As reported by Ragnarok and The Lizard Master, invoking a JS global
hot-key handler (e.g. hitting Ctrl-U or Ctrl-P with the privatemsg
or nodelist modules installed as handlers) would lead to the global
object being cleared/corrupted so other scripts would then fail to
run correctly (e.g. Error: can't open sbbsdefs.js: No such file or directory)

This probably explains the JS_GC/JS_ENDREQUEST change needed when
the hotkey-specific JS runtime/context/scope was introduced last year.
Should revisit that too.

Might as well use their bandwidth before my own.

(e.g. errors become warnings) - I'm tired of seeing this in my
error.log every day:
SEND/TLS ... ERROR 'Received TLS alert message: Handshake failure' (-15) setting session active

Also incremented the rev to 1.60.

Fix problem introduced in
b1ecd9b6 (August 8, 2019)

When sbbs_t::printfile() is used to display a file that is < 2MB in size
and the P_OPENCLOSE mode flag was not specified, the file is read and
displayed one line at a time. This allows for the quick display of very
large files. This was accomlished by calling sbbs_t::putmsg() for each
line. Unfortunately, putmsg() would save and restore console mode flags
like the auto-screen pause control flags stored in sbbs_t::sys_misc
so this would render some features like the POFF @-code, inoperable.
This problem was reported by WitNik with files displayed from
Bulleeye! Bulletins.

Other issues were discovered with the WORDWRAP/WRAPOFF @-codes.

Fixed with the addition of sbbs_t::putmsgfrag() which is used by
printfile() to display a message fragment (one line) without
saving/restoring console state flags.

Fix massive regression introduced in previous commit.

Iterate over RSS feed channels and items in a less ugly way.
Don't iterate over all item children and cherry-pick the ones
we want, just target the desired elements instead.

Remove the "This is shitty" comment. It's still shitty overall,
but that particular shitty thing has been de-shitted.

See merge request sbbs/sbbs!5

Iterate over RSS feed channels and items in a less ugly way.
Don't iterate over all item children and cherry-pick the ones
we want, just target the desired elements instead.

Remove the "This is shitty" comment. It's still shitty overall,
but that particular shitty thing has been de-shitted.

Enable CI build for src/sbbs3/*

See merge request sbbs/sbbs!4

<nelgin> !JavaScript : uncaught exception: Unknown scheme! 'undefined'
<nelgin> (error with no line number) is the worst.

Services with descriptions that begin with "PETSCII" or "COMMODORE"
will have a their ScreenMode set to C64 and NoStatus=true. If the
description also contains "80", then the ScreenMode will be set to
C128-80col instead.

And fix Debug build error (/ZI) and warning (/Gm)

That way, when the error is reported (e.g. in sbbs), it'll include
the file name and line number.

Also replaced some unsafe str function calls with safe(r) ones.

This reverts commit b76f4695.

Clones the git repo into $(SBBSIDR)/git building everything there and then
copying the run-time dirs (e.g. ctrl, exec, text, xtrn, web) to the
$(SBBSDIR). This means a sysop cannot simply "git-update" (there is no
such thing) the exec or xtrn dirs as someone might have done previously with
CVS. Update instructions on the wiki will be forthcoming.

Removed the NOCVS option (if the workdir already exists, no clone occurs).
Changed CVSTAG to just TAG. It does not create a local branch for the
checked-out tag (perhasp it should).
Removed the INSTALL type (only CLASSIC was installed anyway), that should
be easy to restore if/when UNIX-style install is ever implemented.

gtkmonitor.c:156:79: warning: iteration 9 invokes undefined behavior [-Waggressive-loop-optimizations]
   sprintf(str,"%-2d  SL: %-2d  F1: %s",i,cfg.val_level[i],ltoaf(cfg.val_flags1[i],flags));
                                                                 ~~~~~~~~~~~~~~^~~
gtkmonitor.c:151:2: note: within this loop
  for(i=1; i<=10; i++) {
  ^~~