Age | Commit message (Collapse) | Author |
|
|
|
|
|
|
|
Fix various compiler warnings
|
|
|
|
|
|
|
|
|
|
|
|
Issue #1006
|
|
NO_CLIENT_LONG_LONG was removed in a recent version of MySQL.
|
|
|
|
|
|
|
|
|
|
|
|
certificates
|
|
module
Use it to simplify logic in all modules using or providing IOHooks
|
|
|
|
Fixes issue reported by @m4rkw on IRC
|
|
remains in the buffer of the SSL library
|
|
|
|
m_sslinfo
|
|
|
|
|
|
|
|
The previous method was causing build errors on CentOS and Solaris.
|
|
|
|
|
|
preferences by default
|
|
configured to do so or when using the default
|
|
|
|
Generating DH params like this blocked the main thread, significantly slowing down startup and it made /rehash ssl freeze the server until the generation finished. External tools such as our genssl should be used to generate DH params.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
disallow it by default
|
|
|
|
copies done by sendq flattening
|
|
Do sendq flattening in SSL modules, move code for it into class SSLIOHook from core
|
|
|
|
Let OnStreamSocketWrite see the entire sendq instead of one element at a time
|
|
Redesign to be more like m_mysql, polling ldap_result from a thread as
done previously is undefined.
|
|
|
|
completed
|
|
Remove enum EventType
|
|
Setting a vector push function is recommended to reduce overhead during handshake
|