Demangler: bad Ghidra symbol name for some templated functions

[robojumper]

The script from #112 came up with "not great" names for some functions

e.g. baseID_Turn<10sStateID_c>__Fv_RC10sStateID_c demangles to sStateID_c const& baseID_Turn<sStateID_c>()

Excpected symbol name: baseID_Turn<sStateID_c>
Current symbol name: const&baseID_Turn<sStateID_c>

[elijah-thomas774]

Also came across this with CalcAnimationFVS<Q44nw4r3g3d27@unnamed@g3d_resanmchr_cpp@46CAnmFmtTraits<Q34nw4r3g3d18ResAnmChrFVS32Data>>__Q34nw4r3g 3d27@unnamed@g3d_resanmchr_cpp@FPCQ34nw4r3g3d16ResAnmChrFVSDataf_f becoming floatnw4r::g3d::anonymous::CalcAnimationFVS<nw4r::g3d::anonymous::CAnmFmtTraits<nw4r::g3d::ResAnmChrFVS32Data>>

[robojumper]

That one demangles to
float nw4r::g3d::(anonymous namespace)::CalcAnimationFVS<nw4r::g3d::(anonymous namespace)::CAnmFmtTraits<nw4r::g3d::ResAnmChrFVS32Data> >(nw4r::g3d::ResAnmChrFVSData const*, float), so I guess there's also some weird return type stripping that's not working correctly. The auto namespace creation is also not working great due to the namespace separator appearing in template arguments...

[elijah-thomas774]

I didnt even think about the namespace creation aspect, I was just thinking about the return type here xD