[rfc] Code cleanup: Permit BIAS to be NULL
roland at hack.frob.com
Tue Oct 9 23:51:06 UTC 2012
I think the reason I made them non-optional is that it's such an easy bug
to write code that ignores the bias when it shouldn't. Often you'll have
to have some complex situation involving separate debuginfo and prelink to
actually notice such a bug, so it's doubly easy to leave such bugs around.
I think I'm still convinced by that logic for the public API. For internal
functions where it simplifies the code and we can be fairly confident we
are considering all the strange cases properly (and ideally testing them
all in the test suite), I wouldn't mind it.
More information about the elfutils-devel