* add riscv64 release

* undo Alpine upgrade

* exclude *bsd/riscv64 releases

---------

Co-authored-by: Meng Zhuo <mzh@golangcn.org>

bump version and changelog for v2.14.0-rc2

fix truncation check

* The message target was not being counted :-(
* The additional character added to the target by STATUSMSG was not counted

don't require a config file for defaultconfig

* changelog for v2.14.0-rc1

* bump version string for rc1

* bump irctest

lower recommended ban list limit to 100

Insp and Libera use 100, seems a bit safer

fix #2081

Increase default/recommended mask list size limit to 150;
SAMODE overrides enforcement of the limit.

Don't send RPL_CHANNELMODEIS for no-op changes to channel-user modes

fix #2043

Add human-readable description parameters to multiline fail messages,
since they are technically required by the standard-replies spec
(although the utility of showing them to users is dubious)

fix #2141

Log errors from attempting to delete a unix domain socket path

fix #2157

fix #2099

remove draft/bearer in favor of IRCV3BEARER

(Haven't decided what to do at the spec level yet)

Embed a copy of the ergo default config in the binary;
add `ergo defaultconfig` to print it to stdout

Add optional support for rfc1459 and rfc1459-strict casemappings

fix login throttle handling

We were checking the login throttle at the beginning of every SASL
conversation. This had several problems:

1. Pidgin (on Windows?) tries every mechanism in order, regardless of
the CAP advertisement. It would use up the default throttle allowance
trying unsupported mechanisms like CRAM-MD5.
2. The throttle was actually checked twice for AUTHENTICATE PLAIN
(once at the start of the conversation and once in AuthenticateByPassphrase).

The general pattern here is that we should check the throttle every time we
do something "expensive" (bcrypt verification, send a reset email) or
"dangerous" (anything that could lead to a bruteforce attack on passwords).
Therefore, delete the check from the AUTHENTICATE handler, and add one at
the beginning of the SCRAM conversation to replace it.

fix deadlock on channel state mutex

add i2pd b32 address directions

Fixes #1686

fix #2144

fix #2142

add a config switch to accept hostnames from WEBIRC

RPL_NAMREPLY should send = for normal channels and @ for secret channels,
as per Modern docs.

See #1686; this allows i2pd to pass the i2p address to Ergo, which may be
useful for moderation under some circumstances.

Allow specifying TCP4 or TCP6 for outgoing email sending, or choosing a
specific local address to send from.

fix #2139

Database backup filenames contained a colon character, which is disallowed
on Windows; use period instead

fix #2135

* fix #2133

Don't record NICK and QUIT in history for invisible auditorium members

Handling of reserved nicknames is special-cased due to #1594, but we want to send
ERR_NICKNAMEINUSE if the nickname is actually in use, since that doesn't pose any
client compatibility problems.

* fix #2129

Don't print the values of environment variable overrides, just the keys

* fix unit tests

fix #2130

We load registered channels unconditionally; reloading them again on rehash
is incorrect. This caused buggy behavior when channel registration was
disabled in the config, but some registered channels were already loaded.

* Gracefully handle NS cert add myself <fp>

A non-operator with the nick "mynick" attempts to register
a fingerprint to their authenticated account.

They /msg NickServ cert add mynick <fingerprint>

NickServ responds with "Insufficient privileges" because
they've accidentally invoked the operator syntax (to action
other accounts).

This patch allows the user to add the fingerprint if the client's
account is identical to the target account.

Signed-off-by: Matt Hamilton <m@tthamilton.com>

* Update nickserv.go

Compare the case-normalized target to Account()

---------

Signed-off-by: Matt Hamilton <m@tthamilton.com>
Co-authored-by: Shivaram Lingamneni <slingamn@cs.stanford.edu>