Re: Unreliable BFD caching heuristic

>>>>> "Pedro" == Pedro Alves <[email protected]> writes:

>> Do you have any proposals on ways to improve this heuristic?

Pedro> Compare st_ino/st_dev, and don't share if the system doesn't
Pedro> provide meaningful bfd_stat data?

Pedro> symfile.c:separate_debug_file_exists does this already,
Pedro> and then does a CRC check if all else fails.  Not sure
Pedro> whether the CRC part would be a good idea here.

Using the nanosecond information given by stat would help, at least on
platforms where this exists.

Tom