Mathieu Carbou

I was able to reproduce. This is caused by a commit that reworked how bytes are sent over the wire to improve perf ((release notes: https://github.com/ESP32Async/ESPAsyncWebServer/releases/tag/v3.9.0)), but sadly it seems...

A fix has been releases: https://github.com/ESP32Async/ESPAsyncWebServer/releases

FYI: https://github.com/ESP32Async/ESPAsyncWebServer/issues/343

> Thank you, this is finally a good async server that does not hang the ESP32 :) What are the ones you tried ?

> But even with this I'm getting out of RAM pretty quickly and it gets worse over the time (maybe some memory is not freed correctly in some cases). Ok!...

yes indeed... would you mind sending a PR ? Your solution makes sense I think and is also as per the doc. I will merge your PR after. Note: you...

@hoeken : the bug is line 110 in v2 in `PsychicUploadHandler.cpp` `index += received;` received can be negative in case the esp-idf api returns an error code. I have opened...

> but the endpoints were not loading/responding on browser. HTML/CSS was not loading like it was on the v1.2.1 Maybe this warrants looking into... It might be a bug in...

Ah, no this is different... In the multipart there is a difference in the code: `_itemSize - _itemBufferIndex`, that is the only place I would see an issue... I did...

FYI, please read: - https://github.com/mathieucarbou/ESPAsyncWebServer?tab=readme-ov-file#dependencies - https://github.com/espressif/arduino-esp32/issues/10526 If you use the right dependencies this is fixed, until the author merges back the changed from our community maintained fork into his...