From 67e0d9dd98a1fd0b003ae387426697ef4fcae42c Mon Sep 17 00:00:00 2001
From: Ingo Karkat <swdev@ingo-karkat.de>
Date: Fri, 13 Jan 2012 23:27:58 +0100
Subject: [PATCH] Support "preserve line numbers" in deduplicate.

---
 tests/t1910-deduplicate.sh | 53 +++++++++++++++++++++++++++++++++++---
 todo.sh                    | 24 ++++++++++++-----
 2 files changed, 67 insertions(+), 10 deletions(-)

diff --git a/tests/t1910-deduplicate.sh b/tests/t1910-deduplicate.sh
index 16dc999..567c235 100755
--- a/tests/t1910-deduplicate.sh
+++ b/tests/t1910-deduplicate.sh
@@ -16,10 +16,38 @@ double task
 three
 EOF
 
-test_todo_session 'deduplicate with duplicates' <<EOF
+test_todo_session 'deduplicate and preserve line numbers' <<EOF
 >>> todo.sh deduplicate
 TODO: 2 duplicate task(s) removed
 
+>>> todo.sh -p ls
+5 double task
+1 duplicated
+7 three
+2 two
+3 x done
+--
+TODO: 5 of 5 tasks shown
+EOF
+
+test_todo_session 'deduplicate without duplicates' <<EOF
+>>> todo.sh deduplicate
+TODO: No duplicate tasks found
+EOF
+
+cat > todo.txt <<EOF
+duplicated
+two
+x done
+duplicated
+double task
+double task
+three
+EOF
+test_todo_session 'deduplicate and delete lines' <<EOF
+>>> todo.sh -n deduplicate
+TODO: 2 duplicate task(s) removed
+
 >>> todo.sh -p ls
 4 double task
 1 duplicated
@@ -30,9 +58,26 @@ TODO: 2 duplicate task(s) removed
 TODO: 5 of 5 tasks shown
 EOF
 
-test_todo_session 'deduplicate without duplicates' <<EOF
+cat > todo.txt <<EOF
+one
+duplicated
+three
+duplicated
+duplicated
+six
+duplicated
+EOF
+test_todo_session 'deduplicate more than two occurrences' <<EOF
 >>> todo.sh deduplicate
-TODO: No duplicate tasks found
+TODO: 3 duplicate task(s) removed
+
+>>> todo.sh -p ls
+2 duplicated
+1 one
+6 six
+3 three
+--
+TODO: 4 of 4 tasks shown
 EOF
 
 cat > todo.txt <<EOF
@@ -50,7 +95,7 @@ TODO: 1 duplicate task(s) removed
 2 a bold action
 1 normal task
 3 something else
-4 something more
+5 something more
 --
 TODO: 4 of 4 tasks shown
 EOF
diff --git a/todo.sh b/todo.sh
index c9a50ad..28dd2e3 100755
--- a/todo.sh
+++ b/todo.sh
@@ -1208,21 +1208,33 @@ note: PRIORITY must be anywhere from A to Z."
     ;;
 
 "deduplicate" )
-    originalTaskNum=$( sed -n '$ =' "$TODO_FILE" )
+    if [ $TODOTXT_PRESERVE_LINE_NUMBERS = 0 ]; then
+        deduplicateSedCommand='d'
+    else
+        deduplicateSedCommand='{ s/^.*//; p; b }'
+    fi
+
+    # To determine the difference when deduplicated lines are preserved, only
+    # non-empty lines must be counted.
+    originalTaskNum=$( sed -e '/./!d' "$TODO_FILE" | sed -n '$ =' )
 
     # Look for duplicate lines and discard the second occurrence.
     # We start with an empty hold space on the first line.  For each line:
     #   G - appends newline + hold space to the pattern space
     #   s/\n/&&/; - double up the first new line so we catch adjacent dups
-    #   /^\([^\n]*\n\).*\n\1/d;
+    #   /^\([^\n]*\n\).*\n\1/
     #       If the first line of the hold space shows up again later as an
-    #       entire line, it's a duplicate.  Delete the current pattern space,
-    #       quit this line and move on to the next
+    #       entire line, it's a duplicate.
+    #       d;                 - Delete the current pattern space, quit this line
+    #                            and move on to the next, or:
+    #       { s/^.*//; p; b }; - Clear the task text, print this line and move on
+    #                            to the next.
     #   s/\n//;   - else, drop the doubled newline
     #   h;        - replace the hold space with the expanded pattern space
     #   P;        - print up to the first newline (that is, the input line)
-    sed -i.bak -n 'G; s/\n/&&/; /^\([^\n]*\n\).*\n\1/d; s/\n//; h; P' "$TODO_FILE"
-    newTaskNum=$( sed -n '$ =' "$TODO_FILE" )
+    sed -i.bak -n 'G; s/\n/&&/; /^\([^\n]*\n\).*\n\1/'"$deduplicateSedCommand"'; s/\n//; h; P' "$TODO_FILE"
+
+    newTaskNum=$( sed -e '/./!d' "$TODO_FILE" | sed -n '$ =' )
     deduplicateNum=$(( originalTaskNum - newTaskNum ))
     if [ $deduplicateNum -eq 0 ]; then
         echo "TODO: No duplicate tasks found"