sbcl Log


Commit Date  
[27d1f9] by Nikodemus Siivola Nikodemus Siivola

git: that terminal output? make it mention the correct revision

*sigh*

2011-06-10 10:55:03 Tree
[cf1edb] by Nikodemus Siivola Nikodemus Siivola

git: more terminal output from post-receive-email script

2011-06-10 10:47:47 Tree
[b0f344] by Nikodemus Siivola Nikodemus Siivola

git: revered merge detection logic in post-receive-email script

How hard can this be?

2011-06-10 10:46:04 Tree
[c5fb0e] by Nikodemus Siivola Nikodemus Siivola

sb-introspect: test-cases for defstruct copier, predicate, and accessor sources

Feh. There were tests for even the cases that have worked for quite a while
now.

2011-06-10 10:26:02 Tree
[a27839] by Didier Verna Didier Verna , pushed by Nikodemus Siivola Nikodemus Siivola

sb-introspect: source locations for structure copiers

Closely paralleling the existing hacks for structure accessors
and copiers.

2011-06-10 10:10:27 Tree
[bddb38] by Nikodemus Siivola Nikodemus Siivola

robustify debugger against bogus lambda-lists

If we don't find a list where we expect a rest-list to be,
substitute a dummy unprintable object.

Provides a workaround for lp#795245.

2011-06-09 21:06:37 Tree
[11cfbf] by Jim Wise Jim Wise

RANGE-REDUCTION also fails-on sunos x86-64.

2011-06-09 20:51:58 Tree
[4d0b87] by Nikodemus Siivola Nikodemus Siivola

sb-posix: make SYSCALL-ERROR's argument optional

The symbol is exported, and adding a required argument to the call broke
backwards compatibility.

Oops.

2011-06-09 18:05:46 Tree
[c90008] by Nikodemus Siivola Nikodemus Siivola

git: adjust whitespace in commit emails

Kill the extra empty line.

2011-06-09 15:16:35 Tree
[0b51e9] by Nikodemus Siivola Nikodemus Siivola

git: post-receive-email hook again

Check all revisions from new to old for merges, not just the topmost.

2011-06-09 15:02:28 Tree
[6345a7] by Nikodemus Siivola Nikodemus Siivola

git: update hook, fix post-receive-email

Add an update hook that prevents merge commits on master.

Fix the post-receive-email script to work properly with merge commits --
don't try to report pushes of merges with one email per commit.

(Since we deny non-fast-forwards, we pretty much need to allow merges on
other branches, or give up keeping them in the main repo.)

2011-06-09 14:39:23 Tree
[b65e65] by Jim Wise Jim Wise

Fix thinko in one of the Solaris-compat changes in last commit.

2011-06-09 12:56:47 Tree
[9888f4] by Jim Wise Jim Wise

Merge branch 'master' of ssh://sbcl.git.sourceforge.net/gitroot/sbcl/sbcl

2011-06-08 20:28:28 Tree
[0e3fa6] by Jim Wise Jim Wise

Tweak to work with SunOS /bin/sh.

2011-06-08 20:27:40 Tree
[4084b6] by Nikodemus Siivola Nikodemus Siivola

armload of DEFINE-HASH-CACHE changes

* To clear a cache, drop the entire vector instead of filling it with
NILs: thread safe, less work, and doesn't add dirty pages to old
generations.

Entering a value after the cache has been dropped allocates a new
one.

Caches are now initialized with 0 instead of NIL -- faster to
allocate.

* Use DEFGLOBAL instead of DEFVAR.

* SAVE-LISP-AND-DIE drops all caches.

* UNSAFE-CLEAR-ROOTS drops caches depending on the depth of the
collection: nursery collection keeps all caches, gen 1 collection
drops the CTYPE-OF cache, gen 2 and deeper collections drop all
caches.

2011-06-07 20:18:57 Tree
[9ca5f2] by Nikodemus Siivola Nikodemus Siivola

minor bug in type= method for arrays of unknown element type

If the type has been specified since our last encounter with it,
reparsing the previously unknown type specifier can result in TYPE=
saying NIL, T.

