Re: Bug#151886: [Gcl-devel] Re: gcl/maxima on hppa
Greetings!
Carlos O'Donell <carlos@baldric.uwo.ca> writes:
> On Fri, Jun 11, 2004 at 01:10:44PM -0400, Camm Maguire wrote:
> > Greetings! It is still the case that the si_addr field of the
> > siginfo_t struct is not filled in on hppa when SA_SIGINFO is used,
> > right?
>
> Please qualify. Kernel versions make a difference. If you are asking,
> "Did you fix all kernel version?" then the answer is no :)
>
I guess I meant on the buildd (sarti). GCL will test whether the
entry is filled in, and disable 'fast garbage collection' if not.
Everything still works, but slower. If the acl2 build, say, happens
on a machine running 2.6, fast gc will be turned on, and at execution
time, it will be left on unless the running kernel fails the tests, in
which case slow gc is again the default. If the build, however, takes
place on a machine mandating slow gc, the resulting binary will never
enable sgc when running on a more capable kernel, as there is no way
to tell if this was the case or the user just wanted the slower
algorithm.
So if/when sarti gets 2.6, it might be worth it to requeue maxima,
acl2, and axiom.
> I've tested almost all the siginfo_t structure entries and they work
> just fine with a recent 2.6 kernel. I'd have to go back to find the
> exact date I did the checkin, but after that it should just work.
>
Thank you so much for your work here!
Take care,
> c.
>
>
>
>
--
Camm Maguire camm@enhanced.com
==========================================================================
"The earth is but one country, and mankind its citizens." -- Baha'u'llah
Reply to: