I haven't yet given it the due consideration. Even if the best
the moment is just to punt, there is probably a cleaner way to do it.
To begin with, I'm still waiting to hear where ET_REL files with
.zdebug_* sections come from.
They are not hard to prepare by hand. A robust toolkit should not
corrupt its own data just because they are unusual.
> Could a dubious enough input file include
> relocation instructions that cause less mellow errors -
> ie. overwriting data other than the .z* section contents?
Relocation is per-section, so that's not possible. (The relocate code will
already diagnose DWFL_E_BADRELOFF if an r_offset field points past the end
of the section, as some might since the compressed section has shrunk since
the relocs were generated.)
Does the code correctly treat the "end of the section" values as being
pre-compression vs. post-compression?