The GNU_SOURCE build of this function was (sometimes?) just returning the
string and not actually copying it to the passed buf. This is consistent
with the GNU manpage on strerror_r():
"This may be either a pointer to a string that the function stores in buf ..."
but was inconsistent with all the other build types of this function.

Also updated to use strlcpy and write the problematic error number to the
default string (if unknown).

As Nelgin pointed out, a Socket.connect() failure for reasons of address/host
lookup failure would report a stale/nonsense Socket.error/error_str value.

This change required us to query/store the socket API/getaddrinfo error string
at the time of failure (for the Socket.error_str property value) rather than
converting from number to string at the time the property is read.

This does mean that sometimes Socket.error is a errno value and sometimes its
a getaddrinfo (EAI_*) error value. Since the EAI_* values are negative, it
should be obvious which is which.

Reads the new(ish) node*/terminal.ini file when it's changed to determine
the connected-client's terminal type/dimensions. Now reports the terminal
details in the spy form title/caption and if it's UTF-8, does a real-time
translation to CP437 (since F. Piette's terminal lib only supports CP437).
There is a bug here where if the data read from the RingBuffer is a
partial UTF-8 sequence, it isn't decoded correctly. Fix that later (?) if it
actually bugs anyone.

There's no PETSCII conversion/support added here. I expect to someday replace
this entirely (likely with something that uses cterm), so didn't want to
invest too much time into this.

But at least now spying on non-traditional ANSI-BBS clients isn't so terrible
to look at.

These are regarding initializing variables at the same time as the
mutex.

Harmless.

Added MSP, MXP, MSDP and a few other protocols

See merge request !412

readHeaderFixed() already assumes the same header pointer will be
passed in next call if the read is incomplete.

If we don't get one of the commands we're waiting for, close the
socket and exit instead of returning with the appearance of having
received one.

Change wait_for() to no longer loop indefinitely.

If nick is rejected (e.g. begins with number), prepend '_' to it.

Increment version number to 1.62

Fixes issue #723

What I haven't really worked out is what if the translated word mnemonic is
a duplicate of another, valid command key.

We've patched this thing enough to have a different identifier.

By default, Cryptlib uses the type of whatever channel happens to
be currently selected, so if the server has sent some data, that
will be shell.  Instead, look up the new channel and use that type
info.

Ugh.

Previously, it was read into a local variable, and in the case of a
partial read, everything would go to hell and SSH would hang.

"Luckily" this was very hard to trigger.

Also, fix an unlikely race condition.  This, combined with the
cryptlib patches, fixes various weird SSH hangs resulting from
partial reads of the SSH packet headers... a rare event that's
very hard to trigger.  Special thanks to MeaTLoTioN for reporting
this, and running a BBS that manages to trigger it reliably.

Check out The Quantum Wormhole and say thanks!

This should work exactly the same, but is more correct.  It's possible
that in the future, packet data could be added to the buffer after
receiveBufPos, and we could know that no more is required.

Fixes Coverity issue.

Since old is a strdup() of list->comment, it's known that it will
fit, but Coverity doesn't know that.

Whill this is perfectly legal and is not a problem, Coverity complains.

This will hopefully keep both Coverity and compilers happy.

The option from the menu (ALT-Z/CTRL-S) was copy/pasted years ago
and never updated while the keyboard shortcut (ALT-L) was updated
quite a bit.

This extracts the ALT-L code into a function and calls it from the
menu.

This should fix the recent issue when a larger or smaller version
is painted inside the window after resize, as well as fix ALT-→
when the current size is not an integer multiple.

This is my happy place on the desktop system running xfce4... any
higher, and it interferes with how I manage my windows.

Hopefully, this value convers the initialization portion perfectly
on all systems (hah!).

Instead of strictly using "first resize", also count flushes and
assume the hack isn't needed after five.  Experimentation on my
ChromeBook shows 3 to be sufficient, so I'm sing five for now.

Odds are that this will still mess with tiling WMs, but I don't use
one, so I don't mind yet.  I expect this will be better, but likely
still not great.

Fixes issue with possible corrupt display when "weird" scaling
changes occured.

Bugfis:   Be sure to terminate value if returning deflt.
Paranoia: Terminate value if iniGetString() returns NULL

I think the only way to trigger the paranoia is to pass NULL as the
default when reading a key that's not present.

The first time we receive a MapNotify or ConfigureNotify that sets
the width and height, resize the window instead of accepting the
current window size.

Silly to pass a 728-byte object as a parameter.

No more using the standard int type and having it work.
Thanks Microsoft.

Use strtoll() to parse number (except when _BORLANDC_ of course)

OpenFile() returns HFILE and is sus, CreateFile() is more complex,
but returns a HANDLE (and is what MSDN suggests I use).