GCC and Coverity (x 2) both flagged this one.

Also, change the default smb_t.retry_delay from 250 to 100ms

Getting a lot of smb_locksmbhdr() failures with sbbs-Linux accessing message
bases (i.e. SYNCPROG in particular) on Samba share.

Similar to what I do with user.tab record locks 2 months ago, this will
extend the total shared/network file open attempt duration from about 5
seconds to about 45 seconds, but with incrementing back-off.

Trying to sbbs-Linux to run from a Samba share without file open/locking
errors.

Re-synced the sbbs_t version of nopen with the nopen.c version (wasn't
special-handling EDEADLOCK failures).

Mostly we're using fnopen() these days anyway, which doesn't have an sbbs_t
version (for logging warning messages about collisions), perhaps it should.

This date (basically 0) means either the dsts.ini file doesn't exist or there
was an issue reading it successfully (that was not detected by getstats()).

If this works, we can use it for EVERYTHING!
Baw ha ha!

Use sdl-config before falling back to the framework.
Move sdjclean into objodir and link with mt stuff.

Also fix last commit

When using MacPorts (and likely Homebrew), this isn't how this works.

New versions still support that.

This was a regression: Socket.recvline() used to not care what the timeout
duration was so long as there were bytes to receive.

Also, remove the ".0" from timeout values in documented mehtods that don't
(any longer) accept floating point timeout durations. We used to support
fractional seconds for some of these methods, and that was implied by using
the floating point default values, but that's no longer the case. poll()
still accepts a floating point timeout.

This was just a typo which didn't allow the single-file upload xfer prot menu
to be shown.

Actual fix for CID 11447 (previous attempt, commit f50c866f was
insufficient). Thank you Coverity!

Otherwise, zero it out.  Most termios implementations don't put the
speed in here, but this *is* where dragons be.

As noticed while trouble-shooting issue #813, calling comOpen() would
(on Linux, at least) set the port baud rate to 0 bps (B0) which in
most or all Linux serial drivers triggers special logic to deassert DTR and
RTS signals (to the modem, to attempt to "hangup" any connection).

If the app (e.g. sexpots) did not explictily set the port baud rate after
calling comOpen(), the port would be unusable. This is not how comOpen() works
on Windows.

So rather than just overwrite all the bits in termios.c_cflag, we clear the
fields we know we want to, set the bits we want, and leave the rest (which
usually includes the current baud rate, if CBAUD is defined) as-is.

Gracefully terminate with SIGQUIT, SIGINT, and SIGTERM

Ignore SIGHUP and SIGPIPE.

SDL_SetWindowSize() takes dimensions in imaginary units it makes up
so it can call something "Retina™".  SyncTERM however lovingly
crafts everything pixel by pixel for an atisanal ode to BBSing.

This change converts pixel sizes to Retina™ sizes before resizing
the window.  Should fix ticket 164.

I can't believe I missed this - this would've caused definite issues with file
transfers.

I saw a discussion in FSX MYS (between apam and GEARBOX)  which clued me into
their being an issue here. I do wish users of SexPOTS would report issues to
me. :-(

Incremented version to v2.1

Fix for issue #809

(ie: gcc 4.4)

More effort to fix the macOS issue.  This could potentially result
in slight problens with the rightmost pixel, but will protect
against any weird FP issues and -ffast-math concerns.

A few extra bytes really helps with human readability.

True, a sysop could add/configure a monthly timed event instead, but that
requires more configuration. Fixed events are nice in their simplicity.

This is intended for running the new Synchronet trashman utility, but could be
used for anything that's simple to execute and the sysop:
- doesn't care which node/instance of sbbs it runs from
- doesn't care exactly what time the event runs
- doesn't need the assorted options available for timed events

Remove undefined behaviour (right-shifting a negative integer) from
the scaler... replace with divide-by-2.  Any modern compiler should
be able to optimize that.

On the enabling the compiler front, const-ify more variables and
use types that allow removing some casts.

I have a bit more hope this will fix the macOS issue than the previous
efforts.

Not sure what macOS is doing, but it looks like something weird is
happening in interpolate_width... maybe some unexpected integer
promotion?

* Allow left and right arrow keys to move between lines (within reason)
* Use the K_USEOFFSET getstr() mode flag to keep cursor position when moving
  between lines with arrow keys
* Be smart about integer padding when /Listing lines with numbers
* Add range checking (!) and better error reporting for /Lx argument value

Fixes a couple CIDs and a GCC warning

for getstr() to be able to return when user right-arrows off end of string
(for use in msgeditor).

CCl4 sends an ENQ character in the login screen, and sends an ENQ
which will auto-fill your username.  Also, if you enter six zeros
into the username, it will take you to the password field.

With ten zeros, this means it's impossible to log in to CCl4.

Fixes Night Owl

And have it return bool, like the others. The error logging is different
(not using errormsg), but I think that's okay.