AfflictionNetworks/depot
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Commit Graph

  • 0fc1d68bbe Fixed up svn:keywords property. Michael Bayne 2008-11-16 08:03:33 +00:00
  • 9c0dbb5b47 Fixed up packages. Michael Bayne 2008-11-16 08:01:19 +00:00
  • c60181a484 We need a GWT module. Michael Bayne 2008-11-16 07:53:22 +00:00
  • 50053f5282 Oh boy, we have to build a GWT source jar file. Michael Bayne 2008-11-16 07:34:30 +00:00
  • 7a23a6bd49 Copy the tests over. I couldn't migrate these with the rest of the SVN migration because svndumpfilter is a royal pain in my ass. Michael Bayne 2008-11-16 07:30:20 +00:00
  • 711fd44911 No longer needed. Michael Bayne 2008-11-16 06:55:13 +00:00
  • 601a15c00f Added snapshot target. Michael Bayne 2008-11-16 06:49:09 +00:00
  • b32b4eecc3 Updated the source headers and added the license file. Michael Bayne 2008-11-16 06:17:45 +00:00
  • b9604c2d56 Ignore the jar dependencies we keep in here. Michael Bayne 2008-11-16 06:09:33 +00:00
  • 40c1480039 More tidying, added Javadoc overview. Michael Bayne 2008-11-16 06:08:56 +00:00
  • 745c380e8d Repackage Depot as com.samskivert.depot as it is now a standalone library. We're also going to allow ourselves the luxury of a Google Collections dependency. Michael Bayne 2008-11-16 05:59:28 +00:00
  • 6959383b6c Unified Query and Modifier under Operation which cleans up some niggles in PersistenceContext. Michael Bayne 2008-11-13 23:35:12 +00:00
  • f653e59a66 That was broken code-in-progress and not meant to be commnitted. Par Winzell 2008-11-05 00:19:59 +00:00
  • 49391b1995 When we reinitialize the Throttle with a smaller number of allowed operations within a time period, we should preserve the oldest rather than the newest ones, or we'll most likely be in immediate and unfair violation of the limit. Par Winzell 2008-11-05 00:18:21 +00:00
  • bac0348951 Change these hard-coded defaults. Should also be made configurable. Par Winzell 2008-11-04 19:20:33 +00:00
  • 05f499736e Nixed unused import, commented out others that go with commented out code. Michael Bayne 2008-11-01 05:39:27 +00:00
  • 8ec25c5a6d Make the 1.5 compiler happy. Michael Bayne 2008-10-31 00:42:53 +00:00
  • 2b8bc0c97f When we can switch entirely to 1.6 we can use Postgres's value = any(?::int[]) form to pass as many damned keys we want in a giant array. At that time we can nix all the >32,768 key fiddly business. Michael Bayne 2008-10-31 00:17:36 +00:00
  • ec76d4d29f Oops, avoid introducing a bug when nulls == zeros == values.length. We want to return null in that case. Michael Bayne 2008-10-24 05:18:50 +00:00
  • 2309c490f7 Compromise on our 0 == null stance for primary key columns. See the code comments for further details. Michael Bayne 2008-10-24 05:13:28 +00:00
  • ba97743754 Make sure we use the up to date metadata when doing our stale column check. Michael Bayne 2008-10-22 01:09:22 +00:00
  • 19287550f4 More javadoc cleanup. Michael Bayne 2008-09-30 00:23:57 +00:00
  • 4c94296b67 Javadoc fixes. Michael Bayne 2008-09-30 00:12:08 +00:00
  • fee6be6c2d Added support for data migrations, which are run after all schema migrations have been run and are also coordinated between disparate processes by an in-database lock system. Michael Bayne 2008-09-25 23:40:01 +00:00
  • 2b21105dc1 Call EntityMigration SchemaMigration because that's what it's for and I want to add a new DataMigration mechanism so that people can stop using schema migrations to do data migration. Michael Bayne 2008-09-25 22:27:58 +00:00
  • 2c9c9b47c1 Let's preserve the ability for callers with uncomplicated needs to create and initialize a persistence context in one fell swoop. Michael Bayne 2008-09-24 19:09:17 +00:00
  • 0e68c8a599 Change the way initialization is handled (a bit). It is now possible to create a PersistenceContext and give it out to all of your repositories and delay activating the context until you are ready for everyone to start talking to the database. In this way you can enforce that no one does any database fooling around before it's OK. Michael Bayne 2008-09-24 18:33:25 +00:00
