SpanWriter now has an argument that allows it to be resizable.
```cs
public SpanWriter(Span<byte> initialBuffer, bool resize = false)
public SpanWriter(int initialCapacity, bool resize = false)
```
If a SpanWriter is set to be resizable, or given an initial capacity instead of a buffer, it must be disposed:
```cs
public static void SomeMethod()
{
using var writer = new SpanWriter(stackalloc byte[512], true);
// stuff
// Automatic Dispose of writer
}
```
This is because the SpanWriter uses `ArrayPool<byte>.Shared` _rented buffers_ to resize. If the writer is not disposed, those buffers will never be reused resulting in a _memory leak_.
It is possible that the SpanWriter will outright ditch the initial buffer if resize is set to true and growing is needed. To check/account for that we can do the following:
```cs
public static void SomeMethod()
{
Span<byte> span = stackalloc byte[512];
using var writer = new SpanWriter(span, true);
// write some stuff that causes the span to grow
span = writer.RawBuffer;
// Do stuff with the span
}
```
Make sure you don't accidentally `Dispose()` the writer or use the `RawBuffer` outside of the `using` block. If you do, bad things will happen! (NullPointerException, or a fresh SpanWriter with no buffer, depending on the situation).
SpanWriter can also now be used with a fixed statement since it has a `PinnableReference()` function.
- [X] Fixes a bug with RunUO where direction is set multiple times and the run flag is not set properly.
TODO: Fix other direction/movement issues and duplicate packets.
Bumps release version
- [X] Adds support for `SkipLocalsInitAttribute`. SkipLocalsInitAttribute skips initializing stack variables including `stackalloc`. I don't think it is wired/working yet but should be implemented soon.
- [X] Optimizes broadcasted packets by checking if the first byte (packet ID) is non-zero. If it is already set, it reuses the buffer.
- [X] Removes unnecessary refs to Spans. I don't think the Span struct itself is mutable, so a ref is not helpful.
- [X] Fixes an issue where a buffer smaller than 8 bytes would not double with enough space in some cases.
- [X] Fixes an issue with dupe copying the savebuffer reference (ugh).
- [X] Streamlines the IGenericWriter API to use better generics.
- [X] Streamlines the IGenericReader API to use better generics.
- [X] Forces `tidying` of a List/HashSet to be done externally since Writers/Readers should not have side effects.
- [X] Fixes an issue where Tidying a list didn't TrimExcess, causing memory leaks.
- [X] Reverted the meaning of `World.Running` to specifically refer to any world state post world loading.
- NOTE: Do not use this if you want to block on world saves. Instead use checks against `WorldState.Saving` states.
- [X] Fixes an issue with serializing negative DateTime deltas.
- [X] Fixes a potential issue with serializing non-UTC DateTime.
Bumps release version
- [X] Removes some string allocations (e.g. split)
- [X] Optimizes some collections
- [X] Converts insensitive to extension methods of built-ins.
- [X] Adds ordinal (case sensitive) string helpers
- [X] Fixes conditionals for in-game commands so they use Ordinal comparisons.
- [X] Replaces ToLower.Contains with InsensitiveContains
- [X] Adds ValueStringBuilder
- [X] Implements ValueStringBuilder in a few places where it makes sense
- [X] Removes the redundant Wrap function and replaces it with an optimized version
- [X] Fixes list conversions in Utility
Closes#351
Bumps release version
- [X] Fixes IPAddresses not having enough space
- [X] Streamlines some code
- [X] Fixes base escortables to not load destination tables in a dangerous way.
- [X] Fixes some bad assumptions about resizing buffers
- [X] Fixes little endian issue with circular buffer writer
This doesn't seem to address #349. That issue requires more investigation because I am not seeing where the issue would be.
Closes#347Closes#322
### Breaking Change
- [X] `Mobile.GetOldPacketFlags()` no longer exists. Instead there is a flag for `Mobile.GetPacketFlags()`.
