Hi all,
It'd be nice if we can make a stability release. I propose to merge the
attached list of changesets. IMO it's reasonably conservative, I avoided
adding large non-bugfix changes.
Cheers,
Peter
--
http://sjamaan.ath.cx
commit c2ea63b340995b1882d9ed400db0019c1686aa11
Author: felix <fe...@call-with-current-continuation.org>
Date: Fri Dec 28 00:13:27 2012 +0100
Corrected behaviour for "C_i_foreign_[unsigned]_integer64_argumentp"
Extract floating-point values from argument and compare with MIN/MAX
for the associated C type). Added limits to chicken.h, which uses stdint.h
now (or inttypes.h on SunOS).
Disabled compiler-test added by Peter for #955 for 32-bit platforms.
Signed-off-by: Peter Bex <peter....@xs4all.nl>
commit 62aac3554955e36fbe31ad7b13846b712805f612
Author: Peter Bex <peter....@xs4all.nl>
Date: Thu Jan 3 23:02:33 2013 +0100
Update copyright year to 2013. This time I used a slightly modified version
of last year's oneliner, to catch banner.scm and README as well:
for i in `git grep -Il '(c) [0-9-]*, The Chicken Team'`; do sed -r
's/\(c\) (([0-9]+)-|(2012))(2012)?,/(c) \2\3-2013,/' "$i" > tmp && mv tmp "$i";
done
banner.scm had to be slightly changed (spaces and commas) in order to get
detected by this script.
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit fe4c184e20d358d2c9befb20a3d53c729b58fc86
Author: Moritz Heidkamp <mor...@twoticketsplease.de>
Date: Wed Jan 2 22:02:10 2013 +0100
Fix TO argument check in subvector procedure
The subvector procedure checked its TO argument as if it was
inclusive. However, its default value is the passed vector's length so
it must be exclusive which also corresponds with the implementation. In
effect, the erroneous check prevented producing a subvector including
the given vector's last element. This patch changes the TO argument
check to be exclusive and adds some test cases for the subvector
procedure.
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit 67ca64e751451880c89af7c7aa0a233ded34ed09
Author: felix <fe...@call-with-current-continuation.org>
Date: Sat Dec 29 23:19:27 2012 +0100
From: Kon Lovett <konlov...@gmail.com>
Date: Sat, 29 Dec 2012 11:41:51 -0800
Subject: stop clang warnings
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit 8ed1ff65f31628b215b435dfd03d6f1ee3ae4b0b
Author: felix <fe...@call-with-current-continuation.org>
Date: Sat Dec 29 23:09:15 2012 +0100
pass "-z origin" to freebsd linker only if "-deployed" is given;
explained -deployed somewhat (csc help output).
commit eb9c5c6a07e423381f56c38b1756127370555d3c
Author: felix <fe...@call-with-current-continuation.org>
Date: Sat Dec 29 23:04:44 2012 +0100
disable integer64 test on 32-bit platforms
commit bbbbc0be8fb56a1c2e6dc2e37a4d582c780da518
Author: Peter Bex <peter....@xs4all.nl>
Date: Sun Dec 23 17:51:41 2012 +0100
Add regression test for #955
commit 1117e01239b1181b0d6ee66d998eb1e83dbd8285
Author: Peter Bex <peter....@xs4all.nl>
Date: Sun Dec 23 17:20:49 2012 +0100
Change one of the duplicated ##sys#foreign-unsigned-integer64-argument
definitions to ##sys#foreign-unsigned-integer-argument (copy/paste mistake?)
commit 42c869949f893baba40c9b0d5e4832eb2004b20e
Author: felix <fe...@call-with-current-continuation.org>
Date: Sun Dec 23 16:08:31 2012 +0100
Bugfix for foreign-argument-check routine for 64-bit unsigned longs. The
function returned a raw floating-point value (coerced to C_word) instead of the
original number.
Reported by Kon Lovett.
