FID hash mismatch between .o file from static lib and statically compiled binary

[f4c31e55]

I was attempting to make analysing a statically compiled binary easier by generating the FID database for the static library files. Some functions appear to work but not all. When I look at an example function which didn't work, the function in the .o seems identical (aside from relative jumps/calls) to that in both the stripped and unstripped binary but the FID hash is different. The FID has for the stripped and unstripped binaries match.

Is there a way I can debug why the FID hash is different when the function instructions appear identical?

An example function:
binary: FID Hash for FUN_005644c0 at 005644c0: FH: 76e1702ecb4e533f (27) +10 XH: c4e227edb35e00eb
binary-unstripped: FID Hash for endl<char,std::char_traits<char>> at 005644c0: FH: 76e1702ecb4e533f (27) +10 XH: c4e227edb35e00eb
.o: FID Hash for _ZSt4endlIcSt11char_traitsIcEERSt13basic_ostreamIT_T0_ES6_ at 00101290: FH: 11134784cd2ca6b4 (27) +10 XH: 14790637a7f0e9fc

[gemesa]

This does not answer your question but I have great success identifying functions with BSim instead of Function IDs. It can recognize similar functions which do not fully match.

[f4c31e55]

Thanks. Switching to BSim feels quite heavyweight when I already know these functions match.