Commit Graph

13125 Commits (962dade15596d0aeff3baba0b0b85c17524766c6)

Author SHA1 Message Date
AlexRa accf8c420d SQLite: CreatorID added (and asset_flags moved) to the migration script 2010-05-19 21:49:30 +03:00
AlexRa c6977cbd4d Added CreatorID to SQLite asset data 2010-05-19 21:49:24 +03:00
AlexRa ccee95552f A bit of harmless refactoring in SQLiteAssetData.cs 2010-05-19 21:49:22 +03:00
AlexRa d71c483bf0 Prebuild: added *.migrations as resources to MSSQL, SQLiteLegacy (just in case) 2010-05-19 21:49:20 +03:00
AlexRa 527a257b9f Kind of fixed Melanie's "Exception(sql)" correction
Throwing an Ex. with SQL command in the message looks weird,
this is a bit better, but I'm still not sure if that's the
proper way to handle.  Also, there is a catch one level up,
so is this one necessary?
2010-05-19 21:49:18 +03:00
Tom Grimshaw fc76b5f589 Fix a null reference issue in cmGetAvatarList() 2010-05-19 00:24:54 -07:00
Tom Grimshaw 71a95900fd Merge branch 'careminster-presence-refactor' of ssh://3dhosting.de/var/git/careminster into careminster-presence-refactor 2010-05-19 00:21:38 -07:00
Tom Grimshaw ba36c62caa Fix a null reference issue in cmGetAvatarList() 2010-05-19 00:20:51 -07:00
Melanie e4b8d76b10 Change appender to deal with line feeds more intelligently.
Change migration error reporting to not truncate the statement when
reporting. It's a bit messier than the old error reporting, but at least
one gets an idea of what could be wrong again. And things look a lot
neater now.
2010-05-19 04:17:56 +01:00
Melanie 0c209a469b Clean up output a bit 2010-05-19 03:48:03 +01:00
Melanie 20642f2f21 Fix the migration message to say "Continuing" again. Remove line feed, which
prevented the full message from displaying.
2010-05-19 03:26:37 +01:00
Melanie 9fa8013ca5 Remove the return that was inserted in the last merge to allow
migrations to continue in the face of an error
2010-05-19 03:07:31 +01:00
Melanie dedc0c0bd4 Revert "Allow migration steps to fail again without bringing down the house"
This reverts commit 167db50259.
2010-05-19 02:47:31 +01:00
Melanie 167db50259 Allow migration steps to fail again without bringing down the house 2010-05-19 02:33:23 +01:00
Melanie 298368dd99 Merge commit 'alex/mssql-estate-bugfix'
This fixes some type conversion bugs and also makes the estate lists
work properly for MSSQL. Strawberry, please check this.

Signed-off-by: Melanie <melanie@t-data.com>
2010-05-18 23:39:09 +01:00
Melanie d2bc673667 Make m_log in migrations private. Define new m_log in derived class 2010-05-18 23:33:05 +01:00
Melanie caf61ab7d8 Binary Guids are 16 chars long. Fix parser. 2010-05-18 23:22:30 +01:00
Melanie 4fee3f9548 Revert a revert? Am i mad? Don't answer that....!
Revert "Revert "Looks like the new files were never added to prebuild.xml""

This reverts commit f253758c2e.
2010-05-18 23:15:29 +01:00
Melanie 397326ddfa Merge commit 'alex/Migrations'
Another stab at the tests

Signed-off-by: Melanie <melanie@t-data.com>
2010-05-18 23:02:10 +01:00
AlexRa deae030145 Some more corrections after MySQL connector update 2010-05-19 02:28:19 +03:00
AlexRa 8a0c5d14d4 All (?) MySQL stores fixed to use DBGuid.FromDB()
This was needed if we want to update to the latest MySQL
connector dll.  It automatically converts CHAR(36) to
Guids, so getting them as strings no longer works.

By using DBGuid.FromDB(), we unlink from any particular
storage format of GUIDs, could even make them BINARY(16)
if we like.

Actually not all MySql units are touched, but the remaining ones don't
seem to be affected (they don't read GUIDs from DB)
2010-05-19 01:33:02 +03:00
AlexRa a27d49b188 Added DBGuids.cs (static func DBGuid.FromDB()
This DBMS-independent function to be used converting UUIDs
from whatever format used in the DB (string/binary/Guid).
This is mostly needed for MySQL, as in MSSQL they are always
UNIQUEIDENTIFIERs and in SQLite always strings (but would look
better if we use it there anyway).
2010-05-19 01:32:50 +03:00
Melanie f253758c2e Revert "Looks like the new files were never added to prebuild.xml"
This reverts commit df49756e7b.
2010-05-18 21:29:02 +01:00
Melanie df49756e7b Looks like the new files were never added to prebuild.xml 2010-05-18 21:15:58 +01:00
Melanie 6eef334059 Merge commit 'alex/Migrations'
Adding what, in the World according to Alex, will fix the tests