Signed-off-by: Peter Bex <peter....@xs4all.nl>
commit 4223b30426c257f3192d780b7d5e4de3a919dcfd
Author: felix <fe...@call-with-current-continuation.org>
Date: Sun Dec 23 16:05:58 2012 +0100
trivial formatting change
commit af7cc16b0f5986c758958b3b5d73885ad2f4217c
Author: felix <fe...@call-with-current-continuation.org>
Date: Sun Dec 23 16:05:43 2012 +0100
trivial change in describe-output
commit 6f95195f4840fe8e1342356b648036713a59e93e
Author: felix <fe...@call-with-current-continuation.org>
Date: Fri Dec 21 22:53:53 2012 +0100
register feature-identifier 64bit when running on a 64-bit CPU (suggested
by Kon Lovett)
commit 7818de84946260fabbaffbc971ac77583832ce83
Author: felix <fe...@call-with-current-continuation.org>
Date: Thu Dec 13 20:27:09 2012 +0100
handle case in make-pathname when the directory-argument is #f and the
file-argument begins with a path separator (reported by kon)
Fixes #959
Signed-off-by: Mario Domenech Goulart <mario.goul...@gmail.com>
commit 564285c33e7a1451084e0ba809d6efd29ae2dab1
Author: Jim Ursetto <zbignie...@gmail.com>
Date: Thu Dec 13 23:53:05 2012 -0600
manual: Note find-files old signature style was made invalid after 4.7.3
commit 709a4ea4e2fe3008c890d80a5b102bc62e9ba306
Author: felix <fe...@call-with-current-continuation.org>
Date: Thu Dec 13 20:15:06 2012 +0100
define keys only for scheme-mode
commit f016cd78979654ef43d7dbd90b0e96382274a1b4
Author: felix <fe...@call-with-current-continuation.org>
Date: Thu Dec 13 20:14:48 2012 +0100
some comments
commit 52c69531d586a39c51f3acc636596c65e181622e
Author: felix <fe...@call-with-current-continuation.org>
Date: Tue Dec 4 19:58:56 2012 +0100
bugfix for unsigned-integer64 type check
commit 0425b20986a049217c599870f7eed70c45cb3071
Author: Peter Bex <peter....@xs4all.nl>
Date: Wed Nov 28 23:57:10 2012 +0100
Irregex: Fixing folds on conditional begin patterns which aren't treated as
searchers. This is the final fix for #686 and synchronizes with upstream
version 0.9.2 (upstream changesets 01058fc79a16 and fad713187dbb)
Signed-off-by: Christian Kellermann <ck...@pestilenz.org>
commit 9d905e1acd87b32b744fc308891bd968bf97f10a
Author: Peter Bex <peter....@xs4all.nl>
Date: Tue Nov 27 20:40:04 2012 +0100
Irregex: Ensure 'bos is matched only once in irregex/fold; this fixes #686
and updates irregex to 0.9.1 (upstream changesets a6444839100c through
1741bbf14d3e)
Upstream changeset message for this bugfix: Don't bother looping in
irregex-fold(/chunked)/fast if the irregex is a searcher. Either it only
matches the beginning, or it already consumed the rightmost match.
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit 27e6458716b7078f8ade4d6d8ef0f5d6a8203eb1
Author: felix <fe...@call-with-current-continuation.org>
Date: Fri Nov 23 19:53:34 2012 +0100
do not make buildid and buildbranch .PHONY on Windows, or buildversion.scm
needs to be recompiled on every make(1) invocation
commit 9b0fac688337d292bb96965bf7c45bd3a74ff2c2
Author: felix <fe...@call-with-current-continuation.org>
Date: Fri Nov 23 19:52:42 2012 +0100
trivial change
commit 9eef92115fba6034a98cc21ef740fd82ea52387b
Author: Peter Bex <peter....@xs4all.nl>
Date: Sun Nov 18 21:03:51 2012 +0100
Fix select() buffer overrun vulnerability by using POSIX poll() on systems
that support it, leaving only those few systems vulnerable that don't (ie, only
Windows).
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit ef8c48db3fb656663e148eac0dad4f78a44af599
Author: Mario Domenech Goulart <mario.goul...@gmail.com>
Date: Mon Oct 15 19:56:05 2012 -0300
install-extension fix for when the second arg contains pairs (SOURCE DEST)
Cases like
(install-extension 'foo '(("orig" "dest")))
break install-extension. See
http://tests.call-cc.org/master/linux/x86/2012/10/15/salmonella-report/install/crunch.html
for example.
Signed-off-by: Peter Bex <peter....@xs4all.nl>
commit 04ef9ef5401d25e40e61f96aeb2f4ccd473a4772
Author: felix <fe...@call-with-current-continuation.org>
Date: Sat Nov 17 23:57:46 2012 +0100
das auge isst ja schliesslich mit...
commit 33c484028ff76bad3143296a31077bd5d038f95a
Author: Peter Bex <peter....@xs4all.nl>
Date: Sat Nov 10 18:39:50 2012 +0100
Add regression test for #952
commit dd46f2348c2c8f86e5358a6e376483bdb7690446
Author: felix <fe...@call-with-current-continuation.org>
Date: Sat Nov 10 14:10:56 2012 +0100
Improve type-matching for list-like types.
canonicalization of list-like types uses "(list ...)" where
appropriate to allow more precise type-matching of combinations
of "list-of"/"list" types.
