Re: cp file with a filter

[Peter Stephenson <p.w.stephenson@ntlworld.com>]

Peter Stephenson wrote:
> If you're dead set on regular expressions you can use the [[ ... =~
> ... ]] syntax, but unfortunately I've just noticed this is a bit broken
> for substitutions since although it sets the variable MATCH it doesn't
> set the variables MBEGIN and MEND, which both is annoyingly inconsistent
> with variable substitution and makes it hard to decide which bit of the
> line you're replacing.  The following fixes that omission.  It's an
> exercise for the reader to use this to replace the part of the history line
> from $MBEGIN to $MEND.
> 
> Comments on the patch should go to zsh-workers.

Talking of annoying inconsistencies, this updates the patch to work when
pcre is used for regex matching and fixes the pcre behaviour (and
documents the behaviour for both cases) so that the array variables
aren't set in =~ matching if there were no parenthesised subexpressions.

Index: Doc/Zsh/cond.yo
===================================================================
RCS file: /cvsroot/zsh/zsh/Doc/Zsh/cond.yo,v
retrieving revision 1.6
diff -u -r1.6 cond.yo
--- Doc/Zsh/cond.yo	15 Jan 2009 09:49:06 -0000	1.6
+++ Doc/Zsh/cond.yo	17 Jan 2010 21:26:03 -0000
@@ -117,13 +117,28 @@
 extended regular expression using the tt(zsh/regex) module.
 Upon successful match, some variables will be updated; no variables
 are changed if the matching fails.
+
+If the option tt(BASH_REMATCH) is not set the scalar parameter
+tt(MATCH) is set to the substring that matched the pattern and
+the integer parameters tt(MBEGIN) and tt(MEND) to the index of the start
+and end, respectively, of the match in var(string), such that if
+var(string) is contained in variable tt(var) the expression
+`${var[$MBEGIN,$MEND]}' is identical to `$MATCH'.  The setting
+of the option tt(KSH_ARRAYS) is respected.  Likewise, the array
+tt(match) is set to the substrings that matched parenthesised
+subexpressions and the arrays tt(mbegin) and tt(mend) to the indices of
+the start and end positions, respectively, of the substrings within
+var(string).  For example, if the string `tt(a short string)' is matched
+against the regular expression `tt(s(...)t)', then (assuming the option
+tt(KSH_ARRAYS) is not set) tt(MATCH), tt(MBEGIN)
+and tt(MEND) are `tt(short)', 3 and 7, respectively, while tt(match),
+tt(mbegin) and tt(mend) are single entry arrays containing
+the strings `tt(hor)', `tt(4)' and `tt(6), respectively.
+
 If the option tt(BASH_REMATCH) is set the array
 tt(BASH_REMATCH) is set to the substring that matched the pattern
 followed by the substrings that matched parenthesised
-subexpressions within the pattern; otherwise, the scalar parameter
-tt(MATCH) is set to the substring that matched the pattern and
-and the array tt(match) to the substrings that matched parenthesised
-subexpressions.
+subexpressions within the pattern.
 )
 item(var(string1) tt(<) var(string2))(
 true if var(string1) comes before var(string2)
Index: Src/Modules/pcre.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/Modules/pcre.c,v
retrieving revision 1.16
diff -u -r1.16 pcre.c
--- Src/Modules/pcre.c	25 Mar 2009 11:29:14 -0000	1.16
+++ Src/Modules/pcre.c	17 Jan 2010 21:26:03 -0000
@@ -138,8 +138,9 @@
 
 /**/
 static int
-zpcre_get_substrings(char *arg, int *ovec, int ret, char *matchvar, char *substravar, 
-    int want_offset_pair, int matchedinarr)
+zpcre_get_substrings(char *arg, int *ovec, int ret, char *matchvar,
+		     char *substravar, int want_offset_pair, int matchedinarr,
+		     int want_begin_end)
 {
     char **captures, *match_all, **matches;
     char offset_all[50];
@@ -154,6 +155,7 @@
     
