regexp, regsub: add support for -expanded

Fixes #311

Signed-off-by: Steve Bennett <steveb@workware.net.au>

jim-regexp.c

@@ -137,10 +137,10 @@ int Jim_RegexpCmd(Jim_Interp *interp, int argc, Jim_Obj *const *argv)
     int eflags = 0;
     int option;
     enum {
-        OPT_INDICES,  OPT_NOCASE, OPT_LINE, OPT_ALL, OPT_INLINE, OPT_START, OPT_END
+        OPT_INDICES,  OPT_NOCASE, OPT_LINE, OPT_ALL, OPT_INLINE, OPT_START, OPT_EXPANDED, OPT_END
     };
     static const char * const options[] = {
-        "-indices", "-nocase", "-line", "-all", "-inline", "-start", "--", NULL
+        "-indices", "-nocase", "-line", "-all", "-inline", "-start", "-expanded", "--", NULL
     };
 
     for (i = 1; i < argc; i++) {
@@ -185,6 +185,15 @@ int Jim_RegexpCmd(Jim_Interp *interp, int argc, Jim_Obj *const *argv)
                     return JIM_ERR;
                 }
                 break;
+
+            case OPT_EXPANDED:
+#ifdef REG_EXPANDED
+                regcomp_flags |= REG_EXPANDED;
+                break;
+#else
+                Jim_SetResultFormatted(interp, "not supported: %#s", argv[i]);
+                return JIM_ERR;
+#endif
         }
     }
     if (argc - i < 2) {
@@ -361,10 +370,10 @@ int Jim_RegsubCmd(Jim_Interp *interp, int argc, Jim_Obj *const *argv)
     const char *pattern;
     int option;
     enum {
-        OPT_NOCASE, OPT_LINE, OPT_ALL, OPT_START, OPT_COMMAND, OPT_END
+        OPT_NOCASE, OPT_LINE, OPT_ALL, OPT_START, OPT_COMMAND, OPT_EXPANDED, OPT_END
     };
     static const char * const options[] = {
-        "-nocase", "-line", "-all", "-start", "-command", "--", NULL
+        "-nocase", "-line", "-all", "-start", "-command", "-expanded", "--", NULL
     };
 
     for (i = 1; i < argc; i++) {
@@ -405,6 +414,15 @@ int Jim_RegsubCmd(Jim_Interp *interp, int argc, Jim_Obj *const *argv)
             case OPT_COMMAND:
                 opt_command = 1;
                 break;
+
+            case OPT_EXPANDED:
+#ifdef REG_EXPANDED
+                regcomp_flags |= REG_EXPANDED;
+                break;
+#else
+                Jim_SetResultFormatted(interp, "not supported: %#s", argv[i]);
+                return JIM_ERR;
+#endif
         }
     }
     if (argc - i != 3 && argc - i != 4) {

jim_tcl.txt

@@ -57,6 +57,7 @@ Changes since 0.83
 1. `aio` - support for configurable read and write buffering
 2. `exec` TIP 424 - support safer +exec | + syntax (also +open "|| pipeline..."+) (see https://core.tcl-lang.org/tips/doc/trunk/tip/424.md)
 3. New `lsubst` command to create lists using subst-style substitution
+4. Add support for `regexp -expanded` and `regsub -expanded`
 
 Changes between 0.82 and 0.83
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -3876,7 +3877,7 @@ See `aio read`
 
 regexp
 ~~~~~~
-+*regexp ?-nocase? ?-line? ?-indices? ?-start* 'offset'? *?-all? ?-inline? ?--?* 'exp string ?matchVar? ?subMatchVar subMatchVar \...?'+
++*regexp ?-nocase? ?-line? ?-indices? ?-start* 'offset'? *?-all? ?-inline? ?-expanded? ?--?* 'exp string ?matchVar? ?subMatchVar subMatchVar \...?'+
 
 Determines whether the regular expression +'exp'+ matches part or
 all of +'string'+ and returns 1 if it does, 0 if it doesn't.
@@ -3949,13 +3950,17 @@ The following switches modify the behaviour of +'regexp'+
     data, plus one element for each subexpression in the regular
     expression.
 
++*-expanded*+::
+    Enables use of the expanded regular expression syntax where whitespace
+    and comments are ignored.
+
 +*--*+::
     Marks the end of switches. The argument following this one will be
     treated as +'exp'+ even if it starts with a +-+.
 
 regsub
 ~~~~~~
-+*regsub ?-nocase? ?-all? ?-line? ?-command? ?-start* 'offset'? ?*--*? 'exp string subSpec ?varName?'+
++*regsub ?-nocase? ?-all? ?-line? ?-command? ?-expanded? ?-start* 'offset'? ?*--*? 'exp string subSpec ?varName?'+
 
 This command matches the regular expression +'exp'+ against
 +'string'+ using the rules described in REGULAR EXPRESSIONS
@@ -4037,6 +4042,10 @@ The following switches modify the behaviour of +'regsub'+
     start matching the regular expression. +'offset'+ will be
     constrained to the bounds of the input string.
 