Signed-off-by: Peter Bex <peter....@xs4all.nl>
commit 84e997b979393600323c19d2149c39c88b4ea71e
Author: Peter Bex <peter....@xs4all.nl>
Date: Fri Nov 9 21:55:24 2012 +0100
Mark all unused procedures (or replaced by compiler-macro) in irregex to
get rid of compile-time warnings
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit 10c77f8c18a569d5f34d9abc558311c38c62dda9
Author: Peter Bex <peter....@xs4all.nl>
Date: Fri Nov 9 21:54:10 2012 +0100
Irregex: Remove bogus pair? checks; they were testing the wrong bit of the
list, and sre-sequence will already return 'epsilon when the passed list is
empty (upstream changeset 4f257a9c3576)
This gets rid of a scrutiny warning
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit 3e72cecc8891e43a998787144d6bcb2757e4f652
Author: felix <fe...@call-with-current-continuation.org>
Date: Sun Nov 4 17:58:12 2012 +0100
Scrutinizer fixes.
a) when matching "list-of"/"vector-of" with "list"/"vector", each
element of the latter must match the element-type of the former
(reported by megane, fixes #948)
b) when matching result-types, allow "undefined" to match "noreturn"
as the "noreturn" property can not be inferred for foreign procedures
(for example) in general
Signed-off-by: Peter Bex <peter....@xs4all.nl>
commit 1b499f9022ca5650bea2510eb1f061e5997e98a1
Author: Peter Bex <peter....@xs4all.nl>
Date: Sat Nov 3 21:47:11 2012 +0100
Simplify and fix fast_read_string_from_file so it will correctly keep
reading when interrupted after having read 0 bytes (by always clearing errors)
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit f97a4e630c1390372c168327ee97838f633d0eac
Author: felix <fe...@call-with-current-continuation.org>
Date: Fri Nov 2 10:50:30 2012 +0100
Make symbol-GC more reliable by
a) clear weak-entry-table on heap-resizing to avoid keeping stale pointers
to the old heap-space(s)
b) use a randomization value recomputed on every GC to reduce the
probability
of unresolvable hashtable collisions which would cause some values never
to be reclaimed
c) only enter symbol objects in the w-e-t - which is, after all, kind of
obvious ...
This change seems to recover all unused symbols, both in interpreted
and compiled code.
Signed-off-by: Peter Bex <peter....@xs4all.nl>
commit b5472d45310ebb1838d6b2b0534bcf31d5857a8a
Author: felix <fe...@call-with-current-continuation.org>
Date: Fri Nov 2 21:34:20 2012 +0100
assert: only add form as argument to error if no message is given
commit 7bef21de09d279b0413ca94dbef978665c7a0f1c
Author: Peter Bex <peter....@xs4all.nl>
Date: Wed Oct 31 22:27:01 2012 +0100
Fix #944 by making the behvior of macro-renamed definitions inside modules
similar to the behavior at toplevel; they unhygienically introduce identifiers
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit 42c6071ada87135a3eca87defd2af061ff1a21b7
Author: felix <fe...@call-with-current-continuation.org>
Date: Wed Oct 31 23:09:26 2012 +0100
removed obsolete definitions
commit 20b7d9ef62480b43a64e04b7381d9cf62e6823a7
Author: felix <fe...@call-with-current-continuation.org>
Date: Mon Oct 29 14:48:45 2012 +0100
comments
commit bc09c5580b3dbd76810bc4cd26c89e0d455c24fd
Author: felix <fe...@call-with-current-continuation.org>
Date: Sun Oct 28 13:29:28 2012 +0100
remove implicit dependency on tcp.scm from posix-common.scm
commit 7cc25c4756558b47e0e695700a7a8217f0c74a14
Author: felix <fe...@call-with-current-continuation.org>
Date: Sat Oct 27 17:25:18 2012 +0200
fixed incorrect signature of _exit in types.db
commit 1ae287896160a006ea648b1234a20553f28d618d
Author: felix <fe...@call-with-current-continuation.org>
Date: Sun Oct 21 20:49:13 2012 +0200
added missing Types.html to manifest (reported by megane)
commit 8181fe9b6a29a40dcffbe3389a8f7bd04564c445
Author: felix <fe...@call-with-current-continuation.org>
Date: Sun Oct 21 20:48:48 2012 +0200
new is a C++ keyword
commit f3aa0f2d861cbafb0c7455da3de219a5be27a280
Author: Peter Bex <peter....@xs4all.nl>
Date: Fri Oct 5 20:20:31 2012 +0200
Convert (errno == EINTR && !feof(fp)) checks into ferror(fp) checks, and
put the errno dispatching code into Scheme, raising an exception on anything
but EINTR
Signed-off-by: Christian Kellermann <ck...@pestilenz.org>
