[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

bugs/comments in/on cmucl

here is david's earlier email to me.  the first part is relative to a
bug in the sunos port which prevents /tmp/empty being mmap'ed properly
if /tmp is a temporary file system (tmpfs).  i have a question in to
sun about whether this is a permanent bug or what.  

Date: Wed, 10 Jul 91 20:38:45 gmt
From: @sunic.sunet.se:davida%isil (David Axmark)
To: ted%nmsu.edu
In-Reply-To: <9107082058.AA02064@NMSU.Edu>
Subject: CMU-CL porting to sunos

ted@NMSU.Edu writes:
 > make sure that your /tmp is _not_ on a tmpfs partition.
 > if you can't do this, then you may be able to patch the binary you
 > have.  change the two occurences of the string /tmp/empty to another
 > file name that has the same number of characters and which is a world
 > writable file on a non-tmpfs filesystem.

Yes that fixed the problem. Thanks!

I have had some thoughts to diffing the kernel config files but I am glad
i don't have to.

I am still trying to build a new lisp.core from sources. Did you make any
special changes to the source before you built yours ? The biggest thing
i been bitten by until know was the hardcoded constant vm-page-size. Then
i changed this to 8192 and skiped hemlock and clx i was able to build a
new core.

This core starts ok but fails loading files.  For example then i try to
load a file consisting of a single newline i get the following.

* (load "foo.lisp")

Error in function DO-INPUT.
Error reading #<Stream for file "/home2/cmucl/sun4_mach/alpha/compiler/generic/\foo.lisp">: Bad address

  0: Return NIL from load of "foo".
  1: Return to Top-Level.

Debug  (type H for help)

0] backtrace

(ERROR "Error reading ~S: ~A"
       #<Stream for file "/home2/cmucl/sun4_mach/alpha/compiler/generic/foo.lis\p">
       "Bad address")
(DO-INPUT #<Stream for file "/home2/cmucl/sun4_mach/alpha/compiler/generic/foo.\lisp">)
(INPUT-CHARACTER #<Stream for file "/home2/cmucl/sun4_mach/alpha/compiler/gener\ic/foo.lisp">
(READ-PRESERVING-WHITESPACE #<Stream for file "/home2/cmucl/sun4_mach/alpha/com\piler/generic/foo.lisp">
(READ #<Stream for file "/home2/cmucl/sun4_mach/alpha/compiler/generic/foo.lisp\">
(SLOLOAD #<Stream for file "/home2/cmucl/sun4_mach/alpha/compiler/generic/foo.l\isp">)
(INTERNAL-LOAD #.(pathname "foo.lisp")
               #.(pathname "/home2/cmucl/sun4_mach/alpha/compiler/generic/foo.l\isp")

Compile on the same file or a empty file:

* (compile-file "foo")
segv_handler: No mapping fault: 0xf77e8014

Error in function MACH::SIGSEGV-HANDLER.
Segmentation Violation at #x10006C4.

  0: Return to Top-Level.

Debug  (type H for help)


This does not happen the i try to load an empty file. 

ldb also sometimes tells my that i have to big environment. I fixed this
by starting without a environment (env - my-ldb -core my-lisp.core). But
i dont understand why it cares.

I can compile/run funtions from the toplevel.

Any clues to thats going on. I would appreciate any ideas.

David Axmark
EMAIL:	davida@isil.detron.se or davida@emil.csd.uu.se
MAIL:	Detron HB, Petterslundsg 11 A, 753 28, UPPSALA, SWEDEN
PHONE:	+ (46) 18 - 11 07 80

(I can't spell in ANY language ...)