Add support for Unicode whitespace in [:blank:]

See Bug#25366.

* src/character.c (blankp): New function for checking Unicode
horizontal whitespace.
* src/regex.c (ISBLANK): Use 'blankp' for non-ASCII horizontal
whitespace.
(BIT_BLANK): New bit for range table.
(re_wctype_to_bit, execute_charset): Use it.
* test/lisp/subr-tests.el (subr-tests--string-match-p--blank): Add
unit test for [:blank:] character class.
* test/src/regex-tests.el (test): Adapt unit test.
* doc/lispref/searching.texi (Char Classes): Document new Unicode
behavior for [:blank:].

3 files changed, 26 insertions, 4 deletions

diff --git a/src/character.c b/src/character.c
index b594af040c1..bc99daf0df0 100644
--- a/src/character.c
+++ b/src/character.c
@@ -1038,6 +1038,23 @@ printablep (int c)
             || gen_cat == UNICODE_CATEGORY_Cn)); /* unassigned */
 }
 
+/* Return true if C is a horizontal whitespace character, as defined
+   by http://www.unicode.org/reports/tr18/tr18-19.html#blank.  */
+bool
+blankp (int c)
+{
+  /* Fast path for ASCII characters that are always assumed to
+     constitute horizontal whitespace.  */
+  if (c == ' ' || c == '\t')
+    return true;
+
+  Lisp_Object category = CHAR_TABLE_REF (Vunicode_category_table, c);
+  if (! INTEGERP (category))
+    return false;
+
+  return XINT (category) == UNICODE_CATEGORY_Zs; /* separator, space */
+}
+
 void
 syms_of_character (void)
 {
diff --git a/src/character.h b/src/character.h
index fc8a0dd74d2..62d252e91ba 100644
--- a/src/character.h
+++ b/src/character.h
@@ -680,6 +680,7 @@ extern bool alphabeticp (int);
 extern bool alphanumericp (int);
 extern bool graphicp (int);
 extern bool printablep (int);
+extern bool blankp (int);
 
 /* Return a translation table of id number ID.  */
 #define GET_TRANSLATION_TABLE(id) \
diff --git a/src/regex.c b/src/regex.c
index ae3fde80c9e..7e70c494f47 100644
--- a/src/regex.c
+++ b/src/regex.c
@@ -310,11 +310,12 @@ enum syntaxcode { Swhitespace = 0, Sword = 1, Ssymbol = 2 };
                      || ((c) >= 'a' && (c) <= 'f')      \
                      || ((c) >= 'A' && (c) <= 'F'))
 
-/* This is only used for single-byte characters.  */
-# define ISBLANK(c) ((c) == ' ' || (c) == '\t')
-
 /* The rest must handle multibyte characters.  */
 
+# define ISBLANK(c) (IS_REAL_ASCII (c)                  \
+                     ? ((c) == ' ' || (c) == '\t')      \
+                     : blankp (c))
+
 # define ISGRAPH(c) (SINGLE_BYTE_CHAR_P (c)                             \
                      ? (c) > ' ' && !((c) >= 0177 && (c) <= 0240)       \
                      : graphicp (c))
@@ -1790,6 +1791,7 @@ struct range_table_work_area
 #define BIT_ALNUM       0x80
 #define BIT_GRAPH       0x100
 #define BIT_PRINT       0x200
+#define BIT_BLANK       0x400
 
 
 /* Set the bit for character C in a list.  */
@@ -2066,8 +2068,9 @@ re_wctype_to_bit (re_wctype_t cc)
     case RECC_SPACE: return BIT_SPACE;
     case RECC_GRAPH: return BIT_GRAPH;
     case RECC_PRINT: return BIT_PRINT;
+    case RECC_BLANK: return BIT_BLANK;
     case RECC_ASCII: case RECC_DIGIT: case RECC_XDIGIT: case RECC_CNTRL:
-    case RECC_BLANK: case RECC_UNIBYTE: case RECC_ERROR: return 0;
+    case RECC_UNIBYTE: case RECC_ERROR: return 0;
     default:
       abort ();
     }
@@ -4658,6 +4661,7 @@ execute_charset (const_re_char **pp, unsigned c, unsigned corig, bool unibyte)
           (class_bits & BIT_ALNUM && ISALNUM (c)) ||
           (class_bits & BIT_ALPHA && ISALPHA (c)) ||
           (class_bits & BIT_SPACE && ISSPACE (c)) ||
+          (class_bits & BIT_BLANK && ISBLANK (c)) ||
           (class_bits & BIT_WORD  && ISWORD  (c)) ||
           ((class_bits & BIT_UPPER) &&
            (ISUPPER (c) || (corig != c &&