  • 81c25ae4f1 Properly handle int[] fields when encountered as unassociated values. We already do the right thing when we process them as fields of persistent records. Michael Bayne 2008-09-23 17:18:23 +00:00
  • f33a376f7e Formatting tweaks. Michael Bayne 2008-09-23 17:17:05 +00:00
  • d7f864dfe4 Don't try to manage the cache manager. Push that burden out to the caller who may have their own needs and arrangements. Michael Bayne 2008-09-20 03:30:49 +00:00
  • 3e8ac68af6 Properly handle @GeneratedValue(strategy=GenerationType.IDENTITY, initialValue=N) for non-zero values of N. This is complicated. Depot used to "initialize" value generators pretty much every time a table schema was migrated. For TableValueGenerator it could just NOOP if the proper row was in the sequence table. IdentityValueGenerator was just broken and ignored any request to configure an initialValue, so it never did anything. Michael Bayne 2008-09-17 19:09:08 +00:00
  • 5ac15a99ca Use varying size blobs for int[] columns like we do byte[]. Michael Bayne 2008-09-16 22:51:59 +00:00
  • b64658fd83 Handle conversion from int[] to byte[] manually rather than relying on the JDBC driver to know how to do it. Michael Bayne 2008-09-16 22:49:11 +00:00
  • 9112b3c7e2 Javadoc fix from Dave. Michael Bayne 2008-09-16 21:54:21 +00:00
  • 84dd3c6b5f Amazingly, we had apparently never tried to insert a preconfigured row into a table that had field generators on its primary key. The code just blindly overwrite the supplied primary key and ran the generators anyway. Michael Bayne 2008-09-16 21:33:22 +00:00
  • e25aa25bd6 Nixed unused imports. Michael Bayne 2008-09-16 02:01:19 +00:00
  • 2b1b2e8334 Missed a spot. Michael Bayne 2008-09-15 18:46:48 +00:00
  • 759fb587dc Name things a bit more clearly to indicate that we're manipulating field definitions, not necessarily field overrides. Michael Bayne 2008-09-15 18:42:12 +00:00
  • a507308c2b Zell reminds me that I want to use FieldDefinition rather than FieldOverride which is meant to be used only in certain circumstances. Michael Bayne 2008-09-15 18:37:04 +00:00
  • 7e8c7fadcb Don't log a warning if we don't get a result for every primary key provided when we're doing a query with primary keys provided by a random caller. We have know idea where they got those keys or where they have been before we got our filthy mitts on them. Michael Bayne 2008-09-15 18:33:21 +00:00
  • 2f57697c32 We're going to allow FieldOverride on @Computed records, just not on @Computed fields. Maybe Zell will remind me why this was disallowed, but it seems that to disallow this is to make FieldOverride useless. Michael Bayne 2008-09-15 18:22:18 +00:00
  • 98ac644ad9 Ah, entities.keySet() is misleading because it contains records that were found in the cache as well as records loaded from earlier passes in cases where we split our query up to avoid database breaking in() clause sizes. Now we track exactly what we got this time around and report that. Michael Bayne 2008-09-12 01:55:54 +00:00
  • 2611a1076b Let's have a stack trace here because these mismatches are just plain bizarre. Michael Bayne 2008-09-12 01:49:06 +00:00
  • d47e85fe9f If we're going to magically convert ByteEnum to an integer type, then we should magically assign it a default value of 0 like we do for our other integer types. Michael Bayne 2008-09-12 00:24:14 +00:00
  • 96bd8b56a0 We're seeing some weirdness here, so let's log the original statement along with the actual set of wanted and obtained keys when we have a mismatch. Michael Bayne 2008-09-10 18:09:23 +00:00
  • 27b619d59b Let's log a stack trace on lazy initialization so that we can see who the culprit is. Michael Bayne 2008-09-10 18:01:24 +00:00
