Unfortunately, the call to readmail() in other places in sbbs
(e.g. logon.ccp, useredit.cpp) is hard-coded without *out* this loadmail
flag, so you get an index in oldest-first order when you logon and answer
"Yes" to the "Read your mail now?" prompt and get the index in reverse
(newest-first) order when you use the 'R' command from the email prompt
which is unexpected and likely confusing to users. Until we can replace
these hard-coded calls to readmail() with scripted logic which can read/
use a user-preference property (e.g. from data/user/*.ini), we will just
default to the old/previous behavior. The user is one key (!) away from
reversing the index anyway.

function of the same name. The dirwrap version (now) behaves identically
when passed a pathname ending in a directory name (ending in a '/') but
differs when passed a pathname ending in a file name (not ending in a '/').
That should be okay.

same name in uifc/filepick.c - specifically with regards to paths that
begin with a / (or \ on Windows).

Make usage-printed option defaults to the actual defaults and not just the
current values of the options (which could have changed).

specification ("<prop>=<format>"). The "-fmt <format>" option syntax is
still supported.
Fixed a bunch of typos and update the "docs" in the comment header.

lists (listing of areas/directories, not files within those directories),
to export your file area configuration to various file formats (e.g.
FILEGATE.ZXX, tab-delimited, CVS, etc.). This utility is much more flexible
than the SCFG "Export Areas..." option.
This module can also be load()'d to capture the listing output for use within
other scripts or shells.
See the comments at the top of the file for detailed documentation.

fancy structure initalization syntax here.

Removed TCP/IP services not supported by BBS clients (e.g. SyncTERM) from the
list (e.g. ftp, nntp, smtp, binkp, irc) and added "modem".

Removed TCP/IP services not supported by BBS clients (e.g. SyncTERM) from the
list (e.g. ftp, nntp, smtp, binkp, irc) and added "modem".
This list is now used by the syncterm_list() function to filter which entries
will be written to the exported syncterm.lst file.
Also added duplicate section name detection and better names for multiple
services per BBS (using service description and port as well as protocol)
to syncterm_list().
Added support to imsg_capable_system() for an over-ride (imsg_capable).
Changed several info-level log messages to debug-level (e.g. for every list
open/read/write).

- separate sections with a blank line (when adding a section)
- place spaces around '=' in value assignments ("key = value").

Mostly just white-space changes (beautification).
Added some comments.

this idea (web-server parsed html files with embedded JS) was never implemented
and made obsolete by web handlers (e.g. xjs).

anyway (here) and never implemented in the web server.

Oddly, the Identity Lookup checkbox was obscurring (hiding) the Hostname
Lookup checkbox, so that becomes visible/available again.

(not practical to disable JavaScript for the terminal server any longer)
and added the NO_SPY_SOCKETS option.

section of the sockopts.ini (if there is one) rather than the [tcp] or
[udp] sections. This resolves the errors:
      term xxxx !ERROR 95 setting socket option (TCP_NODELAY, 1) to 1
which started appearing after the TCP_NODELAY option was added to the
[tcp] section.

a target port number.
The port and hostname arguments can now be supplied in any order.
Added debug-level log output with the host/port we're trying to connect to.
Include the socket error value in the failure to connect log message.

TCP connection actually failed. This bug only appeared to affect *nix
systems. This bug appears to be very old, introduced in rev 1.74 of
this file (Mar-2003) by yours truly.
From the Linux 'connect' man page:
EINPROGRESS
              The  socket  is  nonblocking  and the connection cannot be i
              completed immediately.  It is possible to select(2) or poll(2)
              for completion by selecting the socket for writing.  After
              select(2) indicates writability, use getsockopt(2) to read the
              SO_ERROR option at level SOL_SOCKET to determine whether
              connect() completed successfully (SO_ERROR is zero) or
              unsuccessfully (SO_ERROR is one of the usual error codes listed
              here, explaining the reason for the failure).

We weren't doing the 'getsockopt(SO_ERROR)' part.

Wrap all read_*_cfg() and write_*_cfg() calls with simple functions that
display popup messages that the Reading or Writing is being performed.
There was already copy/pasta everywhere for the Reading ... cnf files, but
nothing for writing.

pathname (directories must end in a '/'). Unlike getfname() which returns an
empty string if passed a path ending in a '/'.
Useful with glob(.., GLOB_MARK).

getfname() on a path that ends in a slash, returns an empty string pointer.
Fixed-up some logging output around listing creations.

- use GLOB_MARK to eliminate calls to isdir() in loops
- use stat() rather than calls to flength() and fdate()
- don't call getfiledat() if getfileixb() failed
- profiling info-level log msgs added for index/listing generation, e.g.
  "detailed listing (3459 bytes) of /main/SBBS (45 files) created in 2 seconds"

Fixed bug with indexes generated for directories with the
"Access Files Not In Database" option enabled, files not in database were
excluded from the generated index (e.g. 00index, 00index.html).

rather than calling both flength() and fdate() (which both call stat()), just
call stat() once and use the result for both file size and date/time.

There's no reason to call access() before stat() since stat() will fail
(return -1) if the file does not exist.

Loads child scripts (finger.js, sbbsimsg.js) quicker.
Note: irc.js can't be load()ed from here because it calls exit().
Now using appropriate strings from text.dat (R_Chat, ON, OFF).
Added modopts.ini [chat_sec] irc option (disable IRC access by setting to false).

console.ctrlkey_passthru state.
As it was, if the irc.js abended, many ctrl-keys (e.g. ^C) would not work
after exiting. A lot of scripts that much with the console.ctrlkey_passthru
have this/similar issue and this is a simple but very reliable fix.

(mostly migrated from exec/finger.js).

tickit.js version pulled from $ID string in .js file.
tickit version used in Created by line.
tickit version used in Path line.
utc time stamp of processing used in Path line per FTS-5006.001.
per area uploader.
address selection:
   per area aka matching in From and Path lines.
   per area source address in From and Path lines.
   global aka matching in From and Path lines (original addressing method).
   global source address in From and Path lines.
   main FTN address from system.fido_addr_list (preferred but last in line with
the above overrides).
can use domains on linked systems' addresses in sbbsecho.ini again.
can use domains on any address in tickit.ini.
domains are NOT currently written to the generated TIC files to avoid creating
possible problems for other TIC processors.
force replace is available globally and per area in the ini and not only the
command line.

the beginning of a SAUCE record and the end of very old MS-DOS text files.
Added support for ESC[1;1f/H (home cursor) and ESC[0J (clear to EOS) sequences
(converted to the equivalent in Ctrl-A codes).
Output Ctrl-A characters in uppercase, as they were documented and as is
common practice.
Covert ESC[xB and ESC[xD to the appropriate Ctrl-A sequences rather than
converting to ASCII ctrl chars (CR, LF, BS).

Ctrl-AJ (clear to EOS)
Ctrl-A` (home cursor)

Oct-14-2018:
It turns out, PabloDraw actually inserts a Ctrl-AZ sequence at the end of .msg
(and presumably Synchronet .asc) files it edits - before the SAUCE record.
This resulted in a printed Ctrl-Z character (arrow pointing right) in most
terminals when viewing text/menu files created or edited with PabloDraw. :-(
So, now Ctrl-AZ (uppercase) will revert to the previous definition:
premature end-of-file (EOF)
and a Ctrl-Az (lowercase) will output a Ctrl-Z (substitute) character.
I'm not a big fan of case-sensitive Ctrl-A codes, but frankly, running out of
chars and I already started this pattern with the Ctrl-AF/f sequences.
Hopefully there's no existing software that is/was putting Ctrl-Az (lowercase)
in files, expecting that to trigger a premature EOF. I certainly was not.