Signed-off-by: Melanie <melanie@t-data.com>
2010-05-18 21:00:10 +01:00
AlexRa 020f38774f MS SQL migrations converted to the new format 2010-05-18 22:16:13 +03:00
AlexRa 6e7b3950d7 Migrations for SQLite converted to new format 2010-05-18 22:16:10 +03:00
AlexRa e4419c34c3 Converted MySQL migration history to the new format
Replaced all NNN_StoreName.sql migration resources with a more
readable, single-file-per-store
2010-05-18 22:16:06 +03:00
AlexRa 4ebb985b46 Added MySqlMigrations.cs (supports stored proc/funcs)
Uses MySqlScript class to correctly run proc/func definitions
that need delimiter change. Requires MySql.Data.dll 6.2 or later.
2010-05-18 22:16:03 +03:00
AlexRa ade2e5a9d2 Embedded MySql.Data.dll updated to 6.2.3.0.
This is necessary to correct a known problem with the
DELIMITER command in previous versions of the client library.
2010-05-18 22:16:00 +03:00
AlexRa f4450ccf4f Migration.cs supports single-file migration history format
Scans for migration resources in either old-style "scattered" (one file per version)
or new-style "integrated" format (single file "Resources/{StoreName}.migrations[.nnn]") with ":VERSION nnn" sections).
In the new-style migrations it also recognizes ':GO' separators for parts of the SQL script
that must be sent to the server separately.  The old-style migrations are loaded each in one piece
and don't support the ':GO' feature.

Status:  TESTED and works fine in all modes!
2010-05-18 22:15:56 +03:00
AlexRa d95f232629 Massive bugfix in MSSQLEstateData (now works!) 2010-05-18 21:08:51 +03:00
Melanie e260610c00 Merge branch 'Migrations'
This merges AlexRa's work on migration streamlining. This merge is experimental.
If it causes issues, feel free to back out.

Signed-off-by: Melanie <melanie@t-data.com>
2010-05-18 19:03:16 +01:00
Melanie Thielker 55f124745f Allow remote admin to be used on a different port from the main region port 2010-05-18 19:25:40 +02:00
Melanie Thielker 3e1f2ddb8e Allow remote admin to be used on a different port from the main region port 2010-05-18 17:01:44 +01:00
Melanie 1c3790a70d Merge branch 'master' into careminster-presence-refactor 2010-05-18 15:31:19 +01:00
Melanie df04ccfdc3 Change an exception reporting to include the stack trace. "e.Message" just
_won't_ do. Might look more elegant, but it's too little information.
2010-05-18 15:30:08 +01:00
Tom Grimshaw bb0806c61e Don't send kill packets to child agents as we close them 2010-05-18 03:44:09 -07:00
Tom Grimshaw 1c040d8c1e Fix to the scenario where we send an agent to a neighbouring sim (via teleport), then tell our neighbours to close the agents.. thereby disconnecting the user. Added a new CloseChildAgent method in lieu of CloseAgent. This has been a long standing problem - with any luck this will cure it. 2010-05-18 03:24:43 -07:00
Tom Grimshaw ed1cfb5245 Missed a nullref in the previous commit 2010-05-18 02:33:41 -07:00
Tom Grimshaw 91b1d17e5b Fix for hanging on "Connecting to region".. caused by packets being processed before the presence has bound to receive events. Fixed this by adding packets to a queue and then processing them when the presence is ready. 2010-05-18 01:09:47 -07:00
Tom Grimshaw ae7c30a0f3 While examining our 10,10,10 issue I discovered that several threads were locked with each other in an unusual place; I've moved the piece of code inside a lock[] context and so far it's looking promising. 2010-05-18 00:07:17 -07:00
Melanie Thielker eb5a95d26b Prevent an "index out of range" error on login 2010-05-18 03:02:36 +02:00
Tom Grimshaw c25c40a3f9 Following on from the last commit, the stop/disconnect packet should NOT be sent when the user logs out. This eliminates the annoying "crash on log out" bug, and allows the client to commit important information at the last moment (attachment data!) 2010-05-17 15:24:28 -07:00
Tom Grimshaw 6bc7e3429f Provide interface to prevent the client close function from sending the stop packet 2010-05-17 14:32:21 -07:00
Tom Grimshaw 8f838c722d When killing a zombie session, don't send the stop packet since it often has the effect of killing a newly connected client. 2010-05-17 14:14:19 -07:00
Melanie a003c36de5 Merge branch 'master' into careminster-presence-refactor 2010-05-16 20:32:58 +01:00
Melanie a6023ea89e Prevent a crash when the two agent dictionaries get out of sync 2010-05-16 20:31:02 +01:00
Diva Canto 2fb79646c6 Fixes mantis #4691 for real. This time I tested it, and it works. 2010-05-16 10:32:57 -07:00
Melanie d8bf370fd6 Merge branch 'master' into careminster-presence-refactor 2010-05-16 17:22:38 +01:00