"right" way.

by the great "Fix warnings" commit by Deuce. :-)

the "SYSOP_ALIAS" keyword in the sbbsecho.cfg file. e.g. can be set to
"Net Coordinator" with the line "SYSOP_ALIAS Net Coordinator". The default
value is "SYSOP" (case insensitive).

Likely cause of truncated filenames to three chars.

Located by DigitalMan, reported on DoVeNet by someone else.

compiler rules anytime you like) in js_get_size().

1. Scan message areas based on those listed in the areas.bbs, not those listed
in the msgs.cnf file (e.g. configured with SCFG->Message Areas). This allows
the same message base to be scanned more than one for a different set of
uplinks (i.e. gating).
2. Store the message base export pointer (in the data/subs/code.ini file) with
the echotag as part of the key-name. This allows the different upstream
networks to have different export pointers (though after a complete
scan the export pointers should be the same for each msg base).

Increment the version number to 2.20.

Use the 0000p###.su0 point system bundle base-filename scheme for FLO-style
mailers (was previously only used for FD/attach-style mailers).

(calling strlen(NULL)).

I've likely broken something though, so if you see weird behaviour with
extended ASCII, let me know please.

Added backtrace (bt) and up/down commands.
Allow specifying breakpoints in file:line format for future load()s.

Breakpoints are a small memory leak right now as there's no way to clear them.

Running jsexec with the -D argument will enter the debugger.  Inside the
debugger, there are the following commands:
"r"                runs the script
"clear"            clears any pending exceptions (mostly useless)
"break ###"        set a break at the specified line number
"e <statement>"    eval()uates <statement> in the current frame
"eval <statement>" also eval()s

This is a proof of concept right now since it's only in jsexec and the command
parser is a joke.  Eventually, this will also work for ;EXEC scripts in the
BBS and be even more awesome.  This should be enough to get mcmlxxix off my
back though.

MAX_POST_LEN.

This makes the previous change effectively useless except for using up
space on the hard drive, but is stops from malloc()ing a copy of the data
that is unusable.

solve all the problems involved in a file upload though, so don't get too
excited.

where the file would/could be left open if fdopen() failed.