Fix workflow for latest ESP-IDF for espressif examples

[gojimmypi]

Description

The Espressif examples began failing today in GitHub workflow with the "docker latest" image. (un-released ESP-IDF v6.0, latest stable is v5.5)

The core compiler problem was some diagnostic memory tools. As a brute-force solution to get workflow working again, I've simply disabled the features for ESP-IDF v6.0 in the esp_sdk_mem_lib.c file.

Unexpected, and not reproduced locally - was a massive number of linker errors:

dangerous relocation: call8: call target out of range: memset

... as seen in one of the prior workflows. This was resolved with conditional CMake criteria for only the Xtensa devices and only on the ESP-IDF v6.0 using a -mlongcalls compiler options: once for the main application, and then again for all components.

Post-Merge Update Notes

I've confirmed my local idf.py build does add the -mlongcalls in numerous places in the build directory. I've not yet determined why GitHub actions needed to have it added manually in this PR CMakeLists.txt project files only for IDF v6.0.

Fixes zd# n/a

Testing

See my https://github.com/gojimmypi/wolfssl/actions/workflows/docker-Espressif.yml

Checklist

• added tests
• updated/added doxygen
• updated appropriate READMEs
• Updated manual and documentation