SphinxSE patches for MariaDB 10.1 and MySQL 5.7

[adriannuta]

Should superseed #13

[arekm]

on percona server (based on mysql 5.7) it fails with:

percona-server-5.7.21-20/storage/sphinx/snippets_udf.cc: In member function 'int CSphUrl::Connect()':
percona-server-5.7.21-20/storage/sphinx/snippets_udf.cc:383:25: error: 'inet_addr' was not declared in this scope
   if ( (int)( ip_addr = inet_addr ( m_sHost ) )!=(int)INADDR_NONE )

and in case if anyone needs a fix:

--- storage/sphinx/snippets_udf.cc~     2016-07-19 12:42:18.000000000 +0200
+++ storage/sphinx/snippets_udf.cc      2018-03-10 20:16:19.241113584 +0100
@@ -34,6 +34,9 @@
 #include "../mysql_priv.h"
 #endif
 
+#if  !defined(MARIADB_BASE_VERSION) && MYSQL_VERSION_ID>=50709
+       #include <arpa/inet.h>
+#endif
 #include <mysys_err.h>
 #include <my_sys.h>

[shodanium]

SphinxSE is not supported anymore starting from 3.0 anyway. Switching to SphinxQL is recommended.

[zobzn]

SphinxSE is not supported anymore

Why? It's such a great feature. It gives the possibility to join mysql and sphinx search results!

[shodanium]

SphinxSE is not supported anymore

Why? It's such a great feature. It gives the possibility to join mysql and sphinx search results!

For several different reasons. Basically, at this time it's an extremely obscure feature, difficult from our end to maintain, difficult from the user end to both deploy and test, not that generic (ie. tied to MySQL specifically), and last but not least, actually almost unused in prod these days, as far as I could tell.

Starting from 3.x you should be able to either store everything you need to store in Sphinx itself, and avoid an extra roundtrip to MySQL.

Or merely fetch a handful of document IDs and pass them to MySQL (or whatever other DB). Normally that should be only a few extra lines of code :)

[midn1]

Your first solution is stupidly wasteful of storage, and your second is slower than if SphinxSE directly takes as much as it needed.