Hard to provoke without dipping into internals, so no NEWS entry.

2011-06-07 17:07:58 Tree
[26627f] by Nikodemus Siivola Nikodemus Siivola , pushed by Nikodemus Siivola Nikodemus Siivola

more conservative bounds in FP interval arithmetic

Make BOUND-BINOP return closed intervals when floating
point rounding can cause an open bound to close.

Consider:

;; Can return zero.
(defun fii (x)
(declare (type (single-float (0.0)) x))
(/ x 2.0))

;; Can return 2.0.
(defun fii (x y)
(declare (type (single-float 2.0) x)
(type (single-float (0.0)) y))
(+ x y))

2011-06-07 12:27:54 Tree
[8573e1] by Nikodemus Siivola Nikodemus Siivola

mark RANGE-REDUCTION test as failing on x86-64 Linux and Darwin

Maybe others too?

2011-06-07 12:26:47 Tree
[69fe69] by Nikodemus Siivola Nikodemus Siivola , pushed by Nikodemus Siivola Nikodemus Siivola

fix whitespace in tests

2011-06-07 12:25:47 Tree
[6372f5] by Nikodemus Siivola Nikodemus Siivola

missing NEWS entry for REMOVE-FD-HANDLER restart

2011-06-07 11:51:45 Tree
[f9b9b7] by Jim Wise Jim Wise

Convert to with-test throughout.

While here, on SunOS x86-64 disable three tests which currently dump core.

With this, SunOS (x86-64 and x86) has no more tests which crash but are not
marked broken.

2011-06-07 02:40:57 Tree
[83a896] by Jim Wise Jim Wise

Run start-test earlier in with-test.

This way, if we bail on a test early because it is disabled, we've still
counted the test and recorded the source file.

Without this, if the first test in a file was disabled, it would be recorded
with the previous source file.

2011-06-07 02:27:19 Tree
[4191c3] by Jim Wise Jim Wise

Tweak to work with SunOS 10's ancient sh implementation.

2011-06-06 17:30:29 Tree
[4c81c6] by Jim Wise Jim Wise

Rework test infrastructure to keep track of tests which are disabled
on the platform/feature combination being tested, and to differentiate
between tests disabled because the test is broken in some way, and those
skipped because the platform in question is broken in some way (or the test
is irrelevant for the given feature set).

This takes the form of two new keywords to WITH-TEST -- :BROKEN-ON, which
notes that the test itself is broken for a given feature expression, or
:SKIPPED-ON, which skips a test for a given feature expression. This
information is noted at the end of a test run like so:

Finished running tests.
Status:
Skipped (broken): debug.impure.lisp / (TRACE ENCAPSULATE NIL)
Skipped (broken): debug.impure.lisp / (TRACE-RECURSIVE ENCAPSULATE NIL)
Expected failure: packages.impure.lisp / USE-PACKAGE-CONFLICT-SET
Expected failure: packages.impure.lisp / IMPORT-SINGLE-CONFLICT
(38 tests skipped for this combination of platform and features)
ok
//apparent success (reached end of run-tests.sh normally)
Thu Jun 2 15:59:31 EDT 2011

Note that there is no :WORKS-ON or :ENABLED-ON, even though many of the
read-time conditionals this replaced were for a given feature instead of for
its absence -- you can still do, eg:

(with-test (:name foo :broken-on '(not :x86)) ...)

but such declarations are almost always too general (one exception being
`:skipped-on '(not :sb-thread)'), and IMO, should be discouraged.

While here, re-enable a bunch of tests previously skipped on Solaris which
now work.

2011-06-06 17:22:34 Tree
[0fd1ec] by Nikodemus Siivola Nikodemus Siivola

check that git supports --count with rev-list

git rev-list --count is probably the newest git feature we use in
the build, so check that it works and complain if not.

2011-06-06 16:13:06 Tree
Older >