GitLab

Replace use-of/linking-with wsock32.lib with ws2_32.lib to fix issues such as this one coming-up with the HAProxy feature merge:
unresolved external symbol __imp__inet_pton@12 referenced in function _xpms_accept 

There's no new opt-in option for this new feature - maybe there should be.
For geneb.

Sorry guys. :(

This one should build even on Win32.

Needed by MinGW32... we'll see if MSVC needs it as well soon.

This doesn't make anything much faster since the bulk of the time
is building ctyptlib which doesn't support -j, but at least Cyan
won't complain every time he creates an ebuild anymore.

This requires anything that includes sbbs.h to have cryptlib and
mozjs already built/isntalled... to track this, I've added the new
extdeps.mk file.

This is pretty terrible though... not sure how important that
dependency really is.

Parse single-quoted "name" portion of email-addr header fields. I don't think this is going to fully solve the issue with FTN netmail gated to SMTP and replied via Outlook, but it's a start. Try that Nelgin.

This is because it includes js_version.h if JAVASCRIPT is defined,
and JAVASCRIPT is defined for everything using CFLAGS.

It may be worth re-working the JAVASCRIPT define, but that's a bunch
of extra work.

This is the first (and easiest) step in getting parallel builds
working.

160M is the value used by the Mozilla/SpiderMonkey JSShell project, so I'm
rationalizing that as an acceptable default. Of course, it would be good
to optimize memory usage in JS scripts, so we'll work on that too. So, if
you're a JS developer, best to play with lower values here to see if/where
you're being potentially wasteful in your memory usage, but most sysops
just want everything "to work".

Not sure how I missed this bit.

Deleted users would be "properly" loaded as 'X' which would mean
they wouldn't show up in rankings and couldn't be attacked.

... but do it do it correctly. This allows for automated game-reset, e.g.
via cron job or timed event.

This can likely be built on Windows, but since I expect I'll be the
only one supporting this for the forseeable future, get the build I
release in here.

As reported by plt from the I, the R, and the C.

Also, abort menu upon Ctrl-C.

Then, add JSDOC support to crypto class constant objects.

Also, since crypto stuff is almost half of all the docs, move all
the crypto classes and objects to the end. It would be nice if these
massive lists were hidden by default line the menu tree at the start.

Previously, properties were only supported via the TinyID mechanism
in the JS API.  This limited the properties to have a unique signed
8-bit integer ID, which prevented more than 256 properties from
existing in a built-in object.

The id value can be a name however, a TinyID is not required.  For
property descriptions with an ID outside the range of 8 bits, define
them using just the name.  When this is done, the getter and setter
functions will need to convert the names to an ID themself.

This isn't ideal... ideally, each would have a unique getter and
setter, but doing it this way puts of requiring an internal JS API
refactor and allows the CryptCert class to have all the defined
properties.

As implied... CryptCert now supports all properties, and JS Docs
will be created for them.  This section is yuge!

Copy/paste without the update required a single character to have
seven different values... which is unlikely to actually happen.

The calls to cryptlib are still broken, so that's a to-do item.
e.g. calling cryptstatuSOK() in st_crypt.c init_crypt() throws an
exception:
"Run-Time Check Failure #0 - The value of ESP was not properly saved across a
function call.  This is usually a result of calling a function declared with
one calling convention with a function pointer declared with a different
calling convention."

<@Deuce> SyncTERM on MSVC really shouldn't be part of CI since it's not expected to work.
<DigitalMan> why not?
<@Deuce> Because it's not supported.
<@Deuce> So it's not part of the integration path.

Have js.on_exit() throw an exciption with a useful(?) error when the
scope already has private data, and that data is not an on_exit list.

This fixes #182.

This should be use in all the other places, but this was enough for
me to debug the issue at hand.