Age | Commit message (Collapse) | Author | Files | Lines |
|
egg info files in non-egg python distributions.
Add a comment that PY_NO_EGG appears wrong. Obviously there is a
reason it was added, but we need to figure out why and if it's still
true.
|
|
use setuptools.
|
|
and install command lines _before_ the target name.
Needed for non-target specific options, used e.g. by gaupol.
|
|
The list of changes is too long, it includes 5 years of development.
See NEWS and HISTORY files for details.
|
|
|
|
and racket{,-textual} each other.
|
|
Add racket and racket-textual.
|
|
Remove MzScheme, successor is textual Racket (lang/racket-textual).
|
|
Import textual Racket version as lang/racket-textual.
Racket (formerly PLT Scheme) is a specific dialect of the Lisp
language based on the Scheme branch of the Lisp family.
|
|
|
|
Overview of changes from 0.8.3 to 0.8.4
=======================================
* Fixes
- Fix alignment check on x86_64 when --enable-debug is set (bug #26950).
- Fix test-storage.c
|
|
pkgsrc changes: desupport build options that are turned into module packages.
2.49 (2010-07-07)
=================
User visible changes
--------------------
* New command line option -disable-readline lets working around bugs and
incompatibilities between readline CLISP was built against and the
library actually installed.
See <http://clisp.cons.org/impnotes/clisp.html#opt-norl> for details.
* FFI:OPEN-FOREIGN-LIBRARY now accepts the :REQUIRE argument.
See <http://clisp.cons.org/impnotes/dffi.html#dffi-open-lib> for details.
* New user variable CUSTOM:*USER-LIB-DIRECTORY* is respected by REQUIRE
and used by "clisp-link install".
Dynamic modules are now the default build option.
See <http://clisp.cons.org/impnotes/require.html> and
<http://clisp.cons.org/impnotes/clisp-link.html> for details.
* Function RENAME-FILE now accepts :IF-EXISTS argument which determines
the action when the destination exists, unless, of course, *ANSI* is T.
See <http://clisp.cons.org/impnotes/file-func.html#rename-file> for details.
* The replacement value entered by the user in STORE-VALUE and USE-VALUE
restarts is now EVALuated.
* The old user variable CUSTOM:*PRINT-CLOSURE* now controls interpreted
closure output too (RFE#3001956). This is a tricky feature, read up!
See <http://clisp.cons.org/impnotes/multi-rep.html#pr-closure>.
* Module readline now supports readline 6.1.
(Older versions are, of course, still supported).
See <http://clisp.cons.org/impnotes/readline-mod.html> for details.
* Module pcre now supports pcre 8.01.
(Older versions are, of course, still supported).
See <http://clisp.cons.org/impnotes/pcre.html> for details.
* Module libsvm does not come with the upstream sources anymore, install
locally and pass --with-libsvm-prefix to the top-level configure instead.
All upstream versions up to 2.91 are supported.
See <http://clisp.cons.org/impnotes/libsvm.html> for details.
* Module berkeley-db now supports Berkeley-DB 4.8.
(Older versions are, of course, still supported).
See <http://clisp.cons.org/impnotes/berkeley-db.html> for details.
* Module postgresql now supports PostgreSQL 8.4.
(Older versions are, of course, still supported).
See <http://clisp.cons.org/impnotes/postgresql.html> for details.
* Module pari has been updated to support both 64 & 32 bit platforms
with and without GMP.
See <http://clisp.cons.org/impnotes/pari.html> for details.
* New functions OS:VERSION-COMPARE et al call strverscmp.
See <http://clisp.cons.org/impnotes/syscalls.html#strverscmp> for details.
* Bug fixes:
+ Do not eliminate function calls which are advertised to have
exceptional situation in unsafe code (bug#2868166).
+ Fix an internal error in DECLAIM on bad OPTIMIZE quality (bug#2868147).
+ CLEAR-INPUT now clears the EOF condition on file streams (bug#2902716).
+ When quitting on a signal, never enter the debugger (bug#2795278).
+ Respect :FULL T in DIRECTORY :WILD-INFERIORS (bug#3009966).
+ Handle TWO-WAY-STREAM and ECHO-STREAM correctly by
(SETF STREAM-EXTERNAL-FORMAT) (bug#3020933).
+ Fix unbuffered output pipe stream initialization (bug#3024887).
* ANSI compliance:
+ Implement the ANSI issue COMPILER-DIAGNOSTICS:USE-HANDLER: use the
CL Condition System for compiler diagnostics.
+ STREAM-ELEMENT-TYPE on empty CONCATENATED-STREAMs now returns NIL
because nothing can be read from such streams (bug#3014921).
|
|
more and he has agreed in private mail to drop maintainership.
|
|
Changes in sbcl-1.0.40 relative to sbcl-1.0.39:
* bug fix: readdir now works on :inode64 darwin builds (lp#592897)
* bug fix: Name conflicts between symbols passed as arguments to a single
call to IMPORT no longer add multiple symbols with the same name to the
package (detectable via DO-SYMBOLS).
