aboutsummaryrefslogtreecommitdiffstats
path: root/Frameworks/SPMySQLFramework
Commit message (Collapse)AuthorAgeFilesLines
* Swap two lines of code so that if a connection attempt is cancelled, SP does ↵Max2017-11-021-2/+2
| | | | not try to open another connection to cancel the pending connection (#2909)
* Because of a change in CMake the updated libmysqlclient was only runnable on ↵Max Lohrmann2017-05-222-2/+8
| | | | | | 10.12+, while 10.6+ was intended (Also adapted the build script)
* SPMySQL will now enforce SSL connections when requested (#2499)Max Lohrmann2017-05-211-0/+19
| | | | Previously we could only *request* SSL, but if the server didn’t support it, libmysqlclient would go ahead anyway. This is fixed in MySQL 5.5.55.
* Update MySQL client to 5.5.56 (was 5.5.42)Max Lohrmann2017-05-216-5/+41
|
* Move a struct from heap to stack, since the caller will outlive its callees ↵Max2017-05-141-8/+9
| | | | anyway
* Expand a lock in the connection keepalive code, because it still allowed a ↵Max2017-05-111-2/+19
| | | | zombie call
* Fix some issues with charset handling during connectMax2017-05-024-24/+81
| | | | | | | * hostname and file paths are places where libmysqlclient ultimately only passes the string to an OS library so we should use whatever charset the OS expects, not mysql * Even though _makeRawMySQLConnectionWithEncoding:isMasterConnection: takes an explicit charset argument most of the conversion logic simply used whatever charset the existing connection currently has, which is not neccesarily the one the new connection should use * Add some remarks about the charset handling with passwords and mysql_error() * Charsets do not apply to sqlstate.
* Silly README file, you can’t be a source code fileMax2017-04-011-2/+0
|
* Fix file links.Stuart Connolly2017-03-261-2/+8
|
* Fix a race condition in debug code that could lead to accessing a freed objectMax2017-03-101-3/+7
|
* * Turn `-[SPMySQLConnection checkConnectionIfNecessary]` into a public ↵Max2017-03-095-34/+36
| | | | | | method, since it is actually the preferable way to `-[SPMySQLConnection checkConnection]` (which forces network IO in a new thread each time) * Change `-[SPDatabaseStructure queryDbStructureWithUserInfo:]` doing a forced connection check inside a tight loop (#2306)
* Add a bit of debug info in Connection pingingMax2017-03-092-3/+8
|
* Workaround for an exception when connecting to MySQL 8.0 (#2699)Max2017-02-231-0/+1
|
* Change the way max_allowed_packet detection worksMax Lohrmann2017-01-252-24/+43
| | | | Sequel Pro will now simply go with whatever query returns a valid result instead of trying to guess the correct query from the server version number
* Add license file and switch to readme to Markdown.Stuart Connolly2017-01-143-69/+69
|
* Make Max Packet Size detection more robust (issue with recent sphinx ↵Max2017-01-101-10/+11
| | | | versions) #2653
* Add possible socket locations for MariaDB via MacPorts.Sebastian Volland2016-08-301-3/+5
| | | | Also rename Darwinports to MacPorts in comments.
* Fix one erroneous method callMax2016-07-211-1/+1
| | | | Not sure if that may has been causing issues for some users...
