Infinite recursion in TypeFormatterForModule::emit_type

[Swatinem]

It looks like we get some new customer pdb files that exhibit some infinite / mutual recursion that looks like this:

13  symbolicator                    0x555cbf11f67d      [inlined] pdb_addr2line::type_formatter::TypeFormatterForModule::emit_type_inner (type_formatter.rs:1130)
14  symbolicator                    0x555cbf11f67d      pdb_addr2line::type_formatter::TypeFormatterForModule::emit_type (type_formatter.rs:1105)
15  symbolicator                    0x555cbf11ee5a      pdb_addr2line::type_formatter::TypeFormatterForModule::emit_type_index (type_formatter.rs:1091)
16  symbolicator                    0x555cbf11e61d      pdb_addr2line::type_formatter::TypeFormatterForModule::emit_arg_list (type_formatter.rs:996)
17  symbolicator                    0x555cbf11f67d      [inlined] pdb_addr2line::type_formatter::TypeFormatterForModule::emit_type_inner (type_formatter.rs:1130)
18  symbolicator                    0x555cbf11f67d      pdb_addr2line::type_formatter::TypeFormatterForModule::emit_type (type_formatter.rs:1105)
19  symbolicator                    0x555cbf11ee5a      pdb_addr2line::type_formatter::TypeFormatterForModule::emit_type_index (type_formatter.rs:1091)
20  symbolicator                    0x555cbf11e61d      pdb_addr2line::type_formatter::TypeFormatterForModule::emit_arg_list (type_formatter.rs:996)

This then repeats for a long time and eventually crashes in alloc::vec::Vec<T>::with_capacity (mod.rs:483) and finally in jemalloc. Its a bit weird that we get a SIGSEGV / SEGV_ACCERR instead of an allocation panic, but here we are 🤷🏻‍♂️

I wasn’t able yet to figure out what symbol is causing this however.

Though I remember I was suggesting a general recursion / memory usage limit that is being checked for so that we do not run into things like this.

[mstange]

I'd be happy to accept a PR which adds a recursion limit.