* bug fix: support building without the dlshim on darwin x86 and x86-64
(lp#533470).
* bug fix: TRACE :ENCAPSULATE NIL now works on ppc/linux.
|
|
|
|
not-affected version
|
|
|
|
ECL 10.4.1:
===========
* Bugs fixed:
- Fixed offset problems in READ/WRITE-SEQUENCE.
- In multithreaded ECL, the signal servicing thread was not created
due to a stale debug statement.
- The tool 'ecl-config' contained unsubstituted values, such as @LDRPATH@
because these variablesls had been renamed in the autoconf file.
- The reimplementation of DIRECTORY introduced a bug in pathnames with ending
double wildcars, such as "/foo/**/".
- EXT:SAFE-EVAL (or cl_safe_eval in C) would attempt to start the debugger
when an error was signaled.
- The logical hostname EXT: did not have a pathname translation relative to
that of SRC:
- After loading the compiler, the COMMON-LISP package was left unlocked
against changes.
- ECL can build again itself using a C++ compiler (--with-cxx thus works).
- ecl_make_file_stream_from_fd() did not set the right mode in the stream
structure.
- ECL did not understand type declarations that involved symbol macros.
- (SETF STREAM-EXTERNAL-FORMAT) inserted the wrong format type in some cases.
* Visible changes:
- The source location annotator records the original pathnames instead of
truenames. This allows Slime to find out the ECL sources even if they
have been moved.
- The signal servicing thread now carries the name SI:SIGNAL-SERVICING, which
is different from the name of the threads that are created to actually
handle those signals. Users should avoid killing SI:SIGNAL-SERVICING.
- FFI:CLINES admits the notation @lisp-object, just like FFI:C-INLINE. The
result is that @lisp-object gets replaced with a pointer to the given
lisp object, created at run time.
- Gray streams implement a new generic function, GRAY:STREAM-FILE-DESCRIPTOR
which is used by SERVE-EVENT to gather the C file descriptor or Windows
handle used to wait on input from the device.
- ECL restores the habit of installing manual pages for ecl and ecl-config.
- We have removed the obsolete versions of the documentation files. The
up to date documentation now lives in a separate source repository,
as explained in http://ecls.sourceforge.net/download.html
- The encoding :DEFAULT implies looking up in the special variable
EXT:*DEFAULT-EXTERNAL-FORMAT* the actual value of the encoding, which may be
a known one, or :pass-through.
- LOAD now accepts the keyword argument :EXTERNAL-FORMAT
* ASDF:
- ECL's contributed modules and ECL's compiler can now be used in standalone
programs, for they are now stored both in FASL form and as statically linked
libraries.
- ASDF:MAKE-BUILD takes a new keyword argument, :MOVE-HERE, that allows moving
the files that ASDF creates to the current directory. :MOVE-HERE can be the
boolean value T, in which case the files are output at ./asdf-output/ or a
pathname designator with the actual destination.
- ASDF:MAKE-BUILD admits a new type, :BINARY, which creates a standalone
system (*.asd) accompanied by two binary files, statically linked library
(.a or .lib), and a FASL (*.fasb). This can be used, to replace existing
systems with precompiled ones.
|
|
New in version 1.0.39
- bug fix: Backtrace from undefined function on x86 and x86-64
now show the calling frame.
- bug fix: linkage-table entries on PPC now no longer overflow
their space allocation (potentially causing crashes if they
are written out of order).
- bug fix: Scrub control stack after scavenging in gencgc on
non-x86oid platforms, preventing the GC from seeing stale
pointers on the control stack in subsequent GCs (which would,
and does, break invariants).
- bug fix: 32-bit unicode external formats now work on big-endian systems.
- bug fix: Literal characters with code points greater than
about 32767 now work on PPC UNICODE builds.
- bug fix: Any noise left by SSE operations (sqrt and
conversions) in the high order bits are explicitly cleared
out. In some contrived situations, this could lead to wrong
results in mixed real/complex float arithmetic.
- bug fix: Fix function/macro redefinition warnings when
building with clisp. (#576787, thanks to Josh Elsasser)
- new platform: experimental support for ppc/openbsd (thanks to
Josh Elsasser).
- bug fix: Floating-point traps now work on ppc/linux.
|
|
Also add some patches to remove use of deprecated symbols and fix other
problems when looking for or compiling against libpng-1.4.x.
|
|
Changes in 1.6.0_20 (6u20)
* OlsonData 2010b
* A Java Network Launch Protocol (JNLP) file without a codebase parameter, such
as the following, will no longer work with the Java SE 6 update 20 release.
<jnlp spec="0.2 1.0" href="draw.jnlp">
This means that developers must specify the codebase parameter in a JNLP file.
See the following example:
<jnlp spec="0.2 1.0"
codebase="http://java.sun.com/javase/technologies/desktop/javawebstart/apps/"
href="draw.jnlp">
* This release contains fixes for security vulnerabilities.
