Fix method parsing inconsistency

[icon-bobk]

This patch addresses two issues with parsing RPC Methods.

The first issue is an incorrect size of the method.flag during the read. In the master branch it is stated as a Uint16, however it is an Uint32. This is already corrected in the V2 branch. Comments were adjusted to represent the correct memory alignment (also corrected in V2).

The second patch uses the RPC method length as the ending position before continuing parsing the method array. If for some reason the data structure adds any additional information beyond the method.parameters value, it would offset the reads creating invalid parsing of the structure. (This was found while connecting to a Beckhoff 4026.8 host)

[jisotalo]

Thank you very much @icon-bobk!

I found that issue with flag size during v2 development, but probably forgot that it should be fixed in v1 too...

The second one: The fix makes sense. I'm going to develop the RPC to v2 soon, so I will check it out and then merge this pull request too. So will be backported to v1 also.

By checking TwinCAT.Ads.dll from nuget, we can see that RPC methods seem to have possibility for attributes after parameters (this is new I think). So you are absolutely right.

/// <summary>Parameters Collection</summary>
internal AdsMethodParaEntry[] parameters = Array.Empty<AdsMethodParaEntry>();
internal AdsAttributeEntry[]? attributes;

int start18 = start17 + SubStructureReader<AdsMethodParaEntry>.Unmarshal((uint) this.parameterCount, marshaler, span.Slice(start17, num2 - start17), out this.parameters);
if (this.flags.HasFlag((Enum) AdsMethodFlags.Attributes))
{
  ushort elementCount = BinaryPrimitives.ReadUInt16LittleEndian(span.Slice(start18));
  int start19 = start18 + 2;
  start18 = start19 + SubStructureReader<AdsAttributeEntry>.Unmarshal((uint) elementCount, marshaler, span.Slice(start19, num2 - start19), out this.attributes);
}

[jisotalo]

Parsing attributes should now work in v2, however I don't have TC 4026 yet installed for testing.

[jisotalo]

Hi @icon-bobk!

Would you be interested to run v2 test suite against your 4026 PLC?

I don't have access to 4026 at the moment and I haven't yet installed it to my PC as many people I know have had issues.

[jisotalo]

Merged and tested. Sorry for delay!