  • c4ddf1853c Created an explicit EntityMigration.Add for when you want to add a column to a database and need to provide a defaultValue at the time that you add it, but don't want to have to define a permanent default value on the field. Michael Bayne 2008-09-10 02:00:19 +00:00
  • 00d6ae5492 Don't try to do the primary key lookup on delete for records that define no primary key. Michael Bayne 2008-09-08 19:01:30 +00:00
  • ffeb6534c4 Cope with the goddamned broken generics implementation in JDK 1.5. Michael Bayne 2008-09-08 18:58:12 +00:00
  • f1dd34e4b3 We need Key to export a Key.Expression which we can use inside the SQL for a KeySet so that the "where" keyword is not repeated every time we bind a key. Michael Bayne 2008-09-08 17:54:11 +00:00
  • 56c4394b29 Beans/frank. Michael Bayne 2008-09-08 16:48:32 +00:00
  • d5619a3b28 Automatically add the shadow to our type set for @Computed(shadowOf=FooRecord) record classes. Michael Bayne 2008-09-06 02:38:20 +00:00
  • 9a7a78080c We need to set up our _builder in WithKeys. Also moved loadRecords() down below loadAndResolve(). Michael Bayne 2008-09-06 02:27:53 +00:00
  • 2f335c4dc8 Actually, if we use Key we don't need to specify the type. It's already in the keys. So we can have loadAll(Class,Collection<Comparable>) and loadAll(Collection<Key>) and not run into type erasure problems. Yay! Michael Bayne 2008-09-06 02:06:44 +00:00
  • 9a21071c82 Type jockeying. Michael Bayne 2008-09-06 01:59:52 +00:00
  • 35f654a8c1 Added loadAll() which takes a collection of primary keys. Since we've already got the primary keys, this goes right to phase two wherein we check the cache for hits and load the remaining records by primary key (and cache them). Michael Bayne 2008-09-06 01:54:08 +00:00
  • 52de64001d Allow the cache to be explicitly bypassed. Michael Bayne 2008-09-06 00:34:29 +00:00
  • 089706427e Be sure to report that we throw DatabaseException for documentation purposes. Michael Bayne 2008-09-06 00:30:25 +00:00
  • fe1aff3842 Tidying. Michael Bayne 2008-09-06 00:17:52 +00:00
  • 84c32cc5e3 We need to use the key in our or clause, not its non-existent where expression. Michael Bayne 2008-09-06 00:13:05 +00:00
  • 263529c929 Use the SelectClause to obtain our types so that the main record class is always properly added. Also made SelectClause freak out if it's provided with bogus clauses. Michael Bayne 2008-09-05 23:17:37 +00:00
  • 8b6f575af8 Nixed unused import. Michael Bayne 2008-09-05 23:03:49 +00:00
  • ae9e0acef8 We need to let you express whether or not you want your keys from the master server or if it's OK to get them from a (potentially slightly out of date) replica. If you're going to modify rows based on the results of findAllKeys(), you should probably get them from the master server, if you're just reading data, you can talk to a replica. Michael Bayne 2008-09-05 23:03:08 +00:00
  • 2e4decf18a Added findAllKeys() for when you only need the keys. Michael Bayne 2008-09-05 22:58:05 +00:00
  • 95d316d28e One more unused import. Michael Bayne 2008-09-05 22:43:38 +00:00
  • 709a6ce203 Unused import patrol. Michael Bayne 2008-09-05 22:43:06 +00:00
  • cd8184f34f Some method ordering. Michael Bayne 2008-09-05 22:37:23 +00:00
  • fb64881c82 We're handling Key query construction directly instead of letting it look like a WhereClause, so we need to add our own where. Michael Bayne 2008-09-05 22:36:53 +00:00
  • 4c41561f0f PrimaryKeySet -> KeySet. Michael Bayne 2008-09-05 22:16:55 +00:00
  • 249f4a5106 Now we don't have to provide our values as an ArrayList. Michael Bayne 2008-09-05 20:07:43 +00:00
  • 4bcfda861d Ah, Key actually is PrimaryKey and does not in fact keep the primary key columns around. Cleaned some things up along those lines. Michael Bayne 2008-09-05 20:05:44 +00:00
  • 84622e56c7 Formalized the two-phase "get primary keys that match an arbitrary query, then load or modify the rows that match those keys" pattern so that we can almost always just magically do the right thing with regard to the cache. Added a version of deleteAll() that makes use of this. Michael Bayne 2008-09-05 19:39:08 +00:00
