This dict was filled with IrcString keys, which is hashed
as lowercase, so when queried with a non-lowercase string,
the key would not be found, and lead to very confusing errors.
If throttle time is set to 0.0 (e.g. in tests) and the reported time
since last dequeue is also 0.0 (maybe because the timer resolution is
too low), takeMsg would throttle. This would make tests fail because
they expect messages immediately.
This issue was observed on Windows 10 20H2 with both a mingw-w64 Python
v3.8.9 from MSYS2 and several version of official Python v3.x.
Confirmed at least as far back as
aa98d987a7.
Since it's a non-integer comparison, the change shouldn't have any
observable effect in normal operation.
ChannelPluginTestCase#setUp accepts the same parameters as its
super-method PluginTestCase#setUp, but it didn't pass the arguments down
the line. Now it does.
by moving the 'smart' reply() method from NestedCommandsIrcProxy
to ReplyIrcProxy.
There is no reason only commands should have a paginated output
and not snarfers defined in PluginRegexp.
1. it binds the port, which is useless while documenting
2. in some circumstances, it prevents supybot-plugin-doc from exiting
(because the server threads are still running)
It's pointless and looks stupid.
It will look even more stupid when we enable multiline, because the
suffixes will be in the middle of the concatenated message.
Otherwise, if some IP addresses don't work (eg. all odd ones), the bot will
consecutively fail because it can't connect, then connect + get STS + reconnect,
then fail again, then connect + get STS, etc.
Fixes a regression in ecc2c32950 that caused
Socket.py to ignore the IP address entirely after computing it, and
to call getSocket() and connect() with the hostname instead.
FAIL: testMoreIsCaseInsensitive (Misc.test.MiscTestCase)
----------------------------------------------------------------------
Traceback (most recent call last):
File /home/travis/virtualenv/python3.8-dev/lib/python3.8/site-packages/supybot/test.py, line 214, in runTest
originalRunTest()
File ./plugins/Misc/test.py, line 260, in testMoreIsCaseInsensitive
self.assertNotError('more %s' % nick.upper())
File /home/travis/virtualenv/python3.8-dev/lib/python3.8/site-packages/supybot/test.py, line 355, in assertNotError
m = self._feedMsg(query, **kwargs)
File /home/travis/virtualenv/python3.8-dev/lib/python3.8/site-packages/supybot/test.py, line 526, in _feedMsg
response = self.irc.takeMsg()
File /home/travis/virtualenv/python3.8-dev/lib/python3.8/site-packages/supybot/log.py, line 368, in m
return f(self, *args, **kwargs)
File /home/travis/virtualenv/python3.8-dev/lib/python3.8/site-packages/supybot/irclib.py, line 1308, in takeMsg
assert not msg.tagged('emulatedEcho')
AssertionError
I don't understand why it's happening or why it's only that specific test,
but there we go.
Plugins implementing inFilter do not set receivedAt, so incoming messages
rewritten by inFilter would be detected as echo messages, and ignored
by most plugins.
Even if they are not trusted. This fixes a regression in
97016b9c55.
This happens because 'user._checkCapability' raises a KeyError
when the user has neither this cap or the anticap; which was mistakenly
caught here by the 'except KeyError' expecting to catch non-existing
users.
(And that why 'try' blocks should be limited to as few lines as possible.)
This reverts commit 1fe414764c.
I initially wrote that commit so that plugins with echoMessage=True can
get the nick/prefix from the message (eg. the SkypeRelay plugin in my
personal repo).
Unfortunately, this breaks any test that does equality comparison on
between the result of getMsg()/takeMsg() and an IrcMsg object they
crafted themselves.
Additionally, the filled in nick and prefix might be inaccurate if the
bot changes nick/host at the same time. So instead plugins should check
for the None value.
(Also, editing IrcMsgs on the fly like this isn't great, and the commit
forgot to reset the `_str` attribute.)
This is not a regression; this was already forbidden before
23417b0675, and this commit was not
tagged/released yet at the moment I'm writing this one.
So far Limnoria relied on detecting 'ERROR :closing link' (see doError
in src/irclib.py), but that's not a standard at all, and fails on
Oragono; so we need to do this to check we're disconnected.
Plus, parsing the argument of ERROR is awful in the first place.
setTimeout may be called as a supybot.drivers.poll callback,
which may by the access to supybot.drivers.poll() in _select;
so a crash in setTimeout will propage up to _run(), which would
cause a random driver to be killed because another one failed
and that's bad.
For example:
INFO 2020-05-27T18:40:18 supybot Received SIGHUP, reloading configuration.
ERROR 2020-05-27T18:40:19 supybot Uncaught exception in in drivers.run:
Traceback (most recent call last):
File "/usr/lib/python3/dist-packages/supybot/drivers/__init__.py", line 104, in run
driver.run()
File "/usr/lib/python3/dist-packages/supybot/drivers/Socket.py", line 194, in run
self._select()
File "/usr/lib/python3/dist-packages/supybot/drivers/Socket.py", line 167, in _select
[], [], conf.supybot.drivers.poll())
File "/usr/lib/python3/dist-packages/supybot/registry.py", line 422, in __call__
self.set(_cache[self._name])
File "/usr/lib/python3/dist-packages/supybot/registry.py", line 476, in set
self.setValue(float(s))
File "/usr/lib/python3/dist-packages/supybot/registry.py", line 495, in setValue
super(PositiveFloat, self).setValue(v)
File "/usr/lib/python3/dist-packages/supybot/registry.py", line 482, in setValue
super(Float, self).setValue(float(v))
File "/usr/lib/python3/dist-packages/supybot/registry.py", line 385, in setValue
callback(*args, **kwargs)
File "/usr/lib/python3/dist-packages/supybot/drivers/Socket.py", line 305, in setTimeout
self.conn.settimeout(conf.supybot.drivers.poll())
OSError: [Errno 9] Bad file descriptor
ERROR 2020-05-27T18:40:19 supybot Exception id: 0x86ecf
INFO 2020-05-27T18:40:21 supybot Removing driver SocketDriver(Irc object for irchaven).
echo-message ended up in self.state.capabilities_req even though it wasn't
requested, so the bot was stuck in state:
"Waiting for ACK/NAK of capabilities: {'echo-message'}".
prefixes can be rather long, they have rather small cardinality, and
they are repeated a lot in the history; so they take up a sizeable
portion of memory.
And let's also internalize commands and server_tags keys
while I'm at it.
Plugin can opt in to getting echo messages by setting the class attribute
'echo_message = True' if they want to get echos.
This defaults to False in order not to break existing plugins, and because
they usually don't need it (there's outFilter for most cases).
If on an insecure connection: reconnect.
If on a secure connect: store it and do nothing else.
For now, stored STS policies are not read when connecting to an
insecure server.
There's no reason to use it anymore instead of Socket.
It's already missing features compared to Socket, and I don't want to
maintain it anymore so it will keep getting worse.
This change builds all the ircmsg objects directly in irc.reply, with
the original arguments.
A side effect is that if a config var is changed between the initial command
call and the call to '@more', this commit makes it use the old values,
but that shouldn't be too much of an issue.
Closes GH-1405.
I don't see any use for this; and it means a 'channel' converter
called after 'channeldb' returns '#' instead of the actual channel
if it's after 'channeldb' in the arg list and
databases.plugins.channelspecific is False.
The connect() method already adds it, so it was in the list twice
(added both by __init__() and connect()).
This caused _select() to call _read() twice on the same instance,
except there is usually nothing to read on the second call,
so it blocks for up to conf.supybot.drivers.poll().
