egcs-19980803 (pre-1.1) powerpc-ibm-aix4.1.4.0 regressions

Kaveh R. Ghazi
Sat Aug 8 21:11:00 GMT 1998

 > From: David Edelsohn <>
 > 	I found your initial report extremely difficult to parse.  I also
 > assume that when someone specifically CC's me as well as sending the note
 > to a public mailinglist, he or she is putting me on notice that action is
 > expected -- from me.  And, as we are approaching a release, he or she
 > expects me to do something about it immediately.

	Well yes, only I would change the word "expect" to "hope".  As
in, I hope/desire/request you do something.

 > 	As you originally reported this in mid-July, 

	I never got an ACK from that report.  Even something as simple
as verifying whether you see the same results would have been nice.  You
are of course free to refuse to fix things I report (as you have in the
past.) But until you actually acknowledge the report and/or make you
wishes clear on the matter, I have no way of knowing you saw it and in
general assume you missed it.  (Lots of maintainers, I won't list any
names :-), often admit to missing reports not addressed to them directly
due to list volume.) Since you didn't respond to the July report sent to
the list, I felt cc:ing you this time was appropriate. 

 > stating "The
 > following bootstrap error is a regression from prior releases/snapshots"
 > is very misleading.  This was not, in fact, a change from your previous
 > report.  If this had been clearer, I would not have been so incredulous
 > and would not have asked about possible external problems on your system
 > that might have caused this symptom.  Even so, asking the basic questions
 > is not unreasonable.

	Again, I sincerely apologize for ambiguity in my report.

 > 	As far as -mminimal-toc, this depends on which AIX linker patches
 > have been applied.  It *should not* be necessary, but still may be
 > necessary on your particular system depending on what patches have been
 > installed. 

	Okay, agreed.  The sysadmin of my system hasn't installed
these patches, so -mminimal-toc is needed.

 > 	I did not interpret your report as a normal regression report but
 > as a "kick the port maintainer" note which actually contained misleading
 > information.  Maybe you should consider how your messages and distribution
 > lists might be interpreted.
 > David

	David, I never meant to "kick" you.  I only thought that as an
rs6000 maintainer, you would be interested (perhaps even grateful?)
for a problem report against a platform for which you have assumed
responsibility.  What you do with it is up to you.  But I don't assume
you've read prior reports unless you respond to them.

Note I cc:ed meissner too since the MAINTAINERS file lists him for
rs6000 in addition to you.  On reports for other platforms, I cc:ed
the people listed as maintainers for those ports.  So far this seems
to have worked well with other maintainers.  I am happy to accommodate
your particular wishes in this area and behave differently with
respect to you.  All I ask is that you make your wishes clear.

	I've conceded my report could have been better worded.  But
the bug is quite real.  So let's get back to making egcs a better
compiler.  Peace, okay? :-)

Kaveh R. Ghazi			Project Manager / Custom Development		Icon CMT Corp.

More information about the Gcc-bugs mailing list