### Non-Breaking Changes
- [X] Consolidates move to world for mobiles
- [X] Consolidates mobile packets between stygian abyss and older
This PR supersedes changes in #337. That PR will need to be redone and broken out anyway.
- [X] Fixes an edge case where the cycle index goes negative and crashes
- [X] Fixes average CPS calculation
- [X] Fixes issue with logout delay
- [X] Bumps release version
- [X] Updates Movement packets
- [X] Adds OSI fastwalk packets. These aren't used on OSi anymore.
- [X] Adds new movement handling, but looks like the client doesn't use it. (Also leaving the 0x4000 Character List Flag off)
- [X] Adds time sync request handler, but also looks like the client doesn't use it.
- [X] Adds time sync response for time sync, just in case, but it isn't used, so not sure about the arguments.
- [X] Reimplements RunUO's fastwalk to use a circular array on the netstate instead of every mobile
- [X] Removes ClearFastwalkStack from RunUO implementation. This shouldn't be needed anymore
Changed fastwalk settings:
```cs
public static int WalkFootDelay { get; set; } = 440;
public static int RunFootDelay { get; set; } = 220;
public static int WalkMountDelay { get; set; } = 220;
public static int RunMountDelay { get; set; } = 110;
public static bool EnableFastwalkPrevention { get; set; } = true;
public static AccessLevel FastwalkExemptionLevel { get; set; } = AccessLevel.Counselor;
// If this is changed during runtime, then the steps array needs resizing.
public static int MaxSteps { get; private set; } = 4;
```
modernuo.json
```json
{
"settings": {
"movement.delay.runFoot": "220",
"movement.delay.runMount": "110",
"movement.delay.walkFoot": "440",
"movement.delay.walkMount": "220",
"movement.enableFastWalkPrevention": "True",
"movement.fastwalkExemptionLevel": "Counselor",
"movement.maxSteps": "4"
}
}
```
Notes about OSI fastwalk:
While it does work, I can't find a benefit in using it because of the variable speeds. If players moved at a single speed then we could refill the stack every X milliseconds with 6 keys and use a naive token bucket implementation.
Unfortunately variable speeds mount/run/walk/etc means we would have to use a leaky bucket algorithm.
If we are using a leaky bucket algorithm with a variable leak, then we don't need to send tokens because we are already tracking it on the server side.
ModernUO vs OSI Fastwalk:
When the fastwalk was implemented on OSI, it used up to 6 tokens. These tokens were probably distributed every 600-750 milliseconds. To get the same effect, the new fastwalk settings might need to be adjusted. I would tweak them and feel free to let me know what worked for you!
Bumps release version
### API Breaking Change
Combined `AsciiMessage` and `UnicodeMessage` into a single function that takes two arguments, `bool ascii` and `string lang`. Lang can be null (or anything) if ascii is true.
- [X] Changes message packets
- [X] Cleans up some code
- [X] Uses benchmarks to determine if the spanwriter + copyfrom
### Change to how DisplayEquipInfo works
* Fixes a bug where raw name wasn't used
* Fixes another bug where an empty, blank, or null string was still sending an empty crafted by property.
Bumps release version
- [X] Fixes several bugs
- [X] Updates more ordinal issues
- [X] Cleans up the code a bit
- [X] Turns classes static that should have been
- [X] Changes TcpServer.Instances to a HashSet
Bumps release version
- [X] Changes effect packets
- [X] Updates playing sounds
- [X] Updates Effects class to have more options so there are less direct calls to building the packets
- [X] Changes bonding status packet
- [X] Removing Kestrel & Libuv
- [X] Cleaning up NetState
- [X] Removing System.IO.Pipelines
- [X] Cleaning up packet reading
- [X] Adds a maximum of 5000 sockets (configurable) to prevent OOM
- [X] Replaces the AsyncState with a thread-safe wrapped boolean called NetworkState
- [X] Removes Parallel.ForEach (no perf gain)
- [X] Removes custom houses compression on another thread
- [X] Test high load scenarios
Bumps release version
