mirror of
https://github.com/esphome/esphome.git
synced 2026-10-22 05:46:12 +00:00
clydeandClaude Opus 5.5
6fddbd1d48
A JPEG had to be downloaded whole, into a buffer grown to the file size, and was then decoded in one call. When the server gives the size, the JPEG decoder now reads the download from the buffer itself and decodes a row of MCUs at a time, within a time budget per loop pass. The buffer only has to hold the headers, so buffer_size bounds the RAM used, and a slow link makes the decoder wait for the rest rather than decode past what has arrived. Downloads without a size keep the old path. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
