1) Don't reset eob counts for sent or got.
2) Ensure M_GET is never sent in v1.0 mode.
3) Only send one M_EOB in v1.0 mode, even if we add to the list with M_GET

This should cover all the bases, but FREQ may be non-conformant.

in either direction.

For v1.0, a single EOB in both directions terminates the session and for
v1.1, two EOBs in both directions terminates.

This should greatly simplify the overly complicated senteob/sentempty/goteob/
gotempty logic and completely eliminate loops.  Biggest risk with this change
is issues with v1.0 servers.

exists in a .QWK packet. Also, delete the other control/metadata files
case-insensitively (just in case).

from the currently sending file, since the pending ack should always be there,
but sending often won't be.  Also, log a warning when pending_ack isn't empty
and we get an M_EOB.

We can't rely on our send buffers being infinite anymore.

them that way.  Also, ensure there's a timeout on the second byte of the
packet header since we toss out the first byte if the second byte times out
(they should pretty much always be in the same packet, but why take the risk?)

the return value if recv() for non-TLS sockets.

TLS, use MSG_WAITALL for reading integers, add a time() based timeout,
return as soon as any bytes are read (including zero), and generally
behave more closely to how recv() itself behaves.

May resolve the error reported by Bill McGarrity (this.extra_headers is
undefined).

first one.

Radius/4.010/21.01.2005,13:56(Final-Release)/Win32

of poll()ing for each byte.

reason.

to be changed to something the individual clients can frob in the near
future.

bytes have been sent!

"Sometimes" the write timeout value is used for read timeouts.  Since we
use a read timeout of zero, and a mutex, this can cause some serious delays
in SSH processing.  As a workaround, we set the write timeout to zero.
However, a flush failure has historically been fatal.

So, we now change the write timeout to one second just before a flush, and
have a flush timeout not be fatal anymore... we set it back to zero after
the flush.

The fiddling with the write timouts will be removed once the bug is fixed,
but the flush timout not being fatal may stay forever... add some comments
to remind us to remove the hack in the future.

attempt.

"unique" login attempt (e.g. hack attempt).

code merely introduced a 5 second delay in the Node 0 thread.

the challenge length is set and add a TODO comment there.

matches, send an M_ERR to the remote and leave the address list empty.

expect the Seen flags to be coherent across them... support this by leaving
the users IMAP status file open, and holding a lock on the first byte.

We now need to flush changes to the status file more often, and we need
to call file.flush() before unlocking.

nomd5 means "plain-text passwords are okay" (AllowPlainPasswords=true)

The *first* packet from the answering side *must* be an M_NUL "OPT TLS".
If this is the case, the originating size responds with an M_NUL "OPT TLS".
After this, the answering system performs a server TLS handshake, and the
originating system performs a client TLS handshake.

OPT CRYPT is not used in this case (ie: not crypt over TLS)

"user@" prefix.

(import netmail) command along with the '-n' option to set the recipient
net address (and network type). Previously, only Internet or QWK netmail
was supported.

bad login.