SBCL 1.0.22 released

[findinglisp]

Richard Kreuter made the announcement last week:

SBCL 1.0.22 has been released. Release notes:

changes in sbcl-1.0.22 relative to 1.0.21:
* minor incompatible change: LOAD-SHARED-OBJECT no longer by default looks
for the shared object in the current directory, but passes the native
namestring of the designated pathname to the operation system's shared
object loading function as-it.
* minor incompatible change: --disable-debugger toplevel option now takes
effect before processing of initialization files and --eval or --load
options.
* new feature: new commandline argument: --script, which supports
shebang lines. See documentation for details. (based on work by
Kevin Reid)
* new feature: SAVE-LISP-AND-DIE can save current values of
--dynamic-space-size and --control-stack-size in the executable core,
causing it to skip normal runtime option processing. See documentation
for details. (thanks to Zach Beane)
* enhancement: inoccous calls to EVAL or generic functions dispatching
on subclasses of eg. STREAM no longer cause compiler notes to appear.
* enhancement: the system no longer resignals errors from --load and
--eval toplevel arguments as SIMPLE-ERRORS, which caused restarts
associated with the original error to be lost. (thanks to Ariel
Badichi)
* enhancement: :DONT-SAVE keyword argument has been added to
LOAD-SHARED-OBJECT for controlling interaction with
SAVE-LISP-AND-DIE, and UNLOAD-SHARED-OBJECT can be used to undo
the effects of an earlier LOAD-SHARED-OBJECT call.
* bug fix: ADJUST-ARRAY on multidimensional arrays used bogusly give
them a fill pointer unless :DISPLACED-TO or :INITIAL-CONTENTS were
provided. (reported by Cedric St-Jean)
* bug fix: circularity handling in the reader did not treat raw
structure slots correctly. (reported by Cedric St-Jean)
* bug fix: SERVE-EVENT occasionally signaled an error about bogus
file descriptors when there were none.
* bug fix: DEFINE-COMPILER-MACRO support of destructuring lambda-lists
was broken. (reporteed by Willem Broekema)
* bug fix: SB-EXT:NATIVE-NAMESTRING on Windows did not work on
pathnames without a directory.
* bug fix: COMPILE-FILE did not warn about undefined variable
references at toplevel, and LOAD of the resulting fasl did
not signal an error.
* bug fix: functions with non-required arguments used to end up with
(SB-C::&OPTIONAL-DISPATCH ...) as their name.
* bug fix: redefining a function with non-required arguments didn't
update the system's knowledge about its call signature properly.
* bug fix: fixed #431; incompatible alien record type redefinitions
are detected and handled. (reported by Neil Haven)
* bug fix: using RESTRICT-COMPILER-POLICY with DEBUG 3 could cause
PROGV miscompilation. (reported by Matthias Benkard, patch by Juho
Snellman)
* bug fix: on x86 and x86-64 backtraces were sometimes truncated
after alien stack frames.
* bug fix: strings passed to mk{s,d}temp lacked null termination

Get it from SourceForge or CVS.

[donkey]

new feature: new commandline argument: --script, which supports
shebang lines. See documentation for details. (based on work by
Kevin Reid)

If this is what I think it is, I'm getting drunk tonight.

*a few seconds later*

Ok, I'm getting drunk

[dmitry_vk]

new feature: new commandline argument: --script, which supports
shebang lines. See documentation for details. (based on work by
Kevin Reid)

If this is what I think it is, I'm getting drunk tonight.

*a few seconds later*

Ok, I'm getting drunk

But the --script option breaks some things becacuse it prevents site-init (/etc/sbclrc) and user-init (~/.sbclrc) from being loaded. E.g., on Gentoo using it breaks asdf.

[TheGZeus]

Dammit, why is Sid stuck at 1.0.18???????????

[ramarren]

But the --script option breaks some things becacuse it prevents site-init (/etc/sbclrc) and user-init (~/.sbclrc) from being loaded. E.g., on Gentoo using it breaks asdf.

This problem is easily solved by using as #! a custom core, with initfiles already loaded. It seems a good idea anyway to avoid reloading common libraries on every script invocation.

[dmitry_vk]

But the --script option breaks some things becacuse it prevents site-init (/etc/sbclrc) and user-init (~/.sbclrc) from being loaded. E.g., on Gentoo using it breaks asdf.

This problem is easily solved by using as #! a custom core, with initfiles already loaded. It seems a good idea anyway to avoid reloading common libraries on every script invocation.

Custom cores => NOT easily solved.

[ramarren]

Custom cores => NOT easily solved.

Well, I suppose if considered for end user deployment or something... but even then, compared to installing SBCL at all, it is easy. Just create a file make-utility-core.lisp,

Code: Select all

(cl:in-package :cl-user)

(when (probe-file "sbcl-utility")
  (delete-file "sbcl-utility"))
(require :iterate)
(require :alexandria)
(require :cl-ppcre)
(require :closure-html)
(require :drakma)
(require :cl-interpol)
;; and whatever else you want to use from --script's

(save-lisp-and-die "sbcl-utility" :executable t)

and then

Code: Select all

sbcl --load make-utility-core.lisp

and point #! to the created "sbcl-utility" file. You probably have known that, I just wanted to show this is not really a barrier for using --script. I haven't used sbcl scripts extensively yet, but the few I have work. I'm running Gentoo, with sbcl from lisp-overlay (with locally bumped version to 1.0.22), and have no problems.