From 37e67777884acea177d793619440f803d2b9979d Mon Sep 17 00:00:00 2001
From: "Eric S. Raymond" <esr@thyrsus.com>
Date: Wed, 10 Sep 2003 09:31:46 +0000
Subject: [PATCH] Added comment stripping.

---
 Makefile       |  6 +++---
 comparator.xml |  6 +++++-
 main.c         | 19 ++++++++++++++-----
 shred.h        |  1 +
 shredtree.c    |  8 ++++++++
 5 files changed, 31 insertions(+), 9 deletions(-)

diff --git a/Makefile b/Makefile
index 36d9ce3..a69cd9c 100644
--- a/Makefile
+++ b/Makefile
@@ -34,11 +34,11 @@ scf-standard.html: scf-standard.xml
 	xmlto html-nochunks scf-standard.xml
 
 test1: comparator
-	@comparator -w -C -d test test1-a test1-b
+	@comparator -C -d test test1-a test1-b
 test2: comparator
-	@comparator -w -C -d test test2-a test2-b
+	@comparator -C -d test test2-a test2-b
 test3: comparator
-	@comparator -w -C -d test test3-a test3-b
+	@comparator -C -d test test3-a test3-b
 
 install: comparator.1 uninstall
 	install -m 755 -o 0 -g 0 -d $(ROOT)/usr/bin/
diff --git a/comparator.xml b/comparator.xml
index f83f789..82fe0f9 100644
--- a/comparator.xml
+++ b/comparator.xml
@@ -17,11 +17,12 @@
 <cmdsynopsis>
   <command>comparator</command>
   <arg choice='opt'>-c</arg>
+  <arg choice='opt'>-C</arg>
   <arg choice='opt'>-d <replaceable>dir</replaceable></arg>
   <arg choice='opt'>-o <replaceable>file</replaceable></arg>
+  <arg choice='opt'>-r</arg>
   <arg choice='opt'>-s <replaceable>shredsize</replaceable></arg>
   <arg choice='opt'>-w</arg>
-  <arg choice='opt'>-C</arg>
   <arg choice='opt'>-x</arg>
   <arg choice='plain' rep='repeat'>path</arg>
   <sbr/>
@@ -99,6 +100,9 @@ applies C-language-specific normalization; { and } characters are
 removed.  This is recommended for comparing C code; it prevents the
 comparison from being fooled by differences in indent style.</para>
 
+<para>The <option>-r</option> option removes comments.  This is 
+only effective in conjunction with <option>-C</option>.</para>
+
 <para>The <option>-x</option> enables some debugging output.  It
 is for developers; if you need to understand it, read the code.</para>
 
diff --git a/main.c b/main.c
index f5c0f3b..e0dd202 100644
--- a/main.c
+++ b/main.c
@@ -86,6 +86,8 @@ static void write_scf(const char *tree, FILE *ofp)
     buf[0] = '\0';
     if (remove_whitespace)
 	strcat(buf, "remove-whitespace,");
+    if (remove_comments)
+	strcat(buf, "remove-comments,");
     if (remove_braces)
 	strcat(buf, "remove-braces,");
     if (buf[0])
@@ -331,10 +333,12 @@ void report_time(char *legend, ...)
 
 static void usage(void)
 {
-    fprintf(stderr,"usage: shredtree [-c] [-d dir ] [-o file] [-s shredsize] [-w] [-x] path...\n");
+    fprintf(stderr,"usage: comparator [-c] [-C] [-d dir ] [-r] [-o file] [-s shredsize] [-w] [-x] path...\n");
     fprintf(stderr,"  -c      = generate SCF files\n");
+    fprintf(stderr,"  -C      = apply C normalizations\n");
     fprintf(stderr,"  -d dir  = change directory before digesting.\n");
     fprintf(stderr,"  -o file = write to the specified file.\n");
+    fprintf(stderr,"  -r      = renove comments\n");
     fprintf(stderr,"  -s size = set shred size (default %d)\n", shredsize);
     fprintf(stderr,"  -w      = remove whitespace.\n");
     fprintf(stderr,"  -x      = debug, display chunks in output.\n");
@@ -352,7 +356,7 @@ main(int argc, char *argv[])
 
     compile_only = file_only = 0;
     dir = outfile = NULL;
-    while ((status = getopt(argc, argv, "cCd:ho:s:wx")) != EOF)
+    while ((status = getopt(argc, argv, "cCd:ho:rs:wx")) != EOF)
     {
 	switch (status)
 	{
@@ -361,7 +365,7 @@ main(int argc, char *argv[])
 	    break;
 
 	case 'C':
-	    remove_braces = 1;
+	    remove_whitespace = remove_braces = 1;
 	    break;
 
 	case 'd':
@@ -372,6 +376,9 @@ main(int argc, char *argv[])
 	    outfile = optarg;
 	    break;
 
+	case 'r':
+	    remove_comments = 1;
+
 	case 's':
 	    shredsize = atoi(optarg);
 	    break;
@@ -480,9 +487,11 @@ main(int argc, char *argv[])
 	dummy_scf.hash_method = "MD5";
 	buf[0] == '\0';
 	if (remove_whitespace)
-	    strcat(buf, "remove-whitespace,");
+	    strcat(buf, "remove-whitespace, ");
+	if (remove_comments)
+	    strcat(buf, "remove-comments, ");
 	if (remove_braces)
-	    strcat(buf, "remove-braces,");
+	    strcat(buf, "remove-braces, ");
 	if (buf[0])
 	    buf[strlen(buf)-1] = '\0';
 	else
diff --git a/shred.h b/shred.h
index 548d5ca..dd52ba7 100644
--- a/shred.h
+++ b/shred.h
@@ -33,6 +33,7 @@ struct sorthash_t
 /* control bits, meant to be set at startup */
 extern int remove_braces;
 extern int remove_whitespace;
+extern int remove_comments;
 extern int debug;
 extern int shredsize;
 
diff --git a/shredtree.c b/shredtree.c
index a2fd18b..f8ca2ee 100644
--- a/shredtree.c
+++ b/shredtree.c
@@ -13,6 +13,7 @@
 /* control bits, meant to be set at startup */
 int remove_braces = 0;
 int remove_whitespace = 0;
+int remove_comments = 0;
 int shredsize = 3;
 
 struct item
@@ -68,6 +69,13 @@ static int eligible(const char *file)
 static int normalize(char *buf)
 /* normalize a buffer in place, return 0 if it should be skipped */
 {
+    if (remove_comments && remove_braces)	/* remove_c_comments */
+    {
+	char *start = strstr(buf, "/*"), *end = strstr(buf, "*/");
+
+	if (start && end && start < end)
+	    strcpy(start, end+2);
+    }
     if (remove_whitespace)	/* strip whitespace, ignore blank lines */
     {
 	char *tp, *sp;
-- 
2.43.0