Re: GDB with Hitachi Linker output?

[jtc@redback.com (J.T. Conklin)]

>>>>> "tm" == Toshiyasu Morita <tm@netcom.com> writes:
tm> The SRCONV utility is actually pretty broken, through no fault of
tm> Cygnus.  

If anything is at fault, I nominate the sysroff object format itself.
Sysroff is an incredibly baroque underspecified format.  I seem to
remember lots of little sections (I seem to recall a 256 or 512 byte
limit), unrelated fields in the same section, related fields in
different sections, undocumented section ordering, etc.

Even with the sysroff specification, a centimeter thick document with
PROPRIETARY stamped across each page, it was nearly impossible to
determine why one object would be accepted by the Hitachi tools and
another would not.  

There wasn't a rationale section in the spec, so I have no idea what
the designers of the format were thinking.  But I can't think of any
advantages sysroff has over a simpler object format, especially
considering its complexity.  That's perhaps the only advantage of
proprietary specifications, you can sweep any ugliness under the
carpet so no one can find it.

	--jtc

-- 
J.T. Conklin
RedBack Networks