Simutrans is not designed to support x86-64. Apparently there is some badly written/old code in it that is incompatible with x86-64 builds. Well it is compatible enough for them to work but it will break under certain conditions, such as pointers/references using memory past the 4 gigabyte range. This really should be fixed one day but considering multithreading and IPv6 is still "optional" I do not think it will happen soon.
I ran 64-bit Simutrans on Linux years ago, which I think was basically when compiling for 64-bit started working in the first place. I never had any crashes because of that. Nor do I remember seeing any code in Simutrans that casts pointers to something else and back (ptrhash_tpl is only one-way, although it failed at that, which has been fixed). The only problem I had with naughty code was an SSE alignment problem, which might actually have been reported as sigsegv, I don't remember exactly. Not that Simutrans is particularly 64-bit suited anyway, but there have been more bugs not related to it.
And I'm thankful threading is optional, because I trust the thread safety less than the 64-bit safety. (Plus I don't need any pthread dependencies, although MSYS2's mingw64 wants to push it on me anyway.)
Unfortunately, I'm not sure if we currently got a resident Mac developer. But as DrSuperGood writes, without the symbols, there is little to get out of it. At best, given the executable, we can figure out what kind of instruction it was that crashed, but it is likely an instruction used all over the place, giving no clue as to what in the source code it corresponds to.
