cmd/mfp-test: add print system testing pipeline skeleton

[ArmxG]

This PR introduces cmd/mfp-test:- the skeleton of the full-stack print system testing pipeline.

What was added
cmd/mfp-test/test/capture.go
DocumentCapture implements abstract.Printer and collects all incoming print documents (job parameters + raw bytes)
for later inspection. It is thread-safe and signals via a channel when the first document arrives.

cmd/mfp-test/test/cupsctl.go
CreateCUPSQueue and RemoveCUPSQueue manage CUPS printer queues via lpadmin. The queue uses the IPP
Everywhere driver (-m everywhere) which works with any standards-compliant IPP printer.

cmd/mfp-test/test/command.go
The main command handler:

• Parses CLI flags: --model, --port, --name, --output, --threshold, --list, --batch, --single, --verbose
• Loads printer model via modeling.NewModel() + model.Load()
• Starts virtual IPP printer via model.NewIPPServer() on a real TCP port (CUPS requires a real network connection)
• Registers the printer with CUPS via CreateCUPSQueue()
• Generates a test PNG image (300x300, RGB colour bands) and sends it through CUPS via lp
• Waits for DocumentCapture to receive the document (30s timeout)
• Reports captured document size and format
• Exits automatically after tests complete
• Always removes the CUPS queue on exit via defer RemoveCUPSQueue(context.WithoutCancel(ctx), name)
  (WithoutCancel ensures cleanup runs even after ctx is cancelled)

cmd/mfp-test/main.go
Thin entry point following project convention:
test.Command.Main(nil)

How to test
cd cmd/mfp-test
go build -o mfp-test .
./mfp-test --model <printer-model-file.py> -v

A model file with ipp.attrs defined is required. Example model files can be generated with mfp-query or obtained from the modeling package.

Expected output:
virtual IPP printer at ipp://localhost:60000/ipp/print
CUPS queue "mfp-test" ready at ipp://localhost:60000/ipp/print
sending test PNG via lp...
captured doc 1: N bytes, format="image/pwg-raster", job="..."

Notes

• PNG is used instead of plain text because CUPS conversion of plain text depends on fonts and gives unpredictable output. PNG/PDF gives consistent results across systems.
• CUPS queue is always removed after each run to prevent accumulation of stale queues in CUPS configuration.
• Phase 3 will integrate Sanskar's ImageComparator framework to compare captured output against the original test image.