>From 6b43c1d0ea6faa460f77555eeeb74e0523e2f3a0 Mon Sep 17 00:00:00 2001
From: Roni Kallio <roni@kallio.app>
Date: Thu, 8 Oct 2020 01:33:38 +0300
Subject: [PATCH] uniq: --count-separator option

Add a count-separator option, that allows us to set the separator used
between occurrences and line content, when the -c option is enabled.
---
 src/uniq.c | 32 ++++++++++++++++++++++++++++++--
 1 file changed, 30 insertions(+), 2 deletions(-)

diff --git a/src/uniq.c b/src/uniq.c
index e0247579b..2e9d83e9d 100644
--- a/src/uniq.c
+++ b/src/uniq.c
@@ -79,6 +79,10 @@ static bool output_later_repeated;
 /* If true, ignore case when comparing.  */
 static bool ignore_case;
 
+/* Separator character used between output lines and the number of times they
+   occur in the input. */
+static char count_separator = ' ';
+
 enum delimit_method
 {
   /* No delimiters output.  --all-repeated[=none] */
@@ -136,12 +140,14 @@ static enum grouping_method grouping = GM_NONE;
 
 enum
 {
-  GROUP_OPTION = CHAR_MAX + 1
+  GROUP_OPTION = CHAR_MAX + 1,
+  COUNT_SEPARATOR_OPTION
 };
 
 static struct option const longopts[] =
 {
   {"count", no_argument, NULL, 'c'},
+  {"count-separator", required_argument, NULL, COUNT_SEPARATOR_OPTION},
   {"repeated", no_argument, NULL, 'd'},
   {"all-repeated", optional_argument, NULL, 'D'},
   {"group", optional_argument, NULL, GROUP_OPTION},
@@ -178,6 +184,9 @@ With no options, matching lines are merged to the first occurrence.\n\
 
      fputs (_("\
   -c, --count           prefix lines by the number of occurrences\n\
+      --count-separator=CHAR   separator character used between output lines\n\
+                                 and the number of times they occur in the\n\
+                                 input \n\
   -d, --repeated        only print duplicate lines, one for each group\n\
 "), stdout);
      fputs (_("\
@@ -308,7 +317,7 @@ writeline (struct linebuffer const *line,
     return;
 
   if (countmode == count_occurrences)
-    printf ("%7" PRIuMAX " ", linecount + 1);
+    printf ("%7" PRIuMAX "%c", linecount + 1, count_separator);
 
   fwrite (line->buffer, sizeof (char), line->length, stdout);
 }
@@ -483,6 +492,7 @@ main (int argc, char **argv)
   char const *file[2];
   char delimiter = '\n';	/* change with --zero-terminated, -z */
   bool output_option_used = false;   /* if true, one of -u/-d/-D/-c was used */
+  bool count_separator_used = false; /* true if --count-separator was used */
 
   file[0] = file[1] = "-";
   initialize_main (&argc, &argv);
@@ -594,6 +604,18 @@ main (int argc, char **argv)
                                   grouping_method_map);
           break;
 
+        case COUNT_SEPARATOR_OPTION:
+          /* Separator for count and line. Interpret --count-separator='' to
+          mean 'use NUL byte as * delimiter.' */
+          if (optarg[0] != '\0' && optarg[1] != '\0')
+            {
+              error (0, 0, _("The separator must be a single character"));
+              usage (EXIT_FAILURE);
+            }
+          count_separator = optarg[0];
+          count_separator_used = true;
+          break;
+
         case 'f':
           skip_field_option_type = SFO_NEW;
           skip_fields = size_opt (optarg,
@@ -656,6 +678,12 @@ main (int argc, char **argv)
       usage (EXIT_FAILURE);
     }
 
+  if (count_separator_used && countmode == count_none)
+    {
+      error (0, 0, _("--count-separator requires -c"));
+      usage (EXIT_FAILURE);
+    }
+
   check_file (file[0], file[1], delimiter);
 
   return EXIT_SUCCESS;
-- 
2.26.2