sbcl Log


Commit Date  
[dc7f84] (sbcl_1_0_46) by Juho Snellman Juho Snellman

1.0.46: will be tagged as sbcl_1_0_46

2011-02-20 04:48:34 Tree
[f2c8e8] by Juho Snellman Juho Snellman

1.0.45.36: Fix solaris x86-64 page size issue

* Patch lp#720800 from Jim Wise

2011-02-18 01:12:02 Tree
[ad4fd2] by Nikodemus Siivola Nikodemus Siivola

1.0.45.35: revert 1.0.45.30 due to regression (lp#677779)

Reported by Ariel Badichi on sbcl-devel.

Using WITH-COMPILATION-UNIT to silence UNDEFINED-FUNCTION style-warnings for
forward-referenced functions was a bad idea after all: it caused cores dumped
using it to be started with **WORLD-LOCK** held by a dead thread.

...and it's not like WITH-COMPILATION-UNIT would ever get to summarize the
unit anyways the way it was used there.

Proper fix for this is to use UNDEFINED-REFERENCE-STYLE-WARNING as separate
condition class and just muffle those -- but this is too much to put in
during the freeze and will wait till after 1.0.46.

2011-02-16 10:35:57 Tree
[e55494] by Nikodemus Siivola Nikodemus Siivola

1.0.45.34: #+sb-thread the new waitqueue printing test

In despite the freeze as this is test-fix and a regression.

Thanks to Harald Hanche-Olsen for the heads-up.

2011-02-15 12:01:55 Tree
[ae469b] by Nikodemus Siivola Nikodemus Siivola

1.0.45.33: optimize CHARPOS on string-streams

The POSITION in STRING-OUT-MISC wasn't getting optimized at all due
to insufficient type-information.

This speeds up pretty-printing on string-streams somewhat.

2011-02-14 15:08:06 Tree
[838316] by Nikodemus Siivola Nikodemus Siivola

1.0.45.32: export SB-PCL:+SLOT-UNBOUND+

Using STANDARD-INSTANCE-ACCESS with this provided is much nicer --
no need to muck around with SLOT-BOUNDP-USING-CLASS and such.

Also add quick SIA tests and some dependant update abuse
to tests.

lp#718039

2011-02-13 20:34:14 Tree
[a51d83] by Nikodemus Siivola Nikodemus Siivola

1.0.45.31: make COPY-PPRINT-DISPATCH have access to a pristine table

...so that it can always be used to restore the original.

Fixes lp#678409.

Also add missing NEWS entry for last commit.

2011-02-13 20:29:21 Tree
[b92e0b] by Nikodemus Siivola Nikodemus Siivola

1.0.45.30: wrap --script loading in a WITH-COMPILATION-UNIT

Avoid UNDEFINED-FUNCTION warnings for code like

(defun foo () (bar))
(defun bar () 42)

by wrapping the LOAD done by --script in a WITH-COMPILATION-UNIT.

Fixes lp#677779.

2011-02-13 20:22:08 Tree
[1fdaa2] by Nikodemus Siivola Nikodemus Siivola

1.0.45.29: fix a copy and paste comment error

Patch by William Halliburton, lp#710024.

Plus missing NEWS entry for last commit.

2011-02-12 15:54:21 Tree
[a801d5] by Nikodemus Siivola Nikodemus Siivola

1.0.45.28: add LIMIT and PRINT-NO-CALL-LIST arguments to SB-PROFILE:REPORT

Patch by William Halliburton, lp#710017.

These arguments are used to control the amount of information printed
by REPORT as many times you are only looking for the top contenders
and the other information is unnecessary and, at times, inundating.

2011-02-12 15:51:47 Tree
[c77617] by Nikodemus Siivola Nikodemus Siivola

1.0.45.27: fix a typo in the manual

Patch by "adicarlo", lp#706139.

Correct name of special var that turns off beginner's help in
the debugger.

2011-02-12 15:49:21 Tree
[8b44c2] by Nikodemus Siivola Nikodemus Siivola

1.0.45.26: fix a long-standing UNINTERN bug

Patch by Stas Boukarev, lp#693796.

UNINTERN is specified to take a symbol, not a symbol
designator: if P1 has the symbol P1:S, which is not EQ
to P2:S, then (UNINTERN 'P2:S :P1) should not remove
P1:S from P1.

2011-02-12 15:45:38 Tree
[846630] by Nikodemus Siivola Nikodemus Siivola

1.0.45.25: better constant folding in arithmetic functions

Patch by Heka Deep, lp#676414, edited to retain identities
and added a random-tester.

* Folds constants in expressions such as (+ 3 a 5 b 7 c).

* Constants are collected by calling the `reduce-constants' function
from the `source-transform-transitive' and
`source-transform-intransitive' functions. Constants adding up to
identities are retained so that SNaNs don't sneak past.

2011-02-12 15:42:48 Tree
[b96ec9] by Nikodemus Siivola Nikodemus Siivola

1.0.45.24: minor MAKE-TIMER cleanups

Patch by Brit Butler, lp#672252.

