From 86be0c3098d8e27807bc618126b8babfe7f93928 Mon Sep 17 00:00:00 2001
Message-Id: <86be0c3098d8e27807bc618126b8babfe7f93928.1716352610.git.mdw@distorted.org.uk>
From: Mark Wooding <mdw@chiark.greenend.org.uk>
Date: Sat, 29 Dec 2007 16:48:25 +0000
Subject: [PATCH] build default stopword list into server
Organization: Straylight/Edgeware

From: rjk@greenend.org.uk <>

---
 CHANGES                    |  3 ++
 debian/etc.disorder.config |  7 ----
 doc/disorder_config.5.in   |  3 ++
 examples/config.sample.in  | 10 ------
 lib/configuration.c        | 67 ++++++++++++++++++++++++++++++++++++++
 tests/search.py            |  2 ++
 6 files changed, 75 insertions(+), 17 deletions(-)

diff --git a/CHANGES b/CHANGES
index d558d9d..6b1fc20 100644
--- a/CHANGES
+++ b/CHANGES
@@ -6,6 +6,9 @@ See ChangeLog.d/* for detailed revision history.
 
 Users are now stored in the database rather than a configuration file.
 
+The server now has a built-in list of stopwords, so only additions need
+be mentioned in the configuration file.
+
 ** Web Interface
 
 The web interface now uses cookies to remember user identity, and allows
diff --git a/debian/etc.disorder.config b/debian/etc.disorder.config
index f5fd10c..55e42d1 100644
--- a/debian/etc.disorder.config
+++ b/debian/etc.disorder.config
@@ -44,12 +44,5 @@ mixer /dev/mixer
 # The channel to control
 channel pcm
 
-# Stopwords (i.e. ignored words) for the track search facility.
-stopword 01 02 03 04 05 06 07 08 09 10
-stopword 1 2 3 4 5 6 7 8 9
-stopword 11 12 13 14 15 16 17 18 19 20
-stopword 21 22 23 24 25 26 27 28 29 30
-stopword the a an and to too in on of we i am as im for is
-
 # include debconf configuration
 include /etc/disorder/conf.debconf
diff --git a/doc/disorder_config.5.in b/doc/disorder_config.5.in
index 3cb36af..8bc1963 100644
--- a/doc/disorder_config.5.in
+++ b/doc/disorder_config.5.in
@@ -571,6 +571,9 @@ over track names.
 .IP
 If \fBstopword\fR is used without arguments then the list of stopwords is
 cleared.
+.IP
+There is a default set of stopwords built in, but this option can be used to
+augment or replace that list.
 .TP
 .B tracklength \fIPATTERN\fR \fIMODULE\fR
 Specifies the module used to calculate the length of files matching
diff --git a/examples/config.sample.in b/examples/config.sample.in
index 1bf4088..6f214be 100644
--- a/examples/config.sample.in
+++ b/examples/config.sample.in
@@ -44,13 +44,3 @@ user jukebox
 mixer /dev/mixer
 # The channel to control
 channel pcm
-
-# URL of the web interface
-url http://jukebox.anjou.terraraq.org.uk/
-
-# Stopwords (i.e. ignored words) for the track search facility.
-stopword 01 02 03 04 05 06 07 08 09 10
-stopword 1 2 3 4 5 6 7 8 9
-stopword 11 12 13 14 15 16 17 18 19 20
-stopword 21 22 23 24 25 26 27 28 29 30
-stopword the a an and to too in on of we i am as im for is
diff --git a/lib/configuration.c b/lib/configuration.c
index 7ed9276..bb30c2f 100644
--- a/lib/configuration.c
+++ b/lib/configuration.c
@@ -1044,12 +1044,77 @@ static int config_include(struct config *c, const char *path) {
   return ret;
 }
 
+static const char *const default_stopwords[] = {
+  "stopword",
+
+  "01",
+  "02",
+  "03",
+  "04",
+  "05",
+  "06",
+  "07",
+  "08",
+  "09",
+  "1",
+  "10",
+  "11",
+  "12",
+  "13",
+  "14",
+  "15",
+  "16",
+  "17",
+  "18",
+  "19",
+  "2",
+  "20",
+  "21",
+  "22",
+  "23",
+  "24",
+  "25",
+  "26",
+  "27",
+  "28",
+  "29",
+  "3",
+  "30",
+  "4",
+  "5",
+  "6",
+  "7",
+  "8",
+  "9",
+  "a",
+  "am",
+  "an",
+  "and",
+  "as",
+  "for",
+  "i",
+  "im",
+  "in",
+  "is",
+  "of",
+  "on",
+  "the",
+  "to",
+  "too",
+  "we",
+};
+#define NDEFAULT_STOPWORDS (sizeof default_stopwords / sizeof *default_stopwords)
+
 /** @brief Make a new default configuration */
 static struct config *config_default(void) {
   struct config *c = xmalloc(sizeof *c);
   const char *logname;
   struct passwd *pw;
+  struct config_state cs;
 
+  cs.path = "<internal>";
+  cs.line = 0;
+  cs.config = c;
   /* Strings had better be xstrdup'd as they will get freed at some point. */
   c->gap = 2;
   c->history = 60;
@@ -1082,6 +1147,8 @@ static struct config *config_default(void) {
   c->dbversion = 2;
   c->cookie_login_lifetime = 86400;
   c->cookie_key_lifetime = 86400 * 7;
+  if(config_set(&cs, (int)NDEFAULT_STOPWORDS, (char **)default_stopwords))
+    exit(1);
   return c;
 }
 
diff --git a/tests/search.py b/tests/search.py
index 3be6ec9..dab3939 100755
--- a/tests/search.py
+++ b/tests/search.py
@@ -89,6 +89,8 @@ def test():
     check_search_results([u"fi\u0300rst"], first)
     check_search_results([u"THI\u0301RD"], third)
     check_search_results([u"thI\u0301rd"], third)
+    # stopwords shouldn't show up
+    check_search_results(["01"], [])
     
     if failures > 0:
         sys.exit(1)
-- 
[mdw]