Age | Commit message (Collapse) | Author |
|
|
|
|
|
|
|
This is causing problems such as duplicate 904 messages.
|
|
This will not cause any compatibility issues as we do not rely on
this attribute internally.
|
|
|
|
* Send to remote servers upon setting
* Send to halfop if available upon expiry
|
|
2.0: Inform the client when a SASL message cannot be sent
|
|
This replaces the devoice module which has now been removed. If you
want users to be able to devoice themselves then you can load the
customprefix module add the following config tag:
<customprefix name="voice" depriv="yes">
If you wish to keep identical behaviour rather than allowing users
to use "MODE #YourChannel -v TheirNick" then you can load the alias
module and add the following config tag:
<alias text="DEVOICE" format="#*" replace="MODE $2 -v $nick">
|
|
- Rework the code flow to be more readable.
- Ensure that the name is well formed.
- Allow prefix modes with a rank of zero.
|
|
|
|
|
|
When SASL is properly configured with a 'target' server, we are able
to inform the client when the message fails to send.
Currently if a target is configured and it is offline, no response is
sent. This can cause some clients to time out while waiting for a response.
If a target isn't configured, behaviour will not change with this commit.
The default of '*' will still send to all servers.
Updated example config with 'target' variable.
|
|
Previously we used FOREACH_MOD to call OnRehash which handled any
thrown exceptions. When we switched to ReadConfig this stopped
being the case.
This bug was introduced in c202dea024.
|
|
This was done to the string variant in d3b82b2f7f.
|
|
|
|
- Use the same dhparams name in m_ssl_openssl as in other modules.
- Fix building m_ssl_mbedtls.
|
|
This removes the need to do lookups in the cgiirc module. This is
useful as relying on gateways to do proper DNS checks is unreliable
and has resulted in issues like 5fc4403f62. Its more sane if we do
our own lookups.
This change has been okayed by multiple WEBIRC gateway authors so
I don't think it will cause many problems.
|
|
|
|
This removes the need to invalidate the cache after changing a
user's hostname.
|
|
|
|
See: 66f82ccf926aac39273bfc652c85c08080cc9a46
Fixes inspircd/inspircd-extras#110
|
|
|
|
|
|
|
|
After discussing this with Adam we have decided that there is very
little point in keeping this now that Solaris has been killed. Any
legacy Solaris users will have to make do with poll.
This change does not affect users of Solaris forks like Illumos as
they have support for epoll which will have already been selected
over Solaris IOCP by configure.
|
|
|
|
Using unique letters for each flag makes it easier to read when
using fonts which have similar glyphs for upper and lower case
characters.
|
|
|
|
|
|
|
|
The latter is more suited to the things we have previously been
using std::stringstream for.
|
|
|
|
- Switch to using ExtensionItem::ExtensibleType for the type instead
of TargetTypeFlags.
- Pass the extensible to OnCleanup as an Extensible pointer
instead of a void pointer.
- Call OnCleanup for memberships as well as channels and users.
- Rewrite event documentation to remove outdated references.
|
|
This allows us to do user friendly things like loading the alias
module in the example alias files.
|
|
|
|
- Read the disabled command list in ApplyDisabledCommands() instead
of storing it in a global which is only accessed on rehash.
- Write debug messages to the log when disabling commands.
- Use irc::spacesepstream instead of std::stringstream.
|
|
These being constant causes problems on older compilers that don't
elide the temporary copy when adding to the host list.
|
|
|
|
|
|
|
|
This mirrors a similar feature in ircd-hybrid but using a different
numeric that is not already taken in InspIRCd.
|
|
- Fix the extremely broken indentation.
- Simplify the HandleLocal code flow.
- Improve the notices sent to operators.
|
|
|
|
- Store ident hosts separately to WebIRC hosts.
- Throw exceptions when the configuration is invalid.
|
|
When we release 3.0 these links will break as they will point to
the pages for 3.0 rather than 2.0.
|
|
- INSPIRCD_SOCKETENGINE_NAME is not really something that needs to
be user facing. If opers want to know this kind of internal info
then they can look at the build configuration.
- INSPIRCD_SYSTEM causes problems for reproducible builds and is
only accurate for the state of the system when InspIRCd is built
which is not useful in the slightest.
|
|
We don't use C strings for storing the ident anymore.
|
|
|
|
|