* Improve the docstring for SB-EXT:MAKE-TIMER.

* Add a minor comment to unschedule-timer.

...with minor edits by yours truly, including dropping
the bogus SCHEDULE-TIMER change from the original patch.

2011-02-12 15:38:32 Tree
[9b122e] by Nikodemus Siivola Nikodemus Siivola

more pedantic clean.sh

Patch by Jim Wise, lp#666884.

Takes care of the few cases of temporary files which are currently
left behind after running clean.sh.

While there, it normalizes the name used for scratch files used in
contrib testing, and updates .cvsignore to match.

This is obviously pedantry, but it's nice to have a clean tree at
the end of clean.sh.

2011-02-12 15:22:25 Tree
[eb3a71] by Nikodemus Siivola Nikodemus Siivola

1.0.45.22: non-racy RUN-PROGRAM :PTY on OpenBSD

Patch by Josh Elsasser, lp#669485.

Lifts out the pty-setting logic into its own function
for clarity.

2011-02-12 15:18:59 Tree
[f82141] by Nikodemus Siivola Nikodemus Siivola

1.0.45.21: whitespace damage from ctor patches

Die tabs, die!

2011-02-12 15:12:01 Tree
[cb3058] by Nikodemus Siivola Nikodemus Siivola

1.0.45.20: optimize ERROR and CERROR

They were both surprisingly slow as we looked for a *STACK-TOP-HINT*
before calling SIGNAL.

The hint is needed (and was used) only for the INVOKE-DEBUGGER case,
however, so there is no need to pay that price if SIGNAL is enough
-- eg. when there's a handler.

Fixes lp#715191.

Also missing NEWS items for the CTOR hackery.

2011-02-11 17:41:46 Tree
[4cfe6b] by Nikodemus Siivola Nikodemus Siivola

1.0.45.19: more comprehensive CTOR optimization, part 2

Extend CTOR optimizations to cover classes with :AROUND
methods on INITIALIZE-INSTANCE.

Happily SBCL's CALL-NEXT-METHOD is implemented so that we can just
stick a function in the list of next-methods instead of mucking
about with MAKE-METHOD &co...

2011-02-11 17:34:46 Tree
[793d57] by Nikodemus Siivola Nikodemus Siivola

1.0.45.18: more comprehensive CTOR optimization

Allows using optimized constructors in the presence of (SETF
SLOT-VALUE-USING-CLASS) and SLOT-BOUNDP-USING-CLASS methods.

Simply generate calls to appropriate generic functions
instead of using CLOS-SLOTS-REF directly.

2011-02-11 17:30:50 Tree
[ca8135] by Nikodemus Siivola Nikodemus Siivola

1.0.45.17: give CONSTANTLY some love

Replace the source-transform with a DERIVE-TYPE optimizer: there are
essentially no cases where the source transform improves performance.

Make sure the out-of-line version has the right lambda-list.

Fixes lp#713626.

2011-02-11 17:25:55 Tree
[621e38] by Nikodemus Siivola Nikodemus Siivola

1.0.45.16: *features* tweaks for Solaris targets

A mistake in make-config.sh ("amd64" for "x86-64") was preventing
:sb-lutex from being enabled on Solaris x86-64.

While here, enable :largefile for Solaris (all platforms)

Patch by Jim Wise, lp#667297.

2011-02-11 17:20:10 Tree
[7717fe] by Nikodemus Siivola Nikodemus Siivola

1.0.45.15: make waitqueue printing prettier

Fixes lp#673630: just print the name, if any. The token is an
implementation detail, and can blow the stack if *print-circle* is
not set.

2011-02-11 17:15:26 Tree
[29dd8b] by Nikodemus Siivola Nikodemus Siivola

1.0.45.14: fix handling of multibyte character encoding errors

In encodings defined using DEFINE-MULTIBYTE-ENCODING, if there is no
translation for the character, report the size as 0 -- replacements
get their size accounted for elsewhere.

Fixes lp#713063.

2011-02-11 16:59:56 Tree
[3930f3] by Alastair Bridgewater Alastair Bridgewater

1.0.45.13: physenvanal: Treat all functions without XEPs as being D-X.

* In order for a function to be returned or passed as a parameter,
it must have an XEP.

* Functions without XEPs, therefore, can only be called directly
from within their lexical scope. They are, therefore,
dynamic-extent.

* But wait, you say, they could be called from a closure that is
not dynamic-extent, which clearly shows such an analysis to be false.

* It turns out that this doesn't matter, because the non-dynamic-
extent closure also has to close over the variables passed to the
supposedly-dynamic-extent closure, and that will cause explicit
value-cells to be allocated anyway.

* So, it's a bit of an abuse to say that the functions have dynamic
extent, but it does no harm (and quite a bit of good) to treat them
as if they do.

2011-01-21 16:40:53 Tree
Older >