[PATCH v2 09/10] cli: address: Add --output=count

[Michal Sojka <sojkam1@fel.cvut.cz>]

This output prints how many times was each address encountered during
search.
---
 completion/notmuch-completion.bash |  2 +-
 completion/notmuch-completion.zsh  |  2 +-
 doc/man1/notmuch-address.rst       |  7 ++++++
 notmuch-search.c                   | 49 ++++++++++++++++++++++++++++++++------
 test/T095-address.sh               | 48 +++++++++++++++++++++++++++++++++++++
 5 files changed, 99 insertions(+), 9 deletions(-)

diff --git a/completion/notmuch-completion.bash b/completion/notmuch-completion.bash
index 94ea2d5..db152f3 100644
--- a/completion/notmuch-completion.bash
+++ b/completion/notmuch-completion.bash
@@ -332,7 +332,7 @@ _notmuch_address()
 	    return
 	    ;;
 	--output)
-	    COMPREPLY=( $( compgen -W "sender recipients" -- "${cur}" ) )
+	    COMPREPLY=( $( compgen -W "sender recipients count" -- "${cur}" ) )
 	    return
 	    ;;
 	--sort)
diff --git a/completion/notmuch-completion.zsh b/completion/notmuch-completion.zsh
index c606b75..8968562 100644
--- a/completion/notmuch-completion.zsh
+++ b/completion/notmuch-completion.zsh
@@ -61,7 +61,7 @@ _notmuch_address()
 {
   _arguments -s : \
     '--sort=[sort results]:sorting:((newest-first\:"reverse chronological order" oldest-first\:"chronological order"))' \
-    '--output=[select what to output]:output:((sender recipients))'
+    '--output=[select what to output]:output:((sender recipients count))'
 }
 
 _notmuch()
diff --git a/doc/man1/notmuch-address.rst b/doc/man1/notmuch-address.rst
index 96512b7..18473a7 100644
--- a/doc/man1/notmuch-address.rst
+++ b/doc/man1/notmuch-address.rst
@@ -48,6 +48,13 @@ Supported options for **address** include
             Output all addresses from the *To*, *Cc* and *Bcc*
             headers.
 
+	**count**
+	    Print the count of how many times was the address
+	    encountered during search.
+
+	    Note: With this option, addresses are printed only after
+	    the whole search is finished. This may take long time.
+
     ``--sort=``\ (**newest-first**\ \|\ **oldest-first**)
         This option can be used to present results in either
         chronological order (**oldest-first**) or reverse chronological
diff --git a/notmuch-search.c b/notmuch-search.c
index 741702a..d99e530 100644
--- a/notmuch-search.c
+++ b/notmuch-search.c
@@ -33,6 +33,7 @@ typedef enum {
     /* Address command */
     OUTPUT_SENDER	= 1 << 5,
     OUTPUT_RECIPIENTS	= 1 << 6,
+    OUTPUT_COUNT	= 1 << 7,
 } output_t;
 
 typedef enum {
@@ -59,6 +60,7 @@ typedef struct {
 typedef struct {
     const char *name;
     const char *addr;
+    int count;
 } mailbox_t;
 
 /* Return two stable query strings that identify exactly the matched
@@ -247,17 +249,24 @@ is_duplicate (const search_context_t *ctx, const char *name, const char *addr)
 {
     notmuch_bool_t duplicate;
     char *key;
+    mailbox_t *mailbox;
 
     key = talloc_asprintf (ctx->format, "%s <%s>", name, addr);
     if (! key)
 	return FALSE;
 
-    duplicate = g_hash_table_lookup_extended (ctx->addresses, key, NULL, NULL);
+    duplicate = g_hash_table_lookup_extended (ctx->addresses, key, NULL, (gpointer)&mailbox);
 
-    if (! duplicate)
-	g_hash_table_insert (ctx->addresses, key, NULL);
-    else
+    if (! duplicate) {
+	mailbox = talloc (ctx->format, mailbox_t);
+	mailbox->name = talloc_strdup (mailbox, name);
+	mailbox->addr = talloc_strdup (mailbox, addr);
+	mailbox->count = 1;
+	g_hash_table_insert (ctx->addresses, key, mailbox);
+    } else {
+	mailbox->count++;
 	talloc_free (key);
+    }
 
