Supply or retrieve the encoded BODY bytes.

Initialize, set and get, UBYTE pointer. Default NULL.

The buffer is borrowed, never copied or freed. Keep complete readable

data available for every Setup; changing it while a decoded bitmap

exists does not update that image. Get returns the original pointer. No

length is stored or checked. Provide readable padding for the planar

converter prefetch as described in the class entry.

Select raw or ByteRun1 decoding.

Initialize, set and get, UBYTE. Default 0. Stored 0 means raw; 1 means

ByteRun1; larger stored values make Setup fail.

Tag data is converted to UBYTE before validation, so values outside that

type are truncated. Set does not decode again. ByteRun1 uses nonnegative

control bytes for literal runs, -1..-127 for repeated-byte runs and -128

as a no-op, without checking encoded input bounds.

Set or retrieve the number of color bitplanes.

Initialize, set and get, LONG. Default zero; Setup accepts 1 through 8.

The decoded bitmap uses this color depth. Masking 1 adds a separate

consumed plane to the input row but does not increase decoded color

depth. Set performs no immediate validation or regeneration. Supply

dimensions, plane data and optional palette consistent with this value.

Indicate whether each source row includes an extra mask plane.

Initialize, set and get, UBYTE. Default 0.

Only value 1 has special handling: one extra plane is consumed after the

color planes. It is discarded, not converted to a drawing mask. Other

stored values are treated as no extra plane and are not rejected. Set

does not regenerate decoded data. This attribute does not itself select

a transparent color or enable Bitmap alpha drawing.

Bitmap.mui

Decode borrowed ILBM-style BODY data into an owned bitmap during Setup,

then display it through Bitmap. This class does not read an IFF file,

parse chunk headers or infer dimensions. Supply Bitmap_Width,

Bitmap_Height, Bodychunk_Depth and Bodychunk_Body consistently;

dimensions and depth default to zero, Body to NULL. Compression and

Masking default to zero.

Rows contain word-aligned plane bytes: bytes per plane row equal ((width

+ 15) & ~15) / 8. Color planes are interleaved by row, followed by one

extra plane only when Masking equals 1. Compression 0 reads that layout

directly; compression 1 expands ByteRun1 data for the combined planes of

each row. The extra mask plane is consumed but not used for drawing

transparency. Use Bitmap's separate transparency mechanism if needed.

Setup rejects NULL Body, nonpositive dimensions, depth outside 1..8, or

stored Compression above 1. It returns FALSE on those checks or required

allocation failure. It creates the source bitmap before calling Bitmap

Setup, supplies it through Bitmap_Bitmap, and frees it if superclass

setup fails. Temporary scanline bitmap allocation can fall back to

per-pixel writes.

Cleanup first invokes Bitmap Cleanup, clears Bitmap_Bitmap and then

frees the decoded bitmap after WaitBlit. Dispose frees any remaining

decoded bitmap, but does not replace the complete superclass cleanup

lifecycle. Body is never freed. Setters update stored fields only; they

do not decode again, request redraw or relayout. Width/Height tags

update both this class and Bitmap, so live changes can disagree with the

old decoded bitmap.

There is no encoded-data length parameter or decoder input bound.

ByteRun1 -128 is a no-op and malformed streams can read indefinitely

beyond input; a run ending beyond the requested output is cut short

without structural validation. Caller-provided data must be complete and

correctly encoded.

The decompression workspace has width + 16 bytes available, but a

decoded row needs bytes-per-plane-row times total plane count. That

requirement is not checked and can exceed the buffer, including depth

eight plus a mask at some widths. Planar conversion also prefetches the

next source byte at eight-pixel boundaries, so it can read one byte past

the final required plane data. Dimensions use WORD iteration and

unchecked arithmetic. This implementation is not a bounds-checked

decoder for arbitrary BODY data.