dfa: fix handling of ranges in multibyte character sets

* src/dfa.c (parse_bracket_exp_mb): Add separate ranges for
lowercase and uppercase endpoints if folding case.
* tests/Makefile.am (TESTS): Add case-fold-char-range.
* tests/case-fold-char-range: New.

4 files changed, 44 insertions, 7 deletions

diff --git a/NEWS b/NEWS
index 2ade4fc7..4f9c7787 100644
--- a/NEWS
+++ b/NEWS
@@ -4,11 +4,11 @@ GNU grep NEWS                                    -*- outline -*-
 
 ** Bug fixes
 
-  grep -i with a character class would malfunction in multi-byte locales.
-  For example, echo Y | LC_ALL=en_US.UTF-8 grep -i '[y]' would print nothing.
-  Character types would malfunction in multi-byte locales similarly; for
-  example, echo Y | LC_ALL=en_US.UTF-8 grep -i '[[:lower:]]' would print
-  nothing.
+  Character classes would malfunction in multi-byte locales when using grep -i.
+  Examples which would print nothing for LC_ALL=en_US.UTF-8 include:
+  - for ranges, echo Z | grep -i '[a-z]'
+  - for single characters, echo Y | grep -i '[y]'
+  - for character types, echo Y | grep -i '[[:lower:]]'
 
   Various bugs in grep -P, caused by expressions such as [^b] or \S matching
   newlines, were fixed.  grep -P also supports the special sequences \Z and
diff --git a/src/dfa.c b/src/dfa.c
index 6c7494ec..90e3c187 100644
--- a/src/dfa.c
+++ b/src/dfa.c
@@ -569,6 +569,8 @@ parse_bracket_exp_mb (void)
 	      wc1 = fetch_wc(_("unbalanced ["));
 	    }
 
+      	  /* When case folding map a range, say [m-z] (or even [M-z]) to the
+     	     pair of ranges, [m-z] [M-Z].  */
 	  if (range_sts_al == 0)
 	    {
 	      MALLOC(work_mbc->range_sts, wchar_t, ++range_sts_al);
@@ -576,10 +578,22 @@ parse_bracket_exp_mb (void)
 	    }
 	  REALLOC_IF_NECESSARY(work_mbc->range_sts, wchar_t,
 			       range_sts_al, work_mbc->nranges + 1);
-	  work_mbc->range_sts[work_mbc->nranges] = (wchar_t)wc;
 	  REALLOC_IF_NECESSARY(work_mbc->range_ends, wchar_t,
 			       range_ends_al, work_mbc->nranges + 1);
-	  work_mbc->range_ends[work_mbc->nranges++] = (wchar_t)wc2;
+	  work_mbc->range_sts[work_mbc->nranges] =
+            case_fold ? towlower(wc) : (wchar_t)wc;
+	  work_mbc->range_ends[work_mbc->nranges++] =
+            case_fold ? towlower(wc2) : (wchar_t)wc2;
+
+	  if (case_fold)
+            {
+              REALLOC_IF_NECESSARY(work_mbc->range_sts, wchar_t,
+                                   range_sts_al, work_mbc->nranges + 1);
+              work_mbc->range_sts[work_mbc->nranges] = towupper(wc);
+              REALLOC_IF_NECESSARY(work_mbc->range_ends, wchar_t,
+                                   range_ends_al, work_mbc->nranges + 1);
+              work_mbc->range_ends[work_mbc->nranges++] = towupper(wc2);
+            }
 	}
       else if (wc != WEOF)
 	/* build normal characters.  */
diff --git a/tests/Makefile.am b/tests/Makefile.am
index b71b2ee4..6a56dcff 100644
--- a/tests/Makefile.am
+++ b/tests/Makefile.am
@@ -18,6 +18,7 @@ TESTS =						\
   backref.sh					\
   bre.sh					\
   case-fold-char-class				\
+  case-fold-char-range				\
   case-fold-char-type				\
   dfaexec-multibyte				\
   empty.sh					\
diff --git a/tests/case-fold-char-range b/tests/case-fold-char-range
new file mode 100644
index 00000000..9b3120fb
--- /dev/null
+++ b/tests/case-fold-char-range
@@ -0,0 +1,22 @@
+#!/bin/sh
+# This would fail for grep-2.5.3
+: ${srcdir=.}
+. "$srcdir/init.sh"; path_prepend_ ../src
+
+printf 'A\nZ\n'      > exp1 || framework_failure
+fail=0
+
+for LOC in en_US.UTF-8 zh_CN $LOCALE_FR_UTF8; do
+  printf 'A\n1\nZ\n.\n' | LC_ALL=$LOC grep -i '[a-z]' > out1 || fail=1
+  compare out1 exp1 || fail=1
+done
+
+# This actually passes also for grep-2.5.3
+printf 'a\nz\n'      > exp2 || framework_failure
+
+for LOC in en_US.UTF-8 zh_CN $LOCALE_FR_UTF8; do
+  printf 'a\n1\nz\n.\n' | LC_ALL=$LOC grep -i '[A-Z]' > out2 || fail=1
+  compare out2 exp2 || fail=1
+done
+
+Exit $fail