From: Ken Hornstein <kenh@pobox.com>
Date: Thu, 14 Nov 2013 16:43:33 +0000 (-0500)
Subject: Break out the unquote functionality to a separate function.
X-Git-Url: https://diplodocus.org/git/nmh/commitdiff_plain/7e8ff2e745cc775339c53d29658f358786f4a500?ds=sidebyside;hp=-c

Break out the unquote functionality to a separate function.
---

7e8ff2e745cc775339c53d29658f358786f4a500
diff --git a/Makefile.am b/Makefile.am
index 1f87ac26..439ce152 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -547,7 +547,8 @@ sbr_libmh_a_SOURCES = sbr/addrsbr.c sbr/ambigsw.c sbr/atooi.c sbr/arglist.c \
 		      sbr/seq_setcur.c sbr/seq_setprev.c sbr/seq_setunseen.c \
 		      sbr/showfile.c sbr/signals.c sbr/smatch.c \
 		      sbr/snprintb.c sbr/ssequal.c \
-		      sbr/strindex.c sbr/trimcpy.c sbr/uprf.c sbr/vfgets.c \
+		      sbr/strindex.c sbr/trimcpy.c sbr/unquote.c \
+		      sbr/uprf.c sbr/vfgets.c \
 		      sbr/fmt_def.c sbr/mf.c sbr/utils.c sbr/ctype-checked.c \
 		      sbr/m_mktemp.c sbr/getansreadline.c sbr/vector.c \
 		      config/config.c config/version.c
diff --git a/h/prototypes.h b/h/prototypes.h
index 57e51d7b..6db99958 100644
--- a/h/prototypes.h
+++ b/h/prototypes.h
@@ -212,6 +212,23 @@ int ssequal (char *, char *);
 int stringdex (char *, char *);
 char *trimcpy (char *);
 int unputenv (char *);
+
+/*
+ * Remove quotes and quoted-pair sequences from RFC-5322 atoms.
+ *
+ * Currently the actual algorithm is simpler than it technically should
+ * be: any quotes are simply eaten, unless they're preceded by the escape
+ * character (\).  This seems to be sufficient for our needs for now.
+ *
+ * Arguments:
+ *
+ * input	- The input string
+ * output	- The output string; is assumed to have at least as much
+ *		  room as the input string.  At worst the output string will
+ *		  be the same size as the input string; it might be smaller.
+ *
+ */
+void unquote_string(const char *input, char *output);
 int uprf (char *, char *);
 int vfgets (FILE *, char **);
 char *write_charset_8bit (void);
diff --git a/sbr/fmt_scan.c b/sbr/fmt_scan.c
index d0bb16df..408b1093 100644
--- a/sbr/fmt_scan.c
+++ b/sbr/fmt_scan.c
@@ -868,32 +868,10 @@ fmt_scan (struct format *format, char *scanl, size_t max, int width, int *dat,
 		/* UNQUOTEs RFC-2822 quoted-string and quoted-pair */
 	case FT_LS_UNQUOTE:
 	    if (str) {	  	
-		int m;
 		strncpy(buffer, str, sizeof(buffer));
 		/* strncpy doesn't NUL-terminate if it fills the buffer */
 		buffer[sizeof(buffer)-1] = '\0';
-		str = buffer;
-	
-		/* we will parse from buffer to buffer2 */
-		n = 0; /* n is the input position in str */
-		m = 0; /* m is the ouput position in buffer2 */
-
-		while ( str[n] != '\0') {
-		    switch ( str[n] ) {
-			case '\\':
-			    n++;
-			    if ( str[n] != '\0')
-				buffer2[m++] = str[n++];
-			    break;
-			case '"':
-			    n++;
-			    break;
-			default:
-			    buffer2[m++] = str[n++];
-			    break;
-			}
-		}
-		buffer2[m] = '\0';
+		unquote_string(buffer, buffer2);
 		str = buffer2;
             }
 	    break;
diff --git a/sbr/unquote.c b/sbr/unquote.c
new file mode 100644
index 00000000..f9fa9112
--- /dev/null
+++ b/sbr/unquote.c
@@ -0,0 +1,45 @@
+/*
+ * unquote.c: Handle quote removal and quoted-pair strings on
+ * RFC 2822-5322 atoms.
+ *
+ * This code is Copyright (c) 2013, by the authors of nmh.  See the
+ * COPYRIGHT file in the root directory of the nmh distribution for
+ * complete copyright information.
+ */
+
+#include <h/mh.h>
+
+/*
+ * Remove quotes (and handle escape strings) from RFC 5322 quoted-strings.
+ *
+ * Since we never add characters to the string, the output buffer is assumed
+ * to have at least as many characters as the input string.
+ *
+ */
+
+void
+unquote_string(const char *input, char *output)
+{
+    int n = 0;	/* n is the position in the input buffer */
+    int m = 0;	/* m is the position in the output buffer */
+
+    while ( input[n] != '\0') {
+	switch ( input[n] ) {
+	case '\\':
+	    n++;
+	    if ( input[n] != '\0')
+		output[m++] = input[n++];
+	    break;
+	case '"':
+	    n++;
+	    break;
+	default:
+	    output[m++] = input[n++];
+	    break;
+	}
+    }
+
+    output[m] = '\0';
+
+    return;
+}