Specify MALLOC_CAP_INTERNAL while calloc()ing heap

[SmittyHalibut]

https://github.com/NorthernMan54/rtl_433_ESP/blob/master/src/rtl_433_ESP.cpp#L409

I was getting calloc() failures, even though there was enough heap reported by ESP.getFreeHeap() and ESP.getMaxAllocHeap().

Those two reporting tools however are calling heap_caps_ functions, and hard coding a capability of MALLOC_CAP_INTERNAL to force reporting heap for Internal memory only.

I tested by detecting a calloc(...) failure, then trying a heap_cap_calloc(..., MALLOC_CAP_INTERNAL) which always succeeded.

So I propose replacing calloc(..) with heap_caps_calloc(..., MALLOC_CAP_INTERNAL)

Pull request incoming. :-)

[NorthernMan54]

For this issue, any thoughts on how I could recreate it? During testing I never ran into this.

I won’t be able to merge this for a few weeks, as I’m out of the country, and don’t have access to my test environment

[SmittyHalibut]

It came up for me when I enabled both ZgatewayBT and ZgatewayRTL_433 at the same time in OpenMQTTGateway. I suspect that BT is also using an inspecific malloc() scheme and it's defaulting to something other than INTERNAL that's smaller than INTERNAL.

[SmittyHalibut]

Its not confirmed yet, but 1technophile/OpenMQTTGateway#1529 MIGHT be the same problem I was running into yet. Their reported symptoms sound similar, after doing the same thing I did. Soooo, maybe chances are good that ticket is reproduction?

[NorthernMan54]

This has been implemented