| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
|
|
|
|
|
| |
increases and lower memory usage:
- Add a new SPMySQLStreamingResultStore class to SPMySQL.framework. This class acts as both a result set and a data store for the accompanying data, storing the row information in a custom format in a custom malloc zone.
- Amend SPDataStorage to wrap the new class, so original result information is stored in the one location in the custom format. Any edited information is handled by SPDataStorage for clean separation
- Rework table content and custom query data data stores to use the new class. This significantly speeds up data loading, resulting in faster data loads if they weren't previously network constrained, or lower CPU usage otherwise. The memory usage is also lowered, with the memory overhead for many small cells being enormously reduced.
|
| |
|
|
|
|
| |
enable some more
|
|
|
|
| |
Issue #1708
|
|
|
|
| |
instead of nil if a query produces no result set. This allows per-result-set properties to be preserved, fixing issues where information like query execution time was lost - addressing Issue #1577
|
| |
|
| |
|
|
|
|
|
|
|
| |
result of r3894, attempting to improve some exceptions during aborted connections
- Name threads created in SPMySQL.framework
|
|
|
|
|
|
|
|
|
| |
disconnections from external actions, and use that separation to perform safer disconnects
- When closing a database document, add a new notification, and use that to resolve retain cycles affecting connection processes
- Improve connection controller disconnection when the document is closed, fixing crashes, by building on those two features (addresses Issue #1396)
- Use some of the new functionality to improve SSH and MySQL connection cancellation, making both cancelable in the interface and making both respond much more quickly
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
address Issue #877:
- On servers with very short timeouts set the wait_timeout for the session as well as the interactive_timeout to prevent the connection from dropping frequently
- Improve recovery from connection errors, correctly restoring the connection if appropriate and possible
- Allow reconnections to occur recursively by altering the internal tracking mechanism
- Fix some edge cases where the connection would remain locked incorrectly
- Improve error messaging for the "MySQL Server has gone away" network case
|
|
|
|
|
|
|
|
|
|
|
| |
reviewing crash logs and testing a number of situations:
- Improve stability of closing connections after a connection loss
- Minimise prompting a user for connection state restore if closing windows/tabs
- Allow cancellation of keepalive ping threads to prevent crashes after deallocation of parent
- Manually handle ping thread state struct memory to avoid cross-thread deallocation issues
- Improve disconnection speed and resilience
|
|
|
|
| |
of a reference to the retained error string to avoid any overrelease errors later affecting the framework
|
|
|
|
| |
paths, improving general connection robustness in recent builds
|
|
|
|
|
| |
- Improve debug to catch a structure connection issue
|
|
|
|
|
|
|
|
|
|
|
|
| |
background loss of connections:
- Attempt to fix a condition causing a reconnection loop by fixing the order of connection state check and a query variable
- If a connection is lost in the background, only attempt a single reconnect instead of requiring user intervention at once
- Add a new connection state to handle background disconnects
- If the connection has been lost in the background but is about to be used, reconnect it automatically (informing the user of loss if appropriate)
- Don't attempt background reconnections if the connection has not been used for some time
(Also update localisable strings, and tweak navigator controller connection usage)
|
|
|
|
| |
mysql_thread_init() and mysql_thread_end() on each thread as appropriate. This will improve observed MySQL stability, especially in certain circumstances such as LOAD DATA INFILE.
|
|
|
|
|
|
|
| |
improve memory usage during large exports, particularly improving memory which appeared to never be reclaimed
- Fix some minor memory leaks throughout the application
|
|
|
|
|
|
|
|
| |
up-to-date codebase.
- Update the build-mysql-client.sh MySQL build script within SPMySQL.framework to allow it to easily build 5.5 client libraries
- Fix reconnection within SPMySQL.framework after a connection is lost for MySQL servers < 5 when a query is cancelled by killing the connection
|
|
|
|
|
|
|
| |
getters and manually overridden setters
- Fix autoincrement ID retrieval in SPMySQL, addressing issues previewing rows when the setting for "Reload Table After: Adding a row" is not being used
|
|
|
|
| |
mySQLTickQuotedString declarations
|
|
|
|
| |
settings in SPMySQL when connections time out, addressing Issue #1309
|
| |
|
|
|
|
|
|
|
| |
- Fix a bug where the socket path would not be autodetected if an empty string was supplied
- Fix a bug fetching primary key column names for tables
- Fix a bug building database structures on MySQL 3.x servers
|
|
|
|
|
|
|
|
| |
- Add a ping keepalive managing object to prevent retain cycles from the NSTimer
- Add -[SPMySQLConnection copy] support
- Refactor Hans-Jörg Bibiko's database structure retrieval, moving it out of the MySQL framework and building it around a copy of the connection. This reduces the amount of connections-over-time used by Sequel Pro to two constant connections (addressing Issue #1097) and improves robustness.
- Use the database structure retrieval connection for faster query cancellation without an extra connection required, if possible
|
|
|
|
|
|
|
| |
- Correctly record affected rows
- Fix thread safety/autorelease issues when draining pools during fast iteration
- Improve streaming result processing speed
|
|
|
|
|
|
|
|
|
|
|
|
| |
exceptions as database structure retrieval is currently missing!
Further work on SPMySQLFramework integration:
- Improve SPMySQL framework build settings including correct ppc builds and a Distribution configuration for the build distributions to match
- Add new convenience querying and result methods to the framework
- Amend Sequel Pro source to use the new SPMySQL.framework methods everywhere, replacing MCPKit methods where they differ and improving some functions
- Remove MCPKit from the source
- Fix a number of warnings on Release-style builds
|
|
|
|
|
| |
- Delete incorrectly committed build and pref files
|
|
and ready for use but not yet integrated. This new framework should provide much of the functionality required from MCPKit and is based around its interface for relatively easy integration. The largest missing component is Hans' structure code which I believe is better placed outside the framework.
From the Readme file:
The SPMySQL Framework is intended to provide a stable MySQL connection framework, with the ability to run text-based queries and rapidly retrieve result sets with conversion from MySQL data types to Cocoa objects.
SPMySQL.framework has an interface loosely based around that provided by MCPKit by Serge Cohen and Bertrand Mansion (http://mysql-cocoa.sourceforge.net/), and in particular the heavily modified Sequel Pro version (http://www.sequelpro.com/). It is a full rewrite of the original framework, although it includes code from patches implementing the following Sequel Pro functionality, largely contributed by Hans-Jörg Bibiko, Stuart Connolly, Jakob Egger, and Rowan Beentje:
- Connection locking (Jakob et al)
- Ping & keepalive (Rowan et al)
- Query cancellation (Rowan et al)
- Delegate setup (Stuart et al)
- SSL support (Rowan et al)
- Connection checking (Rowan et al)
- Version state (Stuart et al)
- Maximum packet size control (Hans et al)
- Result multithreading and streaming (Rowan et al)
- Improved encoding support & switching (Rowan et al)
- Database structure; moved to inside the app (Hans et al)
- Query reattempts and error-handling approach (Rowan et al)
- Geometry result class (Hans et al)
- Connection proxy (Stuart et al)
|