https://sourceware.org/bugzilla/show_bug.cgi?id=32846
--- Comment #7 from Michael Matz <matz at suse dot de> --- Ultimately first_hash was introduced for what really is a bug in the compiler (pr31482, where the LTO .o files don't contain a reference to abort in their symbol table). Introducing that doesn't completely work around the compiler bug (my previous comment), but caused the bug I fixed with commit 2707d55e (where no compiler problem is involved). That in turn caused the bug reported here (also without compiler bug involvment). So, one course of action would be to revert the whole first_hash introduction to be in 2.43 state and instead fix the compiler bug. But maybe we can salvage some pieces, so I think we needn't do that yet. -- You are receiving this mail because: You are on the CC list for the bug.