commit 59926ffe6928530a683be56009efc78cdcf66e90
Author: Evan Hanson <ev...@thunktastic.com>
Date: Thu Oct 11 22:31:36 2012 -0500
respect -output-file option when compiling from stdin
Signed-off-by: Peter Bex <peter....@xs4all.nl>
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit c67b2bc55d4fc3eb145110634f8cf528e1b6ad40
Author: felix <fe...@call-with-current-continuation.org>
Date: Fri Oct 12 15:08:07 2012 +0200
added missing tests
commit 1c6f37c00ffd043fe187f6b5a73d589d780bcada
Author: felix <fe...@call-with-current-continuation.org>
Date: Wed Oct 10 22:27:06 2012 +0200
removed obsolete file from helper script
commit 95e32b229cc982113b1dab290761b95d8761bd3f
Author: Peter Bex <peter....@xs4all.nl>
Date: Sat Oct 6 18:26:32 2012 +0200
Fixed scrutiny db type entry for signal-handler: it will return #f when no
handler was installed
Signed-off-by: Mario Domenech Goulart <mario.goul...@gmail.com>
commit a1b972dfe1d9cc238b5fcddc577d357c355d3bbe
Author: Peter Bex <peter....@xs4all.nl>
Date: Fri Oct 5 21:41:54 2012 +0200
standard-extension uses -O3, not -O2, which can make a big difference
commit 0f4a0a3d1f2f3a98fb5c9e03255345f8b1d7aace
Author: Peter Bex <peter....@xs4all.nl>
Date: Wed Oct 3 22:22:55 2012 +0200
Fix handling of EINTR in process-wait by retrying. Add combined test for
this and the getc() EINTR handling bug
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit 6e6afed0e1cd2f57958c126e6192aed3f0f7c8bb
Author: Peter Bex <peter....@xs4all.nl>
Date: Wed Oct 3 21:00:52 2012 +0200
After calling getc() and getting back EOF, properly check feof() before
checking errno; if errno is not cleared we'd get in an endless loop on an
interrupted system call. These are quite rare, that's why this wasn't noticed
before. Also copy the bit already read when looping, after receiving EINTR in
read-line
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit 96c5b48e107a0353902e0eca4ea2eed21d99d97f
Author: felix <fe...@call-with-current-continuation.org>
Date: Thu Oct 4 20:36:20 2012 +0200
small fix in identify.sh, originally by Mario
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit 8efd0784e189c0fe0434a5e1d16c25cff4c827e0
Author: felix <fe...@call-with-current-continuation.org>
Date: Thu Oct 4 20:34:08 2012 +0200
scrutinizer and types.db fixes by sjamaan
Signed-off-by: felix <fe...@call-with-current-continuation.org>
commit 0cca92ea9ca8dfef2b2edafa95948e1afd511beb
Author: felix <fe...@call-with-current-continuation.org>
Date: Mon Oct 1 19:47:54 2012 +0200
Copy-propagation of global that refers to intrinsic
If a variable is known to be bound to a global variable which itself
refers to an intrinsic function (one of the "standard-" or "extended"
bindings, that get specifically treated in call position by the
compiler), than references to the former variable will be replaced
with the latter.
Signed-off-by: Peter Bex <peter....@xs4all.nl>
commit 4e5ba39639abe206f2e8cf0e2a7647d00e759562
Author: felix <fe...@call-with-current-continuation.org>
Date: Tue Sep 25 21:19:35 2012 +0200
Do not clear eof-status of input port when reading from file-port.
Calling "(read-string #f)" or "(read-all)" from csi in a terminal will
not detect EOF, waiting forever for more input. This is caused by
"C_fast_read_string_from_file", which calls clearerr(3) to clear the
error/eof status on encountering EOF. But when the input port is
connected to a terminal, input may still follow and the EOF status
will be lost. The patch removes the call to clearerr(3) and adds
an additional EOF check at the beginning. I'm not 100% sure about
this fix, but it seems to do the job.
Signed-off-by: Christian Kellermann <ck...@pestilenz.org>
commit 49c1af45fdce3881076c7224cb25ceb514c1c0a1
Author: Mario Domenech Goulart <mario.goul...@gmail.com>
Date: Sat Sep 22 13:57:04 2012 -0300
csc.scm: add "-z origin" to linker options in FreeBSD
FreeBSD's linker require tath option to process $ORIGIN. Without this
patch, the deployment test (make check) fails.
Signed-off-by: felix <fe...@call-with-current-continuation.org>
_______________________________________________
Chicken-hackers mailing list
Chicken-hackers@nongnu.org
https://lists.nongnu.org/mailman/listinfo/chicken-hackers
