Compressor: throw on overflow instead of returning a negative size

Compress() and FrameTransformation::CompressImage() returned int64_t with a
negative value meaning "did not fit". That is a footgun: the negative result
silently converts to a huge size_t if a caller forgets to check it. Return
size_t and instead throw a named CompressionBufferTooSmallException (deriving
from JFJochException, Compression category) when the output would not fit the
destination buffer.

The receiver catches it explicitly and drops just that frame, as before; the
offline/GetCompressedImage path uses a worst-case buffer so it never throws.
Add a test that a too-small destination throws and a worst-case buffer does not.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

receiver/FrameTransformation.cpp

@@ -54,7 +54,7 @@ FrameTransformation::FrameTransformation(const DiffractionExperiment &in_experim
     }
 }
 
-int64_t FrameTransformation::CompressImage(void *output, size_t output_size) {
+size_t FrameTransformation::CompressImage(void *output, size_t output_size) {
     return compressor.Compress(output, output_size, precompression_buffer.data(), experiment.GetPixelsNum(),
                                experiment.GetByteDepthImage());
 }
@@ -70,7 +70,7 @@ CompressedImage FrameTransformation::GetCompressedImage() {
                                                    experiment.GetPixelsNum(),
                                                    experiment.GetByteDepthImage()));
         data = reinterpret_cast<uint8_t *>(compressed_buffer.data());
-        size = static_cast<size_t>(CompressImage(compressed_buffer.data(), compressed_buffer.size()));
+        size = CompressImage(compressed_buffer.data(), compressed_buffer.size());
     }
 
     return CompressedImage(data, size,experiment.GetXPixelsNum(),