ChangeLog
+++ b/test/ChangeLog
@@ -1,3 +1,9 @@
+2015-08-25         Arnold D. Robbins     <arnold@skeeve.com>
+
+	* mbstr1.ok: Updated after code change.
+	* Makefile.am (mbstr2): New test.
+	* mbstr2.awk, mbstr2.in, mbstr2.ok: New files.
+
 2015-06-29         Arnold D. Robbins     <arnold@skeeve.com>
 
 	* Makefile.am (dbugeval2, typedregex3): New tests.
diff --git a/test/Makefile.am b/test/Makefile.am
index b9e14468..68fd709f 100644
--- a/test/Makefile.am
+++ b/test/Makefile.am
@@ -547,6 +547,9 @@ EXTRA_DIST = \
 	mbprintf4.ok \
 	mbstr1.awk \
 	mbstr1.ok \
+	mbstr2.awk \
+	mbstr2.in \
+	mbstr2.ok \
 	membug1.awk \
 	membug1.in \
 	membug1.ok \
@@ -1107,7 +1110,7 @@ GAWK_EXT_TESTS = \
 	incdupe incdupe2 incdupe3 incdupe4 incdupe5 incdupe6 incdupe7 \
 	include include2 indirectbuiltin indirectcall indirectcall2 \
 	lint lintold lintwarn \
-	manyfiles match1 match2 match3 mbstr1 \
+	manyfiles match1 match2 match3 mbstr1 mbstr2 \
 	nastyparm negtime next nondec nondec2 \
 	nonfatal1 nonfatal2 nonfatal3 \
 	patsplit posix printfbad1 printfbad2 printfbad3 printfbad4 printhuge procinfs \
@@ -1760,6 +1763,12 @@ mbstr1::
 	AWKPATH="$(srcdir)" $(AWK) -f $@.awk  >_$@ 2>&1 || echo EXIT CODE: $$? >>_$@
 	@-$(CMP) "$(srcdir)"/$@.ok _$@ && rm -f _$@
 
+mbstr2::
+	@echo $@
+	@[ -z "$$GAWKLOCALE" ] && GAWKLOCALE=en_US.UTF-8; \
+	AWKPATH="$(srcdir)" $(AWK) -f $@.awk < "$(srcdir)"/$@.in >_$@ 2>&1 || echo EXIT CODE: $$? >>_$@
+	@-$(CMP) "$(srcdir)"/$@.ok _$@ && rm -f _$@
+
 printfbad2: printfbad2.ok
 	@echo $@
 	@$(AWK) --lint -f "$(srcdir)"/$@.awk "$(srcdir)"/$@.in 2>&1 | sed 's;$(srcdir)/;;g' >_$@ || echo EXIT CODE: $$?  >>_$@
diff --git a/test/Makefile.in b/test/Makefile.in
index 0ae51917..1fe8da66 100644
--- a/test/Makefile.in
+++ b/test/Makefile.in
@@ -804,6 +804,9 @@ EXTRA_DIST = \
 	mbprintf4.ok \
 	mbstr1.awk \
 	mbstr1.ok \
+	mbstr2.awk \
+	mbstr2.in \
+	mbstr2.ok \
 	membug1.awk \
 	membug1.in \
 	membug1.ok \
@@ -1363,7 +1366,7 @@ GAWK_EXT_TESTS = \
 	incdupe incdupe2 incdupe3 incdupe4 incdupe5 incdupe6 incdupe7 \
 	include include2 indirectbuiltin indirectcall indirectcall2 \
 	lint lintold lintwarn \
-	manyfiles match1 match2 match3 mbstr1 \
+	manyfiles match1 match2 match3 mbstr1 mbstr2 \
 	nastyparm negtime next nondec nondec2 \
 	nonfatal1 nonfatal2 nonfatal3 \
 	patsplit posix printfbad1 printfbad2 printfbad3 printfbad4 printhuge procinfs \
@@ -2198,6 +2201,12 @@ mbstr1::
 	AWKPATH="$(srcdir)" $(AWK) -f $@.awk  >_$@ 2>&1 || echo EXIT CODE: $$? >>_$@
 	@-$(CMP) "$(srcdir)"/$@.ok _$@ && rm -f _$@
 
+mbstr2::
+	@echo $@
+	@[ -z "$$GAWKLOCALE" ] && GAWKLOCALE=en_US.UTF-8; \
+	AWKPATH="$(srcdir)" $(AWK) -f $@.awk < "$(srcdir)"/$@.in >_$@ 2>&1 || echo EXIT CODE: $$? >>_$@
+	@-$(CMP) "$(srcdir)"/$@.ok _$@ && rm -f _$@
+
 printfbad2: printfbad2.ok
 	@echo $@
 	@$(AWK) --lint -f "$(srcdir)"/$@.awk "$(srcdir)"/$@.in 2>&1 | sed 's;$(srcdir)/;;g' >_$@ || echo EXIT CODE: $$?  >>_$@
diff --git a/test/mbstr1.ok b/test/mbstr1.ok
index dcb4347b..3fd1bf8e 100644
--- a/test/mbstr1.ok
+++ b/test/mbstr1.ok
@@ -1,2 +1,3 @@
+gawk: mbstr1.awk:2: warning: Invalid multibyte data detected. There may be a mismatch between your data and your locale.
 4
 1
diff --git a/test/mbstr2.awk b/test/mbstr2.awk
new file mode 100644
index 00000000..4f2c8cc6
--- /dev/null
+++ b/test/mbstr2.awk
@@ -0,0 +1 @@
+match($0,/:deathdate=2007....:/) { print substr($0,RSTART+11,RLENGTH-16) }
diff --git a/test/mbstr2.in b/test/mbstr2.in
new file mode 100644
index 00000000..36e971a6
--- /dev/null
+++ b/test/mbstr2.in
@@ -0,0 +1,4 @@
+missile:deathdate=20070306:
+P”rr”:deathdate=20070306:
+wizard:deathdate=20071103:
+Daithí:deathdate=20071103:
diff --git a/test/mbstr2.ok b/test/mbstr2.ok
new file mode 100644
index 00000000..29ac876a
--- /dev/null
+++ b/test/mbstr2.ok
@@ -0,0 +1,5 @@
+2007
+gawk: mbstr2.awk:1: (FILENAME=- FNR=2) warning: Invalid multibyte data detected. There may be a mismatch between your data and your locale.
+2007
+2007
+2007