[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[MIT-Scheme-devel] Fedora Core 3
From: |
Chris Hanson |
Subject: |
[MIT-Scheme-devel] Fedora Core 3 |
Date: |
Fri, 18 Mar 2005 14:50:58 -0500 |
User-agent: |
IMAIL/1.21; Edwin/3.116; MIT-Scheme/7.7.91 |
Date: Fri, 18 Mar 2005 04:50:44 +0000
From: Matt Birkholz <address@hidden>
Me too, on Debian Testing. I assume you, like me, pulled the HEAD CVS
revisions and tried to build from sources. I got a seg fault like
this:
>> A hardware fault has occurred within critical section "garbage
>> collector daemon".
>> [signal 11 (SIGSEGV), code 1]
>> Successful recovery is unlikely.
I eventually recalled that runtime.com contains microcode-specific
things like primitive-procedure offsets.
That's not really a problem, since runtime.com can be rebuilt from
binaries.
The real issue is that there have been a couple of low-level
representation changes since 7.7.90, so the current sources are
completely incompatible with the old binaries.
As I said, I'm happy to provide working snapshots.
I would love to hear how Chris bootstraps bands for a new microcode.
While skimming bintopsb.c I noticed options called "allow_cc" and
"allow_bands". Maybe I can translate a compiled runtime.com...
I haven't used those programs in a long time.
The latest round of changes involved some complicated staging with
special tweaks to the microcode. I had to work with interpreted code,
and compile the system through 4 stages before I got one in which #F
and '() were properly separated. Once that was done, I used the
compiler (interpreted, of course) to compile that code to x86 native.