support xapian ngrams

Xapian supports an "ngrams" option to help with languages/scripts
without explicit wordbreaks, such as Chinese / Japanese / Korean.

Add some plumbing for supporting this in mu as well. Experimental for
now.

mu/mu-cmd-info.cc

@@ -202,6 +202,8 @@ topic_store(const Mu::Store& store, const Options& opts)
 		info.add_row({"ignored-address", c});
 
 	info.add_row({"messages in store", mu_format("{}", store.size())});
+	info.add_row({"support-ngrams",  conf.get<Config::Id::SupportNgrams>() ? "yes" : "no"});
+
 	info.add_row({"last-change", tstamp(store.statistics().last_change)});
 	info.add_row({"last-index",  tstamp(store.statistics().last_index)});
 

mu/mu-cmd-init.cc

@@ -55,6 +55,8 @@ Mu::mu_cmd_init(const Options& opts)
 			conf.set<Config::Id::PersonalAddresses>(opts.init.my_addresses);
 		if (!opts.init.ignored_addresses.empty())
 			conf.set<Config::Id::IgnoredAddresses>(opts.init.ignored_addresses);
+		if (opts.init.support_ngrams)
+			conf.set<Config::Id::SupportNgrams>(true);
 
 		return Store::make_new(opts.runtime_path(RuntimePath::XapianDb),
 				       opts.init.maildir, conf);

mu/mu-options.cc

@@ -457,13 +457,16 @@ sub_init(CLI::App& sub, Options& opts)
 		       "Maximum allowed message size in bytes");
 	sub.add_option("--batch-size", opts.init.batch_size,
 		       "Maximum size of database transaction");
+	sub.add_option("--support-ngrams", opts.init.support_ngrams,
+		       "Support CJK n-grams if for querying/indexing");
 	sub.add_flag("--reinit", opts.init.reinit,
 		       "Re-initialize database with current settings")
 		->excludes("--maildir")
 		->excludes("--my-address")
 		->excludes("--ignored-address")
 		->excludes("--max-message-size")
-		->excludes("--batch-size");
+		->excludes("--batch-size")
+		->excludes("--support-ngrams");
 }
 
 static void

mu/mu-options.hh

@@ -185,13 +185,15 @@ struct Options {
 	 * Init
 	 */
 	struct Init {
-		std::string	maildir;		/**<  where the mails are */
-		StringVec	my_addresses;		/**<  personal e-mail addresses */
-		StringVec	ignored_addresses;	/**<  addresses to be ignored for
+		std::string	maildir;		/**< where the mails are */
+		StringVec	my_addresses;		/**< personal e-mail addresses */
+		StringVec	ignored_addresses;	/**< addresses to be ignored for
 							 * the contacts-cache */
-		OptSize         max_msg_size;		/**<  max size for message files */
-		OptSize		batch_size;		/**<  db transaction batch size */
-		bool		reinit;			/**<  re-initialize  */
+		OptSize         max_msg_size;		/**< max size for message files */
+		OptSize		batch_size;		/**< db transaction batch size */
+		bool		reinit;			/**< re-initialize  */
+		bool            support_ngrams;         /**< support CJK etc. ngrams */
+
 	} init;
 
 	/*

mu/mu.cc

@@ -90,6 +90,14 @@ handle_result(const Result<void>& res, const Mu::Options& opts)
 int
 main(int argc, char* argv[])
 {
+	/*
+	 * We handle this through explicit options
+	 */
+	g_unsetenv("XAPIAN_CJK_NGRAM");
+
+	/*
+	 * set up locale
+	 */
 	setlocale(LC_ALL, "");
 
 	/*