     /* captures[0] will be entire matched string, [1] first substring */
     if (!pcre_get_substring_list(arg, ovec, ret, (const char ***)&captures)) {
+	int nelem = arrlen(captures)-1;
 	/* Set to the offsets of the complete match */
 	if (want_offset_pair) {
 	    sprintf(offset_all, "%d %d", ovec[0], ovec[1]);
@@ -161,8 +163,70 @@
 	}
 	match_all = ztrdup(captures[0]);
 	setsparam(matchvar, match_all);
-	matches = zarrdup(&captures[capture_start]);
-	setaparam(substravar, matches);
+	/*
+	 * If we're setting match, mbegin, mend we only do
+	 * so if there were parenthesised matches, for consistency
+	 * (c.f. regex.c).
+	 */
+	if (!want_begin_end || nelem) {
+	    matches = zarrdup(&captures[capture_start]);
+	    setaparam(substravar, matches);
+	}
+
+	if (want_begin_end) {
+	    char *ptr = arg;
+	    zlong offs = 0;
+
+	    /* Count the characters before the match */
+	    MB_METACHARINIT();
+	    while (ptr < arg + ovec[0]) {
+		offs++;
+		ptr += MB_METACHARLEN(ptr);
+	    }
+	    setiparam("MBEGIN", offs + !isset(KSHARRAYS));
+	    /* Add on the characters in the match */
+	    while (ptr < arg + ovec[1]) {
+		offs++;
+		ptr += MB_METACHARLEN(ptr);
+	    }
+	    setiparam("MEND", offs + !isset(KSHARRAYS) - 1);
+	    if (nelem) {
+		char **mbegin, **mend, **bptr, **eptr;
+		int i, *ipair;
+
+		bptr = mbegin = zalloc(nelem+1);
+		eptr = mend = zalloc(nelem+1);
+
+		for (ipair = ovec + 2, i = 0;
+		     i < nelem;
+		     ipair += 2, i++, bptr++, eptr++)
+		{
+		    char buf[DIGBUFSIZE];
+		    ptr = arg;
+		    offs = 0;
+		    /* Find the start offset */
+		    MB_METACHARINIT();
+		    while (ptr < arg + ipair[0]) {
+			offs++;
+			ptr += MB_METACHARLEN(ptr);
+		    }
+		    convbase(buf, offs + !isset(KSHARRAYS), 10);
+		    *bptr = ztrdup(buf);
+		    /* Continue to the end offset */
+		    while (ptr < arg + ipair[1]) {
+			offs++;
+			ptr += MB_METACHARLEN(ptr);
+		    }
+		    convbase(buf, offs + !isset(KSHARRAYS) - 1, 10);
+		    *eptr = ztrdup(buf);
+		}
+		*bptr = *eptr = NULL;
+
+		setaparam("mbegin", mbegin);
+		setaparam("mend", mend);
+	    }
+	}
+
 	pcre_free_substring_list((const char **)captures);
     }
 
