Age | Commit message (Collapse) | Author |
|
Issue #882
|
|
timing-safe way
|
|
successful auth
|
|
Checking the password before the fingerprint means that even without the
correct cert it's possible to brute force the password or leak
information about it. Checking the fingerprint means attackers must
forge the cert before they can learn any information about the password.
|
|
|
|
|
|
This patch has made me extremely grateful for git-blame as the
GnuTLS developers apparently do not document when deprecation
warnings are introduced.
|
|
|
|
Missing doc added by @attilamolnar
|
|
m_httpd also now uses the correct timestamp format.
Windows-specific fixes by @attilamolnar, original PR #849
|
|
|
|
|
|
pointlessly
|
|
|
|
document what it does
|
|
|
|
containing it
|
|
|
|
|
|
containing it using fakederef
|
|
containing it using fakederef
|
|
ModuleManager is constructed when ServerInstance is non-NULL, test that instead
|
|
containing it using fakederef
|
|
containing it using fakederef
|
|
containing it using fakederef
|
|
|
|
|
|
/USERHOST
|
|
|
|
|
|
|
|
Change append(" ") to push_back(' ')
|
|
|
|
Deduplication makes us do pointless processing for every ISON and offers
no benefit at all to proper clients that send a nick only once
|
|
|
|
variable
|
|
once per chan
|
|
|
|
|
|
|
|
to LocalUser
No remote users were passed to this hook before.
Remove needless IS_LOCAL() checks.
|
|
module denies a nick change
|
|
|
|
server on rehash
|
|
|
|
protocol servers
|
|
VERSION
|
|
Issue #637, suggested by @CuleX
|
|
|
|
|