From c255c34c53dfe5dddd7d12f9694c24eaf8e2b2df Mon Sep 17 00:00:00 2001
From: Francois Fleuret <francois@fleuret.org>
Date: Sun, 11 Dec 2011 06:21:40 +0100
Subject: [PATCH] Added the --trim-first option.

---
 finddup.1 |  3 +++
 finddup.c | 40 +++++++++++++++++++++++++++++++---------
 2 files changed, 34 insertions(+), 9 deletions(-)

diff --git a/finddup.1 b/finddup.1
index 7eb9b76..b279216 100644
--- a/finddup.1
+++ b/finddup.1
@@ -60,6 +60,9 @@ do not show the file group IDs
 \fB-t\fR, \fB--time-sort\fR
 sort files in each group according to the modification times
 .TP
+\fB-q\fR, \fB--trim-first\fR
+do not print the first file in each group
+.TP
 \fB-p\fR, \fB--show-progress\fR
 show progress information in stderr
 .TP
diff --git a/finddup.c b/finddup.c
index 45ba3a6..13d0f66 100644
--- a/finddup.c
+++ b/finddup.c
@@ -75,6 +75,8 @@ int same_inodes_are_different = 0; /* 1 means that comparison between
 int sort_by_time = 0; /* 1 means to sort files in each group according
                          to the modification time */
 
+int trim_first = 0; /* remove the first entry in each group */
+
 char *command_to_exec = 0; /* the name of the command to exec for each
                               group of identical files */
 
@@ -273,7 +275,7 @@ struct file_node *scan_directory(struct file_node *tail, const char *name) {
   return result;
 }
 
-void print_file(FILE *out, struct file_node *node) {
+void write_one_entry_to_file(FILE *out, struct file_node *node) {
   char tmp[PATH_MAX + 1];
   if(show_realpaths) {
     if(realpath(node->name, tmp)) {
@@ -350,8 +352,11 @@ void exec_command(int nb, struct file_node **nodes) {
   n = 0;
   while(n < nb) {
     m = n;
+    if(trim_first) { m++; }
     while(n < nb && nodes[n]->group_id == nodes[m]->group_id) {
-      args[n - m + 1] = nodes[n]->name;
+      if(n >= m) {
+        args[n - m + 1] = nodes[n]->name;
+      }
       n++;
     }
     args[n - m + 1] = 0;
@@ -375,18 +380,23 @@ void exec_command(int nb, struct file_node **nodes) {
 
 void write_groups_in_files(int nb, struct file_node **nodes) {
   FILE *file = 0;
-  int current_group = -1, n;
+  int current_group = -1, n, first_of_group;
   char filename[PATH_MAX + 1];
 
   for(n = 0; n < nb; n++) {
+    first_of_group = 0;
     if(nodes[n]->group_id != current_group) {
       if(file) { fclose(file); }
       sprintf(filename, "%s%06d", result_file_prefix, nodes[n]->group_id);
       file = fopen(filename, "w");
       current_group = nodes[n]->group_id;
       printf("Writing %s.\n" , filename);
+      first_of_group = 1;
+    }
+
+    if(!trim_first || !first_of_group) {
+      write_one_entry_to_file(file, nodes[n]);
     }
-    print_file(file, nodes[n]);
   }
 
   if(file) { fclose(file); }
@@ -395,7 +405,7 @@ void write_groups_in_files(int nb, struct file_node **nodes) {
 void print_result(struct file_node *list1, struct file_node *list2) {
   struct file_node *node1, *node2;
   struct file_node **nodes;
-  int nb, n;
+  int nb, n, first_of_group;
 
   nb = 0;
   for(node1 = list1; node1; node1 = node1->next) {
@@ -433,10 +443,16 @@ void print_result(struct file_node *list1, struct file_node *list2) {
     write_groups_in_files(nb, nodes);
   } else {
     for(n = 0; n < nb; n++) {
-      if(!show_groups && n > 0 && nodes[n]->group_id != nodes[n-1]->group_id) {
-        printf("\n");
+      first_of_group = 0;
+      if(n > 0 && nodes[n]->group_id != nodes[n-1]->group_id) {
+        if(!show_groups) {
+          printf("\n");
+        }
+        first_of_group = 1;
+      }
+      if(!trim_first || !first_of_group) {
+        write_one_entry_to_file(stdout, nodes[n]);
       }
-      print_file(stdout, nodes[n]);
     }
   }
 
@@ -614,6 +630,7 @@ void usage(FILE *out) {
   fprintf(out, "                              those in DIR1\n");
   fprintf(out, "   -g, --no-group-ids         do not show the file groups\n");
   fprintf(out, "   -t, --time-sort            sort according to modification time in each group\n");
+  fprintf(out, "   -q, --trim-first           do not show the first file in each group\n");
   fprintf(out, "   -p, --show-progress        show progress\n");
   fprintf(out, "   -r, --real-paths           show the real file paths\n");
   fprintf(out, "   -i, --same-inodes-are-different\n");
@@ -640,6 +657,7 @@ static struct option long_options[] = {
   { "hide-matchings", no_argument, 0, 'c' },
   { "no-group-ids", no_argument, 0, 'g' },
   { "time-sort", no_argument, 0, 't' },
+  { "trim-first", no_argument, 0, 'q' },
   { "ignore-dots", no_argument, 0, 'd' },
   { "ignore-empty", no_argument, 0, '0' },
   { "show-progress", no_argument, 0, 'p' },
@@ -653,7 +671,7 @@ int main(int argc, char **argv) {
 
   setlocale (LC_ALL, "");
 
-  while ((c = getopt_long(argc, argv, "vhircgtd0pme:f:",
+  while ((c = getopt_long(argc, argv, "vhircgtqd0pme:f:",
                           long_options, NULL)) != -1) {
     switch (c) {
 
@@ -692,6 +710,10 @@ int main(int argc, char **argv) {
       sort_by_time = 1;
       break;
 
+    case 'q':
+      trim_first = 1;
+      break;
+
     case 'p':
       show_progress = 1;
       break;
-- 
2.20.1