Couldn't get a reliable repro, but this is probably what that bug was.
It's not very throughly investigated, but we'll be tweaking those parts
of the codebase anyway, so we can examine if there's a deeper issue
later.
* Restrict fcntl() to only the flags we need
* Non-fatally deny tgkill() and rt_sigaction() so that segfaults don't
result in a SIGSYS. They're debuggable either way, but this way it's
clearer what the issue is right away.
* Allow truncate() and ftruncate() for sqlite's alternate journal modes
* Slight macro cleanup
* Add missing colon in a DB log message
We don't need to worry about compilation problems arising if glibc or
musl-libc add their own wrapper for the seccomp() syscall in the future.
Ours will/would just silently take precedence over the external one
without interfering with compilation. This should work regardless of
whether libc uses weak symbols and regardless of whether libc is
dynamically or statically linked into the executable. The wrapper's
signature has been stripped of its static and inline qualifiers, as it
must match the exact declaration the libc headers will/would use.
Further, if a pre-compiled binary is run on a system which genuinely
doesn't support seccomp(), it'll just return ENOSYS and the server will
terminate with an error. The user can then just disable the sandbox in
the config file. We don't need any special logic for that scenario.
Adding the assembly source files to the build system(s) would be more
trouble than it's worth, considering we don't make 32-bit builds for
much other than satisfying our curiosity.
There are some network configurations in which it's undesirable; such as
reverse tunneling through ssh. These are obscure enough to allow leaving
the option undocumented (in the example config file).
This fixes a UX issue, where if you accidentally capitalized a letter
in the username when logging in, it would instead create a new account.
The behavior was confusing, since to the user it looks as if their
characters were deleted or progress was not saved.
In order for this to work, duplicate accounts (e.g. username and USERNAME)
need to be deleted/renamed. The server will *detect* if any duplicates
exist. If any are found, it will direct the server operator to a pruning
script, or they can choose to resolve the duplicates manually.
This wasn't strictly necessary as this command has outlived its
usefulness, but I had gone ahead and rewritten it because it was (barring
taskStart()) the only place in the codebase that accesses Chunking::chunks
outside of Chunking.cpp. This became apparent during a (currently paused)
effort to improve the API that the Chunking namespace exposes to the
rest of the codebase.
I went ahead and rewrote the rest of this command as it was poorly
implemented anyway. This has been sitting in my working directory
uncommitted for a few months, so I may as well push it.
Explanation: it was uncertain whether mobs could perform critical hits, since the color of damage numbers didn't change at all. However, I found that male characters will actually use a different sound effect when receiving a crit (I confirmed this SFX appeared in old FF videos), so I went ahead and re-enabled it.
The server administrator must now specify which patches they want the
server to load (if deviating from the defaults). There are multiple
reasons for this:
* It's useful to be able to pick and choose which patches you want to
boot the server with; without having to move the directories in and out
of the patch directory
* This way, we can have different default patches for different builds
of the game (104 vs 1013)
* ...ergo, it's easier to rapidly switch builds without having to
rearrange your workspace to properly run them
* This also allows us to remove the std::filesystem stuff, which has
spotty compatibility with slightly older (but still current) versions of
the compilers