  • 67bdfda3fe Changed Depot to throw a RuntimeException on unexpected database error instead of a checked exception. Invariably, we don't do anything with our PersistenceExceptions except let them percolate all the way to the top and then log a warning. We can probably automate the process of logging a warning with useful information and save ourselves the trouble of doing it manually everywhere. Michael Bayne 2008-09-03 16:10:06 +00:00
  • 2fd131c0d2 Any collection of anything that extends QueryClause is generally all we need. Michael Bayne 2008-09-03 00:55:07 +00:00
  • beef018b03 Nixed unused import. Michael Bayne 2008-08-25 18:16:35 +00:00
  • 43e1a38703 Let's go all the way up to Short.MAX_VALUE. Michael Bayne 2008-08-25 16:27:43 +00:00
  • d54c633401 Don't try to use an IN() clause with more than 16738 entries. Michael Bayne 2008-08-25 15:57:43 +00:00
  • 01962836f7 Nixed unneeded log message. Michael Bayne 2008-08-20 18:17:52 +00:00
  • 8dec5386f2 Don't require a schema version for computed records. Michael Bayne 2008-08-20 01:28:31 +00:00
  • a670beb665 A couple of changes: Michael Bayne 2008-08-20 00:35:12 +00:00
  • 64d186d08b Report a better error message if a bogus rename is attempted. Some formatting cleanup. Michael Bayne 2008-08-20 00:22:41 +00:00
  • 22fb614f34 Unfiddle long lines. Michael Bayne 2008-08-16 20:30:52 +00:00
  • 338bb62d9f PreparedStatement.setObject() does a lot of runtime type checking, so we don't have to do setShort(), setString(), etc. But we do have to convert ByteEnum's to their byte representation. Par Winzell 2008-08-07 06:02:41 +00:00
  • 86f3b1e293 The PostgreSQL datatypes for float and double are REAL and DOUBLE PRECISION. Par Winzell 2008-08-07 05:45:25 +00:00
  • a48c93be4a From Dave: a bunch of @Override annotating and some other cleanup. Michael Bayne 2008-08-06 12:59:06 +00:00
  • a0731b34c1 Trimmed trailing whitespace. Michael Bayne 2008-08-04 12:37:23 +00:00
  • f07694895e Fixed some bogus null checks and assignments. Michael Bayne 2008-08-04 00:49:37 +00:00
  • 8f2cf98c17 Le metric crapload of generics cleanup. Michael Bayne 2008-07-30 13:53:24 +00:00
  • 4e002054e4 Unused import pruning. Michael Bayne 2008-07-26 20:14:38 +00:00
  • e3adfea99e Handle SEQUENCE strategies, which are not yet implemented. This'll fix David Hoover's super-strict compilation urges, besides being a generally good idea. Par Winzell 2008-07-21 19:36:47 +00:00
  • eac0e214ef Always do a (potentially distributed) cache remove on invalidation, regardless of what's in the local cache. From Kyle Sampson. Par Winzell 2008-07-17 12:16:39 +00:00
  • 8f1b3ad8b0 Ray Greenwell 2008-07-15 00:03:10 +00:00
  • 33e159047d Use new logging shim. Michael Bayne 2008-05-27 18:40:39 +00:00
  • c16003177d Ray Greenwell 2008-05-19 18:32:58 +00:00
  • 7e93da6fc0 Ray Greenwell 2008-05-16 03:20:27 +00:00
  • 195fabe2cc We need parens so that operators are evaluated in the right order. This adds extra params around a bunch of things. I'll let the gods of Depot sort that out. Ray Greenwell 2008-05-16 02:10:07 +00:00
  • 31b886aed2 Michael Bayne 2008-04-25 18:32:44 +00:00
  • 4eb7d9ecd3 Pass -Dcom.samskivert.jdbc.depot.debug=true to have Depot log queries before they are executed. Michael Bayne 2008-04-23 23:29:07 +00:00
  • 513ce9aa60 Yank our nomenclature into consistency -- indices, not indexes; fields(), not fieldNames(), as per Ray's suggestion. Par Winzell 2008-04-14 20:18:45 +00:00
  • 13121249cb The query is in no way guaranteed to uniquely keyed rows, so let's stuff the result into a Set. We should probably add the 'unique' modifier to the actual query, but that requires a bit more cross-dialect-and-version research than I'm willing to give it just now. Par Winzell 2008-04-06 14:57:14 +00:00