Re: build failure on Irix 6.5/gcc

[Dave Love]

"Richard M. Stallman" <address@hidden> writes:

> I appreciate your work.

I'm afraid I couldn't have guessed so, and experimentally I couldn't
get fixes made, even being thick-skinned about the response.

> If you send a useful and clear bug fix, we
> will certainly use it.

I can't be any clearer than saying that you need to treat the sections
I mentioned like .sdata in unexelf and copy the stack-marking
definitions from irix6-0.h to irix6-5.h.

> At times when we want new features, we would
> be glad accept them from you too (assuming we like the features).

The last time I offered a TODO item, the story was there was a feature
freeze, but all sorts of features were going in from others.  Anyway,
I would rather that features I wrote several years ago looked like
getting released someday.  It badly needs a feature freeze to
stabilize it and get a release made.  I'm actually surprised people
can work on improvements when you continually have to address
regressions.

> Did you send a patch for this that we neglected?

I don't want to send a patch since I've been banned from working on
free software on lab systems even in my own time, and I wouldn't want
contributions to be in question even though I was addressing something
affecting users.  I'll be out of work soon, but I won't have the
proprietary systems available then.  However the changes are trivial
to make.  I since noticed the unexec ones are even in TODO.