Embarassingly, I'd added SS_NEW_DAY recently (for a different purpose) with
such a similiar name. So let's rename this flag since it serves a very
different purpose and avoid the confusion between the two flags.

uncrustify nl_split_if_one_liner setting

... using uncrustify mod_paren_on_return config

White-space changes only, exception being the rare insertion of NL before
closing brace (couldn't find the option to disable that behavior).

I excluded some header files (e.g. sbbs.h) since uncrustify seemed to be doing
more harm than good there. I might just end up applying different set of rules
to .h files.

Rob Swindell authored 2 months ago and Deucе committed 2 months ago

I don't know if these seek failures are actually happening or not, but
reading from or writing to the wrong offset in the node.dab file could
explain some of the node.dab corruption I'm seeing from macOS (over
SMB share).

Include O_CREAT access mode flag in opennodedat().

I experimented with excluding O_DENYNONE when the NM_CLOSENODEDAB flag is set
(to hopefully work-around MacOS Samba node.dab corruption issue), but that
didn't work with SBBSCTRL leaving the file open in SH_DENYNONE mode, so will
have revist that, but using a common open function helps.

Looking into potential causes of issue #843, I found several instances where
we call getuserdat() without checking the return value and restoring the
useron.number to the current user number upon error: getuserdat() zeroes out
the user struct/number upon error, a bad API choice made 23 years ago.

Replace those instances with calls to sbbs_t::getuseron() which logs any
open/lock/read failures of the user.tab and does not modify/zero-out the
sbbs_t::useron struct upon error.

Since we we're not using opennodeext(), we don't have the path/fname for
any failure error message here.

CID 515714

Custom NodeStatus and NodeAction/Activity strings are now displayed pretty
much everywhere possible (except the node utility, since it doesn't read
any configuration files). This fixes issue #618, finally.

If you have custom (non-blank) NodeAction* strings in your text.dat/text.ini
file, or you change these strings during runtime (e.g. using bbs.replace_text)
this will impact you:
- do not include the user name or any other fields normally included in
  printed node status in your string - just the customized node action and
  really, you should only use/override these strings for dynamically changing
  node status (i.e. you're using NodeActionCustom and you're setting different
  custom strings depending on current node activity). Consider using the new
  NodeActivity* text.dat strings instead if your customized string doesn't
  change during run-time.
- don't use the %s/%u/%d specifiers any more - use @-codes instead, if needed

If you want static/permanent node action strings, set new new NodeActivity*
strings in your text.dat or (preferrbaly) text.ini file instead.
This means the node actions can now be localized for other languagues!

I'm still using the node.exb file (I did consider alternatives, e.g.
node.ini or node*/activity.asc), but decided to postpone such a change.
The node.exb file is still used, but the strings represent just the custom
(extended) node *activity* and not the entire node status string.

Use of Ctrl-A codes needs to be considered/accounted-for (currently isn't).

"performing custom action"

Normally, the sysop would set the NodeActionCustom text.dat string
to override this.

Addresses Clang warnings about deprecation

getnodedat() now performs a non-locked read by default.

Perform more checks of getnodedat() return value before calling putnodedat()
to avoid unintentionally zeroing out node.dab records.

Add/use unlocknodedat() method for unlocking a node.dab record without
writing.

Note: The userdat.c getnodedat() and putnodedat() still return int (i.e. 0 on
success), so that can be confusing.

Upon any node.dab lock or read failure, this code would cause errormsg() which
would often/usually end up claling getnodedat() which would block forever trying
to acquire the ndoefile_mutex (introduced in commit b9633069, I'm not clear why).

Unlock/release the mutex *before* calling errormsg().

Fix new date editing bug in user editor: we must use the numeric format
for dates when creating strings to be edited.

... this was the cause of some observed unnecessarily high disk/file server
(Samba) utilization, as we call getnodedat() a lot. utime() opens and closes
the file, which was already open - and we're not modifying the file, so
updating the 'modification time' here is wrong anyway.

Disabling this 21-year old bit of logic resulted in a pretty dramatic
reduction in Samba (smbd) CPU utilization on Vertrauen.

If a BBS actually needes this hack (e.g. for NFS compatibility, as Deuce
eluded in the comment), they'd be better off just setting the "Keep Node File
Open" node setting (in SCFG->Nodes) to "No".

For SFTP sessions, there's no shell/terminal, so no need to run command
shells (with their input timeouts, etc.). Reflect the node connection as
"via sftp" in the node status. Handle node interruption signal. Probably more
to do here with node status/actions (e.g. it'd be nice to set the action to
"uploading or "downloading" when appropriate).