     return duplicate;
 }
@@ -267,6 +276,7 @@ print_mailbox (const search_context_t *ctx, const mailbox_t *mailbox)
 {
     const char *name = mailbox->name;
     const char *addr = mailbox->addr;
+    int count = mailbox->count;
     sprinter_t *format = ctx->format;
     InternetAddress *ia = internet_address_mailbox_new (name, addr);
     char *name_addr;
@@ -276,6 +286,10 @@ print_mailbox (const search_context_t *ctx, const mailbox_t *mailbox)
     name_addr = internet_address_to_string (ia, FALSE);
 
     if (format->is_text_printer) {
+	if (count > 0) {
+	    format->integer (format, count);
+	    format->string (format, "\t");
+	}
 	format->string (format, name_addr);
 	format->separator (format);
     } else {
@@ -286,6 +300,10 @@ print_mailbox (const search_context_t *ctx, const mailbox_t *mailbox)
 	format->string (format, addr);
 	format->map_key (format, "name-addr");
 	format->string (format, name_addr);
+	if (count > 0) {
+	    format->map_key (format, "count");
+	    format->integer (format, count);
+	}
 	format->end (format);
 	format->separator (format);
     }
@@ -294,7 +312,7 @@ print_mailbox (const search_context_t *ctx, const mailbox_t *mailbox)
     g_free (name_addr);
 }
 
-/* Print addresses from InternetAddressList.  */
+/* Print or prepare for printing addresses from InternetAddressList. */
 static void
 process_address_list (const search_context_t *ctx,
 		      InternetAddressList *list)
@@ -319,17 +337,21 @@ process_address_list (const search_context_t *ctx,
 	    mailbox_t mbx = {
 		.name = internet_address_get_name (address),
 		.addr = internet_address_mailbox_get_addr (mailbox),
+		.count = 0,
 	    };
 
 	    if (is_duplicate (ctx, mbx.name, mbx.addr))
 		continue;
 
+	    if (ctx->output & OUTPUT_COUNT)
+		continue;
+
 	    print_mailbox (ctx, &mbx);
 	}
     }
 }
 
-/* Print addresses from a message header.  */
+/* Print or prepare for printing addresses from a message header. */
 static void
 process_address_header (const search_context_t *ctx, const char *value)
 {
@@ -353,6 +375,15 @@ _my_talloc_free_for_g_hash (void *ptr)
     talloc_free (ptr);
 }
 
+static void
+print_hash_value (unused (gpointer key), gpointer value, gpointer user_data)
+{
+    const mailbox_t *mailbox = value;
+    search_context_t *ctx = user_data;
+
+    print_mailbox (ctx, mailbox);
+}
+
 static int
 _count_filenames (notmuch_message_t *message)
 {
@@ -448,6 +479,9 @@ do_search_messages (search_context_t *ctx)
 	notmuch_message_destroy (message);
     }
 
+    if (ctx->addresses && ctx->output & OUTPUT_COUNT)
+	g_hash_table_foreach (ctx->addresses, print_hash_value, ctx);
+
     notmuch_messages_destroy (messages);
 
     format->end (format);
@@ -685,6 +719,7 @@ notmuch_address_command (notmuch_config_t *config, int argc, char *argv[])
 	{ NOTMUCH_OPT_KEYWORD_FLAGS, &ctx->output, "output", 'o',
 	  (notmuch_keyword_t []){ { "sender", OUTPUT_SENDER },
 				  { "recipients", OUTPUT_RECIPIENTS },
+				  { "count", OUTPUT_COUNT },
 				  { 0, 0 } } },
 	{ NOTMUCH_OPT_INHERIT, &common_options, NULL, 0, 0 },
 	{ 0, 0, 0, 0, 0 }
@@ -702,7 +737,7 @@ notmuch_address_command (notmuch_config_t *config, int argc, char *argv[])
 	return EXIT_FAILURE;
 
     ctx->addresses = g_hash_table_new_full (g_str_hash, g_str_equal,
-					    _my_talloc_free_for_g_hash, NULL);
+					    _my_talloc_free_for_g_hash, _my_talloc_free_for_g_hash);
 
     ret = do_search_messages (ctx);
 
diff --git a/test/T095-address.sh b/test/T095-address.sh
index 8a256d2..92e17b0 100755
--- a/test/T095-address.sh
+++ b/test/T095-address.sh
@@ -96,5 +96,53 @@ notmuch address --output=sender --output=recipients '*' >OUTPUT
 # Use EXPECTED from previous subtest
 test_expect_equal_file OUTPUT EXPECTED
 