++*-expanded*+::
+    Enables use of the expanded regular expression syntax where whitespace
+    and comments are ignored.
+
 +*--*+::
     Marks the end of switches. The argument following this one will be
     treated as +'exp'+ even if it starts with a +-+.

jimregexp.c

@@ -948,6 +948,27 @@ static int regatom(regex_t *preg, int *flagp)
 
 			ret = regnode(preg, EXACTLY);
 
+			if (preg->cflags & REG_EXPANDED) {
+				/* Skip leading white space */
+				while ((ch = *preg->regparse) != 0) {
+					if (strchr(" \t\r\n\f\v", ch)) {
+							preg->regparse++;
+							continue;
+					}
+					break;
+				}
+				if (ch == '#') {
+					/* And skip comments to end of line */
+					preg->regparse++;
+					while ((ch = *preg->regparse) != 0) {
+						preg->regparse++;
+						if (ch == '\n') {
+							break;
+						}
+					}
+				}
+			}
+
 			/* Note that a META operator such as ? or * consumes the
 			 * preceding char.
 			 * Thus we must be careful to look ahead by 2 and add the
@@ -993,6 +1014,12 @@ static int regatom(regex_t *preg, int *flagp)
 					break;
 				}
 
+				/* For REG_EXPANDED, if we hit white space, stop */
+				if ((preg->cflags & REG_EXPANDED) && n == 1 && strchr(" \t\r\n\f\v", ch)) {
+					preg->regparse += n;
+					break;
+				}
+
 				/* No, so just add this char normally */
 				regc(preg, ch);
 				added++;

jimregexp.h

@@ -73,6 +73,7 @@ typedef regexp regex_t;
 #define REG_ICASE 2
 
 #define REG_NOTBOL 16
+#define REG_EXPANDED 32
 
 enum {
 	REG_NOERROR,      /* Success.  */

tests/regexp.test

@@ -199,9 +199,9 @@ test regexp-6.1 {regexp errors} {
 test regexp-6.2 {regexp errors} {
     list [catch {regexp -nocase a} msg] $msg
 } {1 {wrong # args: should be "regexp ?-switch ...? exp string ?matchVar? ?subMatchVar ...?"}}
-test regexp-6.3 {regexp errors} jim {
+test regexp-6.3 {regexp errors} -constraints jim -body {
     list [catch {regexp -gorp a} msg] $msg
-} {1 {bad switch "-gorp": must be --, -all, -indices, -inline, -line, -nocase, or -start}}
+} -result {1 {bad switch "-gorp": must be --, -all, -expanded, -indices, -inline, -line, -nocase, or -start}}
 test regexp-6.4 {regexp errors} {
     catch {regexp a( b} msg
 } 1
@@ -367,7 +367,7 @@ test regexp-11.4 {regsub errors} {
 } {1 {wrong # args: should be "regsub ?-switch ...? exp string subSpec ?varName?"}}
 test regexp-11.5 {regsub errors} -constraints jim -body {
     list [catch {regsub -gorp a b c} msg] $msg
-} -result {1 {bad switch "-gorp": must be --, -all, -command, -line, -nocase, or -start}}
+} -result {1 {bad switch "-gorp": must be --, -all, -command, -expanded, -line, -nocase, or -start}}
 test regexp-11.6 {regsub errors} {
     catch {regsub -nocase a( b c d} msg
 } 1

tests/regexp2.test

@@ -463,12 +463,12 @@ test regexpComp-9.6 {-all option to regsub} {
     }
 } {1 123xxx}
 
-#test regexpComp-10.1 {expanded syntax in regsub} {
-#    evalInProc {
-#	set foo xxx
-#	list [regsub -expanded ". \#comment\n  . \#comment2" abc def foo] $foo
-#    }
-#} {1 defc}
+test regexpComp-10.1 {expanded syntax in regsub} {
+    evalInProc {
+	set foo xxx
+	list [regsub -expanded ". \#comment\n  . \#comment2" abc def foo] $foo
+    }
+} {1 defc}
 test regexpComp-10.2 {newline sensitivity in regsub} {
     evalInProc {
 	set foo xxx
@@ -523,11 +523,11 @@ test regexpComp-11.4 {regsub errors} {
 	list [catch {regsub a b c d e f} msg] $msg
     }
 } {1 {wrong # args: should be "regsub ?-switch ...? exp string subSpec ?varName?"}}
-#test regexpComp-11.5 {regsub errors} {
-#    evalInProc {
-#	list [catch {regsub -gorp a b c} msg] $msg
-#    }
-#} {1 {bad switch "-gorp": must be -all, -nocase, -expanded, -line, -linestop, -lineanchor, -start, or --}}
+test regexpComp-11.5 {regsub errors} {
+    evalInProc {
+	list [catch {regsub -gorp a b c} msg] $msg
+    }
+} {1 {bad switch "-gorp": must be --, -all, -command, -expanded, -line, -nocase, or -start}}
 test regexpComp-11.6 {regsub errors} {
     evalInProc {
 	list [catch {regsub -nocase a( b c d} msg] $msg