@@ -238,7 +302,8 @@
     if (ret==0) return_value = 0;
     else if (ret==PCRE_ERROR_NOMATCH) /* no match */;
     else if (ret>0) {
-	zpcre_get_substrings(*args, ovec, ret, matched_portion, receptacle, want_offset_pair, 0);
+	zpcre_get_substrings(*args, ovec, ret, matched_portion, receptacle,
+			     want_offset_pair, 0, 0);
 	return_value = 0;
     }
     else {
@@ -297,7 +362,9 @@
 		    break;
 		}
                 else if (r>0) {
-		    zpcre_get_substrings(lhstr, ov, r, NULL, avar, 0, isset(BASHREMATCH));
+		    zpcre_get_substrings(lhstr, ov, r, NULL, avar, 0,
+					 isset(BASHREMATCH),
+					 !isset(BASHREMATCH));
 		    return_value = 1;
 		    break;
 		}
Index: Src/Modules/regex.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/Modules/regex.c,v
retrieving revision 1.5
diff -u -r1.5 regex.c
--- Src/Modules/regex.c	19 Jan 2009 08:26:21 -0000	1.5
+++ Src/Modules/regex.c	17 Jan 2010 21:26:03 -0000
@@ -108,11 +108,65 @@
 	    if (isset(BASHREMATCH)) {
 		setaparam("BASH_REMATCH", arr);
 	    } else {
+		zlong offs;
+		char *ptr;
+
 		m = matches;
 		s = ztrduppfx(lhstr + m->rm_so, m->rm_eo - m->rm_so);
 		setsparam("MATCH", s);
-		if (nelem)
+		/*
+		 * Count the characters before the match.
+		 */
+		ptr = lhstr;
+		offs = 0;
+		MB_METACHARINIT();
+		while (ptr < lhstr + m->rm_so) {
+		    offs++;
+		    ptr += MB_METACHARLEN(ptr);
+		}
+		setiparam("MBEGIN", offs + !isset(KSHARRAYS));
+		/*
+		 * Add on the characters in the match.
+		 */
+		while (ptr < lhstr + m->rm_eo) {
+		    offs++;
+		    ptr += MB_METACHARLEN(ptr);
+		}
+		setiparam("MEND", offs + !isset(KSHARRAYS) - 1);
+		if (nelem) {
+		    char **mbegin, **mend, **bptr, **eptr;
+		    bptr = mbegin = (char **)zalloc(nelem+1);
+		    eptr = mend = (char **)zalloc(nelem+1);
+
+		    for (m = matches + start, n = start;
+			 n <= (int)re.re_nsub;
+			 ++n, ++m, ++bptr, ++eptr)
+		    {
+			char buf[DIGBUFSIZE];
+			ptr = lhstr;
+			offs = 0;
+			/* Find the start offset */
+			MB_METACHARINIT();
+			while (ptr < lhstr + m->rm_so) {
+			    offs++;
+			    ptr += MB_METACHARLEN(ptr);
+			}
+			convbase(buf, offs + !isset(KSHARRAYS), 10);
+			*bptr = ztrdup(buf);
+			/* Continue to the end offset */
+			while (ptr < lhstr + m->rm_eo) {
+			    offs++;
+			    ptr += MB_METACHARLEN(ptr);
+			}
+			convbase(buf, offs + !isset(KSHARRAYS) - 1, 10);
+			*eptr = ztrdup(buf);
+		    }
+		    *bptr = *eptr = NULL;
+
 		    setaparam("match", arr);
+		    setaparam("mbegin", mbegin);
+		    setaparam("mend", mend);
+		}
 	    }
 	}
 	else
Index: Test/C02cond.ztst
===================================================================
RCS file: /cvsroot/zsh/zsh/Test/C02cond.ztst,v
retrieving revision 1.23
diff -u -r1.23 C02cond.ztst
--- Test/C02cond.ztst	26 Nov 2008 10:50:07 -0000	1.23
+++ Test/C02cond.ztst	17 Jan 2010 21:26:03 -0000
@@ -251,6 +251,39 @@
   fi
 0:regex tests shouldn't crash
 
+  if zmodload -i zsh/regex 2>/dev/null; then
+    string="this has stuff in it"
+    bad_regex=0
+    if [[ $string =~ "h([a-z]*) s([a-z]*) " ]]; then
+      if [[ "$MATCH $MBEGIN $MEND" != "has stuff  6 15" ]]; then
+	print -r "regex variables MATCH MBEGIN MEND:
+  '$MATCH $MBEGIN $MEND'
+  should be:
+  'has stuff  6 15'" >&2
+        bad_regex=1
+      else
+	results=("as 7 8" "tuff 11 14")
+	for i in 1 2; do
+	  if [[ "$match[$i] $mbegin[$i] $mend[$i]" != $results[i] ]]; then
+	    print -r "regex variables match[$i] mbegin[$i] mend[$i]:
+  '$match[$i] $mbegin[$i] $mend[$i]'
+  should be
+  '$results[$i]'" >&2
+	    break
+	  fi
+	done
+      fi
+    else
+      print -r "regex failed to match '$string'" >&2
+    fi
+    (( bad_regex )) || print OK
+  else
+    # if it didn't load, tough, but not a test error
+    print OK
+  fi
+0:MATCH, MBEGIN, MEND, match, mbegin, mend
+>OK
+
 %clean
   # This works around a bug in rm -f in some versions of Cygwin
   chmod 644 unmodish


-- 
Peter Stephenson <p.w.stephenson@ntlworld.com>
Web page now at http://homepage.ntlworld.com/p.w.stephenson/