Tags: vapor/mysql-nio
Tags
Several very long overdue crash fixes (#95) * Update package support files (dockerignore, readme, CI workflows) * Minimal code cleanup - replace `fatalError()`s and `assert()`s with thrown errors, add escapes in comments so Xcode's inline formatting doesn't choke, add `any` to existential usage, other very minor tweaks. * Fix the "server sends error packet when nothing else is going on causes crash" problem - fixes #87 * Add fix for the bad protocol version bug - fixes #91 * Fix the tests bug that causes problems with the `foos` table still existing when the integration tests for FluentMySQLDriver run. * Work around weird deadlock on Linux * Remove many unneeded !s from try!s in tests
Add package minimum versions for all platforms (#83) * ADD package minimum versions for all platforms * Update CI Co-authored-by: elGeekalpha <[email protected]> Co-authored-by: Gwynne Raskind <[email protected]>
Clean up MySQL's copy of some ByteBuffer utilities - make readNullTer… …minatedString() much more efficient, make the write*() methods return the length written as they should, simplified the read and write methods for length-encoded values, removed a couple of unneeded/unused utilities. (#65)
Fix hard crash when describing "zero" datetime (#64) * Fix #63: Hard crash in MySQLData.description when a zero-value datetime is received from older MySQL configurations which still do this. Returns a description of the epoch instead. * Use the thread-safe gmtime_r() instead of gmtime() in MySQLTime.init(date:). * Pass a more sensible capacity value to ByteBufferAllocator for SHA digests. * Fix deprecation warnings for TLSConfiguration and declare the explicit dependency on the update NIOSSL version. * Another round of long-overdue CI updates * Don't crash if server capabilities are not available during connection close.
PreviousNext