+test_begin_subtest "--output=sender --output=count"
+notmuch address --output=sender --output=count '*' | sort -n >OUTPUT
+cat <<EOF >EXPECTED
+1	Adrian Perez de Castro <aperez@igalia.com>
+1	Aron Griffis <agriffis@n01se.net>
+1	Chris Wilson <chris@chris-wilson.co.uk>
+1	François Boulogne <boulogne.f@gmail.com>
+1	Ingmar Vanhassel <ingmar@exherbo.org>
+1	Israel Herraiz <isra@herraiz.org>
+1	Olivier Berger <olivier.berger@it-sudparis.eu>
+1	Rolland Santimano <rollandsantimano@yahoo.com>
+2	Alex Botero-Lowry <alex.boterolowry@gmail.com>
+2	Jjgod Jiang <gzjjgod@gmail.com>
+3	Stewart Smith <stewart@flamingspork.com>
+4	Alexander Botero-Lowry <alex.boterolowry@gmail.com>
+4	Jan Janak <jan@ryngle.com>
+5	Lars Kellogg-Stedman <lars@seas.harvard.edu>
+5	Mikhail Gusarov <dottedmag@dottedmag.net>
+7	Keith Packard <keithp@keithp.com>
+12	Carl Worth <cworth@cworth.org>
+EOF
+test_expect_equal_file OUTPUT EXPECTED
+
+test_begin_subtest "--output=sender --output=count --format=json"
+# Since the iteration order of GHashTable is not specified, we
+# preprocess and sort the results to keep the order stable here.
+notmuch address --output=sender --output=count --format=json '*' | \
+    sed -e 's/^\[//' -e 's/]$//' -e 's/,$//' | sort >OUTPUT
+cat <<EOF >EXPECTED
+{"name": "Adrian Perez de Castro", "address": "aperez@igalia.com", "name-addr": "Adrian Perez de Castro <aperez@igalia.com>", "count": 1}
+{"name": "Alex Botero-Lowry", "address": "alex.boterolowry@gmail.com", "name-addr": "Alex Botero-Lowry <alex.boterolowry@gmail.com>", "count": 2}
+{"name": "Alexander Botero-Lowry", "address": "alex.boterolowry@gmail.com", "name-addr": "Alexander Botero-Lowry <alex.boterolowry@gmail.com>", "count": 4}
+{"name": "Aron Griffis", "address": "agriffis@n01se.net", "name-addr": "Aron Griffis <agriffis@n01se.net>", "count": 1}
+{"name": "Carl Worth", "address": "cworth@cworth.org", "name-addr": "Carl Worth <cworth@cworth.org>", "count": 12}
+{"name": "Chris Wilson", "address": "chris@chris-wilson.co.uk", "name-addr": "Chris Wilson <chris@chris-wilson.co.uk>", "count": 1}
+{"name": "François Boulogne", "address": "boulogne.f@gmail.com", "name-addr": "François Boulogne <boulogne.f@gmail.com>", "count": 1}
+{"name": "Ingmar Vanhassel", "address": "ingmar@exherbo.org", "name-addr": "Ingmar Vanhassel <ingmar@exherbo.org>", "count": 1}
+{"name": "Israel Herraiz", "address": "isra@herraiz.org", "name-addr": "Israel Herraiz <isra@herraiz.org>", "count": 1}
+{"name": "Jan Janak", "address": "jan@ryngle.com", "name-addr": "Jan Janak <jan@ryngle.com>", "count": 4}
+{"name": "Jjgod Jiang", "address": "gzjjgod@gmail.com", "name-addr": "Jjgod Jiang <gzjjgod@gmail.com>", "count": 2}
+{"name": "Keith Packard", "address": "keithp@keithp.com", "name-addr": "Keith Packard <keithp@keithp.com>", "count": 7}
+{"name": "Lars Kellogg-Stedman", "address": "lars@seas.harvard.edu", "name-addr": "Lars Kellogg-Stedman <lars@seas.harvard.edu>", "count": 5}
+{"name": "Mikhail Gusarov", "address": "dottedmag@dottedmag.net", "name-addr": "Mikhail Gusarov <dottedmag@dottedmag.net>", "count": 5}
+{"name": "Olivier Berger", "address": "olivier.berger@it-sudparis.eu", "name-addr": "Olivier Berger <olivier.berger@it-sudparis.eu>", "count": 1}
+{"name": "Rolland Santimano", "address": "rollandsantimano@yahoo.com", "name-addr": "Rolland Santimano <rollandsantimano@yahoo.com>", "count": 1}
+{"name": "Stewart Smith", "address": "stewart@flamingspork.com", "name-addr": "Stewart Smith <stewart@flamingspork.com>", "count": 3}
+EOF
+test_expect_equal_file OUTPUT EXPECTED
 
 test_done
-- 
2.1.1