DD Archive Viewer now supports the built-in Archive class in Synchronet 3.19

See merge request !299

If the search up the scope tree for js.exec_dir fails, the script 'path' would
be left blank leading to a weird error message.
e.g. !JavaScript default.js line 249: Error: Script file () does not exist

There's a weird issue where the scope-walk to find a "js.exec_dir" isn't
working (for Nick Young or Nightfox) after invoking DDMsgReader. This should
fix (or work around) that issue.

And js.exec() doesn't search the mods_dir first when a full script path isn't
passed, so with this change, child scripts that are in the mods_dir should now
work as a sysop would expect.

When there are no changes (areas added or removed) via an areafix message, the
data/AREASxxxxxx temp file would be left behind. The file was only removed if
areas were added or removed from the area file via areafix message.

The orphaned data/AREASxxxxxxx files were reported via DOVE-Net by Gamgee (PALANTIR)
If you have these stale files, you can safely delete them.

DDMsgAreaChooser fix for exiting immediately if the first message group has no sub-boards. Fixes #589

Closes #589

See merge request !298

DDMsgAreaChooser fix for exiting immediately if the first message group has no sub-boards. Fixes #589

As reported by Pasta68K:
FreeBSD's ln doesn't support -r

ln -sfr clang.freebsd..exe.release/* /sbbs/exec
ln: illegal option -- r

So use realpath instead. Hopefully all the targeted *nix platforms include
'realpath'.

Make behavior consistent with smbutil and fixmsb.

Eliminates a SANITIZE build error

Hopefully Cyan agrees that 1.10 was the logical next version number.
Hopefully these are the only 2 places where the version is duplicated.

Someone should add IPv6 exempted addresses too

Maybe perform a check against ctrl/dnsbl_exempt.cfg too?

Socket.close() always sets the descriptor to -1 which is/would be logged here
as an unsigned int (%u) or 4294967295.

If this continues to log 4294967295, then a check of descriptor >= 0 would be
prudent.

And as Cyan said:
<Cyan> the IP would be more useful though

Always log (a notice-level log message) when there's a DNSBL match, along with
the address it actually resolved to (indicates which list the IP address was
found in).
No log message otherwise.

Note: resolve_ip() doesn't actually support IPv6 yet, so I don't think this
IPv6 address parsing/reformatting logic has actually been tested.

Added RBL lookups on connection via DroneDL.

See merge request !297

JavaScript is lovely at data validation ... not!

Synchronet IRC services uses RAKILL to remove any previouslly

See merge request !296

I don't believe them, but I'll fix it anyway.

Should fix ASan "crash".

This reverts commit 0a6fdb95.

This simply caused the callback to never be called.

Check size of X11 property before dereferencing as an Atom.
This fixes:
==10203==ERROR: AddressSanitizer: heap-buffer-overflow on address
0x602000065970 at pc 0x5c5834fb8812 bp 0x7819c717c220 sp 0x7819c717c218
READ of size 8 at 0x602000065970 thread T2 (X11 Events)
    #0 0x5c5834fb8811 in net_wm_state_is_cb ../conio/x_events.c:1541

But I'm not clear when nitems_return (nir) might represent > 8-bit items or how
to detect that, so perhaps there's a *better* fix?

Caught be GCC ASan.

This does look like a false positive to me: the if clauses that use fptr appear
to only cover the same values as the switch-case statements. But harmless
defensive change here.

Updated Tech question pool. Now valid until 6/30/2026

See merge request !295

warning: format not a string literal and no format arguments [-Wformat-security]
<nelgin> gcc version 9.4.0

The commit just the zero-length behaviour from uninitialized use
with possible undefined behaviour to guaranteed undefined behaviour.

It's unlikely this is what Digitalman reported from ASan and UBSan

Fix multiple warning: ignoring return value of ‘asprintf’
    declared with attribute ‘warn_unused_result’

This was really just a test to make sure I had Geany successfully reconfigured
to no longer auto-strip trailing whitespace. Spoiler: It wasn't.

As Deon pointed out on DOVE-Net::Synchronet Discussion. Thanks!

This macro hasn't done anything meaningful since we stopped using really old
versions of Borland compilers (and std libraries) where strerror() returned a
string terminated with a line-feed (\n) character.

These should only be used in the sbbs_t methods called by node_thread or
event_thread. Any other thread-shared sbbs_t instances should not use these
methods since they share a single error description string buffer.

As reported by Deon on DOVE-Net: type 2.2 packet generation was not (since
SBBSecho v3.0) setting the origin and destination point values correctly:
was leaving the packet type 2.0 date values in these fields.

Also, zero-out the reserved bytes in the type 2.2 packet header (don't leave
the type 2.0 date values in there).

Thanks for the report Deon!

If the oldlen param value is less than one, set to a sane non-zero value (79).
Fix issue #576

this is recommended to reliably display the call stack when using the
print_stacktrace option.