Jest-Image-Snapshot image labelling

ideas.md

Image Format

• Row 1, Col 1-3: RGB-encoded snapshot and version information.
• Row 1, Col 4: 3-byte representation of header height, encoded as 1 px dot of an RGB color
• Rows 2 - n: Header, plus a bottom border of arbitrary color
• rows n+1 - ? : actual image

Reading

1. Read color at row 1, column 1-3. This tells us if the existing snapshot used a label header and, if so, what version it used.
2. Read color at row 1, column 4. This tells us how big the header is.
3. Skip the number of rows we found in step 2.
4. The remaining rows are our actual image.

Writing

1. Write an image containing the label text, with a width matching the snapshot.
2. Get height of image from 1.
3. Create final image with width = width_snapshot and height = height_snapshot + height_label + 1 (for header row)
4. Encode current version info and height of # 1 into first row of final image
5. Copy label image into final image, rows 2 through (height_label + 1)
6. Copy actual snapshot into the rest of the final image