* Added shared scheme for SPMySQL FrameworkMarius Ursache2016-02-291-0/+99
|
* Updated SPMySQL Framework SettingsMarius Ursache2016-02-293-3/+12
|
* * Change a while() to do while; to help code analysisMax2016-02-182-4/+6
| | | | * Add log message (possibly related to #2258)
* Add another fallback for charset detection (which might improve ↵Max2016-02-061-1/+18
| | | | compatibility with Sphinx)
* Fix a rare crash due to calling a NULL method pointer (#2390)Max2016-02-011-4/+4
|
* Try to fix what could be the cause of #2353 ("Attempted to connect a ↵Max2016-01-154-1/+24
| | | | connection that is not disconnected")
* Add more debug info for "Attempted to connect a connection that is not ↵Max2016-01-132-3/+26
| | | | disconnected" crash (#2353)
* Another attempt at the 0x338 crash (#2258)Max2016-01-131-1/+1
|
* Fix wrong charset mapping for "latin1"Max2015-11-191-3/+3
|
* Add basic support for JSON type (part of #2199)Max2015-11-155-2/+12
|
* In order to enable even more people to shoot themselves in the foot, this ↵Max2015-11-122-0/+24
| | | | | | commit adds support for remote server shutdown 💣
* * Basic math is hard sometimes (fixes an issue introduced by me in ↵Max2015-10-311-1/+1
| | | | | | | b2d798ba9282d3acf1a2d65de30849e529d4d255) * Fix an exception that could occur when trying to view a damaged table * Fix a theoretical use-after-free issue by a wrongly structured retain/release in a setter
* * Add a wrapper method for 3 repeatetly used callsMax2015-10-305-12/+17
|
* * Lock connection during disconnect to prevent some race conditionsMax2015-10-304-59/+41
| | | | | * Always use the server version name provided by mysql_get_server_info() as that should me more reliable * Use mysql_get_server_version() for version comparisons. Less code, official API and closer to what mysql does
* * Remove some ivars that are only used as local varsMax2015-10-264-11/+9
| | | | * Also I take back what I said about 4d97cbd4df1ebb451d89d6c4e12dd7c622d00b84 and say it will happen now *duckundweg*
* This change will either reduce the amount of crashes or increase the amount ↵Max2015-10-171-0/+3
| | | | of beachballing. In the latter case please file an issue.
* Minor changeMax2015-10-173-17/+21
|
* Move a mysql_affected_rows() to see how the 0x338 crash behaves (#2258)Max2015-10-081-1/+7
|
* Add a few more unit tests to SPMySQLMax2015-10-043-5/+98
|
* Configure SPMySQL project to actually be able to run Unit TestsMax2015-10-043-6/+320
| | | | | | | | * Add a new configuration "Unit Tests" * Add a new Target for the Unit Test code * Assign Unit Test Target to the Test step of SPMySQL scheme * Also remove a compiler setting for PowerMac G5
* Remove a duplicate @interface and prepare some functions for unit testingMax2015-10-044-25/+19
|
* Add logging for a possible issue in Connection keepaliveMax2015-10-041-0/+4
|
* Add some debug code for another crash that has been happening latelyMax2015-10-021-1/+1
|
* Fix conversion of BIT fields (fixes #2254)Max2015-10-022-14/+84
|
* Remove a bit duplicate codeMax2015-09-301-5/+4
|
* Clarify thread namesMax2015-09-291-1/+3
|
* Added fallback code for server version identification.Max2015-08-074-17/+29
| | | | | Good news: Sequel Pro can now connect to sphinxql server. Bad news: It will treat it like a MySQL 2 server (or whatever version of Sphinx is running)
* * Expand the debugging code from prev. commitMax2015-08-061-0/+7
| | | | * Also Sequel Pro will now crash on the "Attempted to connect a connection that is not disconnected" issue. That is sadly necessary to get a stack trace of ALL threads as for exceptions Feedback Reporter only gives us the current thread.
* Add code to disable mysql protocol compression (no UI) to connect to Amazon ↵Max2015-07-304-5/+39
| | | | Aurora (see #2122)
* Fix an issue where Sequel Pro added a NUL byte to the end of every query ↵Max2015-07-292-9/+12
| | | | | | (fixes #2184) We tried to convert the query string into a c string that could contain NUL bytes - which by definition a c string cannot (making it a byte buffer with a terminating NUL byte) and then tried to pass that to mysql_real_query() which expects a byte buffer anyway.
* that also belongs to the previous commit...Max2015-07-141-0/+2
|
* Fix an exception caused by empty column names in query results (fixes #2173)Max2015-07-141-5/+10
| | | | Issue likely introduced in 2b52f76e.