So Clang-FreeBSD was warning (in compiles of scfg/scfg*.c by Deuce):
result of comparison of constant 100000 with expression of type 'uint16_t'
(aka 'unsigned short') is always true

Why? Cause a uint16_t's max value is 65535 (less than 100000). Sure we could
have just lowered the UIFC max number of config items to 65535, but that would
have been too easy. And why are these compared-with values of type uint16_t to
begin with? Because most ctrl/*.cnf lists (of configuration items) were
limited to 65535 entries cause ... 16-bit DOS, historically. Now that *.cnf
files aren't used, we could just increase these scfg_t.*_total type sizes from
16 to 32-bits, yeah? The result is this commit.

I went to (signed) int so we could still keep -1 as the special illegal
sub/dir num value (e.g. INVALID_SUB, which is sometimes used to indicate the
email message base). Theoretically, 2 billion configuration items could be
supported in these lists, but SCFG will limit you to 100000 anyway. So there's
a whole lot of s/uint/int in this commit.

I'd be very surprised if this doesn't result in some new GCC/Clang warnings,
but at least the old "comparison of constant 100000" warnings are now gone!

A bunch of possible (but often, not really) use of undefined values.
Some ignored return values (e.g. of chsize/ftruncate, read, write, fgets).

Other than some added diagnostics upon some of these unexpected syscall
failures, there should be no change in behavior from this commit.

That's all these macros were doing anyway, so no change.

Vestiges of MS-DOS and 16-bit ints and local keyboard access.

This should address issue (feature request) #196.

The last 20 sets of displayed messages are stored as data/msgs/<user-num>.last.#.msg. This number is currently hard-coded, but could be configurable in the future. I say "sets" because messages are batched-up and displayed together normally, unless a user is actively polling for new users (e.g. while at the Ctrl-P/PrivateMsg prompt).

This involved getting rid of some copy/pasta in sbbs_t:getsmsg() as well by creating/using/reusing readsmsg().

The current node action was not saved/restored when being forced into node-to-node private chat (by a sysop). This could lead to subsequent chat attempts by the same node to lead to a false indication that the nodes were chatting with each other.
As reported by Altere (ATHEL).

When a sysop wants to chat, he wants to chat right now, damnit!

automatically revert to Raw/TCP mode. This resolves the issue of users that
connect using modem/serial emulators over TCP that do not actually support Telnet and then either try to transfer files, which usually will not work due to escaped (or unescaped) IAC (0xFF) characters, or the user may be sent Telnet Go Ahead (GA) sequences which may display as periodic garbage in their terminal program.

This seemed like a better solution than having (yet another) dedicated terminal server port for "raw" connections over TCP, though a "raw" port still might not be a bad idea for a future enhancement. RLogin is already super close to "raw", so it's probably not really necessary to have another port configured/listening.

- nodelist (default: "nodelist")
- who's online (default: "nodelist -active")
- private message (default: "privatemsg")

This allows the nodelist and private message modules to be "installed" without
over-riding global hot key events (SCFG->External Programs->Global Hot Key
Events), allows the Baja functions NODELIST_ALL, NODELIST_USERS, and
PRIVATE_MESSAGE and JS bbs.list_nodes(), bbs.whos_online(), and
bbs.private_message() methods to use external modules (JS or Baja)
automatically.

The old C++ logic remains for these features if the module's are cleared (set
to a blank string) in SCFG, but at some point, that code will likely be removed.

optionally including the current (your) node.
New @-codes:
"ANODES" (active-nodes count), aliases: "ANODE" and "AN"
"ONODES" (other-active-nodes count), aliases: "ONODE" and "ON"

Make "NN" an alias for "NODE" and "TN" and alias for "TNODE".
Add "TNODES" alias for "TNODE" as well, for symmetry.

the current column number. Use the correct member variable name (column).

record in the node.dab - only the node(s) that have the requested user
number.
In sbbs_t::getsmsg() and ::getsmsg(), strip all invalid Ctrl-A attribute
codes (e.g. clear screen, pause, etc.) since short-messages (telegrams)
can come from external sources.

getsmsg() is now based on the current console column (!=0), not the current
node action.

When using the MSVC2017 CRTL, an unlock() of a region that was not previously
locked is a blocking call causing major slowness in the MSVC2017 build.

lock across things with timeouts in them, but since if another thread has
the file locked we shouldn't be fillding with it anyway, this should be
fine.

Fixes weird (and relatively rare) timing issues with node.dab accesses on a
many core host.