Potential fix for #272

[TMRh20]

Ran the reassembly code through AI again and it came up with these code changes. Currently testing...

[2bndy5]

Makes sense in the sense that each platform allocates different sizes for data types. We've certainly seen divergences in ESP cores on this numerous times before. I suspect this sizeof(Header) + sizeof(Addr) math is done at compile time. Might be cleaner to have a constexpr to signify that this math is calculating the... whatchamacallit... FRAG_MSG_OFFSET?

[TMRh20]

Yeah, if this works, we can def do that

[TMRh20]

Been 6 days running w/no problems, going to give it a while longer:

***********Reset reason************: External System
Free: 31952 | Max Block: 31632 | Frag: 2%

Memory is holding steady and functionality is right on.

[TMRh20]

So after my last comment, there was a Hardware WDT timer reset the next day. Tempting fate by stating that since then, I made a slight modification to my code and have now been running for a new record of 8 days solid with no WDT timer resets or obvious exploits. Letting it ride...

[TMRh20]

Ok, 2 weeks now with nothing but stability.

05:23:38.877 -> ***********Reset reason************: External System
05:23:38.914 -> Free: 30928 | Max Block: 29832 | Frag: 4%

Now, once I find time, I will have to do a little more testing on the ESP32 platform instead of just the ESP8266, but the ESP32 doesn't seem as delicate as the ESP8266, so I expect things to work.