feat(go): implement a binary reader

[chengxilo]

Which issue does this PR close?

N/A

Rationale

Iggy, when read/write binary, use little endian. When there is a string, it has a prefix provide its length, and the actually string body. So instead of using the binary.LittleEndian.Uint32(payload[processIDPos : processIDPos+4]) and use position ++ , why not just wrap everything in a reader to make it simple and clear.

As the Stats deserialize function need to be updated, I think this change would be helpful. #2964 (comment)

Here is a example that shows the difference:
Before:

iggy/foreign/go/binary_serialization/stats_serializer.go

Lines 53 to 90 in 9c8b5cc

	func (stats *TcpStats) Deserialize(payload []byte) error {
	stats.ProcessId = binary.LittleEndian.Uint32(payload[processIDPos : processIDPos+4])
	stats.CpuUsage = math.Float32frombits(binary.LittleEndian.Uint32(payload[cpuUsagePos : cpuUsagePos+4]))
	stats.TotalCpuUsage = math.Float32frombits(binary.LittleEndian.Uint32(payload[totalCpuUsagePos : totalCpuUsagePos+4]))
	stats.MemoryUsage = binary.LittleEndian.Uint64(payload[memoryUsagePos : memoryUsagePos+8])
	stats.TotalMemory = binary.LittleEndian.Uint64(payload[totalMemoryPos : totalMemoryPos+8])
	stats.AvailableMemory = binary.LittleEndian.Uint64(payload[availableMemoryPos : availableMemoryPos+8])
	stats.RunTime = binary.LittleEndian.Uint64(payload[runTimePos : runTimePos+8])
	stats.StartTime = binary.LittleEndian.Uint64(payload[startTimePos : startTimePos+8])
	stats.ReadBytes = binary.LittleEndian.Uint64(payload[readBytesPos : readBytesPos+8])
	stats.WrittenBytes = binary.LittleEndian.Uint64(payload[writtenBytesPos : writtenBytesPos+8])
	stats.MessagesSizeBytes = binary.LittleEndian.Uint64(payload[messagesSizeBytesPos : messagesSizeBytesPos+8])
	stats.StreamsCount = binary.LittleEndian.Uint32(payload[streamsCountPos : streamsCountPos+4])
	stats.TopicsCount = binary.LittleEndian.Uint32(payload[topicsCountPos : topicsCountPos+4])
	stats.PartitionsCount = binary.LittleEndian.Uint32(payload[partitionsCountPos : partitionsCountPos+4])
	stats.SegmentsCount = binary.LittleEndian.Uint32(payload[segmentsCountPos : segmentsCountPos+4])
	stats.MessagesCount = binary.LittleEndian.Uint64(payload[messagesCountPos : messagesCountPos+8])
	stats.ClientsCount = binary.LittleEndian.Uint32(payload[clientsCountPos : clientsCountPos+4])
	stats.ConsumerGroupsCount = binary.LittleEndian.Uint32(payload[consumerGroupsCountPos : consumerGroupsCountPos+4])
	position := consumerGroupsCountPos + 4
	hostnameLength := int(binary.LittleEndian.Uint32(payload[position : position+4]))
	stats.Hostname = string(payload[position+4 : position+4+hostnameLength])
	position += 4 + hostnameLength
	osNameLength := int(binary.LittleEndian.Uint32(payload[position : position+4]))
	stats.OsName = string(payload[position+4 : position+4+osNameLength])
	position += 4 + osNameLength
	osVersionLength := int(binary.LittleEndian.Uint32(payload[position : position+4]))
	stats.OsVersion = string(payload[position+4 : position+4+osVersionLength])
	position += 4 + osVersionLength
	kernelVersionLength := int(binary.LittleEndian.Uint32(payload[position : position+4]))
	stats.KernelVersion = string(payload[position+4 : position+4+kernelVersionLength])
	return nil
	}

After

func (s *Stats) UnmarshalBinary(payload []byte) error {
	r := newReader(payload)
	s.ProcessId = r.u32()
	s.CpuUsage = r.f32()
	s.TotalCpuUsage = r.f32()
	s.MemoryUsage = r.u64()
	s.TotalMemory = r.u64()
	s.AvailableMemory = r.u64()
	s.RunTime = r.u64()
	s.StartTime = r.u64()
	s.ReadBytes = r.u64()
	s.WrittenBytes = r.u64()
	s.MessagesSizeBytes = r.u64()
	s.StreamsCount = r.u32()
	s.TopicsCount = r.u32()
	s.PartitionsCount = r.u32()
	s.SegmentsCount = r.u32()
	s.MessagesCount = r.u64()
	s.ClientsCount = r.u32()
	s.ConsumerGroupsCount = r.u32()
	s.Hostname = r.u32LenStr()
	s.OsName = r.u32LenStr()
	s.OsVersion = r.u32LenStr()
	s.KernelVersion = r.u32LenStr()
	return r.Err()
}

What changed?

A reader for byte slice is created.

Local Execution

• Passed
• Pre-commit hooks ran

AI Usage

1. smart Claude Sonnet 4.6 and my rotten brain
2. entire implementation, manually modified some part to make sure the error output readablity.
3. Reviewed the tests and code, actually modified some logic of reading binary and works well.
4. yes

[codecov]

Codecov Report

❌ Patch coverage is 82.60870% with 16 lines in your changes missing coverage. Please review.
✅ Project coverage is 72.00%. Comparing base (f0e8578) to head (bcf7a53).
⚠️ Report is 2 commits behind head on master.

Files with missing lines	Patch %	Lines
foreign/go/internal/codec/reader.go	82.60%	8 Missing and 8 partials ⚠️

Additional details and impacted files

@@             Coverage Diff              @@
##             master    #2986      +/-   ##
============================================
+ Coverage     71.81%   72.00%   +0.18%     
+ Complexity      930      925       -5     
============================================
  Files          1116     1114       -2     
  Lines         92616    92426     -190     
  Branches      70139    69885     -254     
============================================
+ Hits          66512    66547      +35     
+ Misses        23543    23314     -229     
- Partials       2561     2565       +4     

Flag	Coverage Δ
go	37.28% <82.60%> (+0.90%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
foreign/go/internal/codec/reader.go	82.60% <82.60%> (ø)

... and 19 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.