For more information, please see Oracle Security Alert for CVE-2010-0886
Changes in 1.6.0_19 (6u19)
* OlsonData 2010b
* Root Certificates
* Ensuring Application and Applet Security when Mixing Signed and Unsigned Code
* Interim Fix for the Transport Layer Security (TLS) Man-in-the-Middle Attack
* Bug Fixes
Changes in 1.6.0_18 (6u18)
* OlsonData 2009s
* VisualVM 1.2
* Java DB 10.5.3.0
* Performance Improvements
* Deployment Updates
* JSR-173 StAX 1.2 API Upgrade
* Bug Fixes
|
|
|
|
"no-profit").
|
|
|
|
Bump PKGREVISION.
|
|
* New compile option '-x'. This causes the compiler to produce an
executable program. '-fmain' is deprecated.
* Remove long option --verbose. Use '-v' for verbosity. Problem is
with getopt_long_only which does not like eg. -mv
* New conformity option -std=bs2000.
* FUNCTION is implemented. See cobc/reserved.c for a list of what is
implemented.
* Nested programs are partially supported.
* LINAGE is implemented.
* EXTERNAL on FD is implemented.
* SAME RECORD AREA is implemented.
* New config variables -
* Support for non-gcc compilers.
* Large file support, system dynamic loading and Berkeley DB inclusion
are default for the configure.
* New configure option --with-patch-level=<n>
* At run time, version checking is done. ie. When executing/loading
Cobol programs, the version (eg. 0.33) and the patch level (eg. 0)
are checked against the OC library version/patch level.
* Libtool is not required for systems that support native dynamic
loading. This includes Linux, Cygwin and MingW amongst others.
* Note to developers : See README for required software versions.
|
|
pulseaudio plugin unless option selected. Option off by default
so no change to default package build.
Fixes build if you happen to have pulseaudio package installed when
building.
|
|
|
|
|
|
|
|
|
|
|
|
==================================
New features and changes since 2.3
==================================
(1) Maude 2.4 provides an order-sorted Ax-unification algorithm for
all order-sorted theories (CSigma, E U Ax) such that:
- the signature Sigma is preregular modulo Ax;
- the axioms Ax associated to function symbols are as follows:
- there can be arbitrary function symbols and constants with
no equational attributes;
- the iter equational attribute can be declared for some unary
symbols;
- the comm or assoc comm attributes can be declared for some
binary function symbols, but then no other equational
attributes must be given for such symbols.
Explicitly excluded are theories with binary function symbols having
either: (i) the id:, left id:, or right id: attributes; or (ii) the
assoc attribute without the comm one; or (iii) a combination of (i)
and (ii).
(2) Unification is reflected in the META-LEVEL module by two descent
functions:
op metaUnify : Module UnificationProblem Nat Nat ~>
UnificationPair? [special (...)].
op metaDisjointUnify : Module UnificationProblem Nat Nat ~>
UnificationTriple? [special (...)].
(3) Statements (rule, equations and membership axioms) can now take a
print attribute. In print attribute mode, when a statement is executed
the items in its print attribute are printed, with variables taking
their value in the current substitution.
(4) Parsing of file names in the commands load, in, cd and pushd now
allows spaces using either of two syntactic conventions:
If the file name starts with " then all following characters will be
taken literally up to the terminating ", line feed or form feed.
If a file name starts with other than ", the following escape
sequences are recognized
\\ becomes \
\<space> becomes <space>
\" becomes "
(5) For operators in the C, CU, CI and CUI theories, if both arguments
are the same, the rewrite, srewrite and search commands and the model
checker will only consider one of the arguments for that step since
only one rewriting step is made per pass, the choice of argument is
irrelevant. This already happened for operators in the AC and ACU
theories. The frewrite command still always considers all arguments
even if they are identical since multiple rewriting steps can happen
in each pass.
(6) The GNU libsigsegv library is used to distinguigh between true
segmentation faults and stack overflows and so stack overflows are now
reported with an informative message.
(7) Several optimizations, bug fixes and improvements.
|
|
|
|
Changes since 1.2 release:
- Try to fetch X86 CPU identification using "cpuid" instruction
- Detect Ubuntu version
- Only use 64-bit libraries when running on 64-bit kernel
- Fixes for builds with older versions of Visual C
- Dectect Windows 7
- Fix Garbage Collection crashes after calling an external function
- Fix HTTP w/ DOC/WatTCP
- Fix MinGW compliation
|
|
Changes in GNU Prolog version 1.3.1:
* add working sigaction detection to detect fault addr (e.g. Mac OS X)
* add gplc option --no-mult-warn
* add prolog flags suspicious_warning, multifile_warning
* detect integer underflow/overflow in the parser
* fix a memory leak in catch/3
* increase limits (MAX_VAR_NAME_LENGTH=1024 and MAX_VAR_IN_TERM=10240)
* add PL_INT_LOWEST_VALUE and PL_INT_GREATEST_VALUE to gprolog.h
* prefix all global symbols, constants and types with Pl_ PL_ Pl
* fix a bug in the byte-code due to new max number of atoms
* provide a minimal gprolog.h
* detect if struct sigcontext needs asm/sigcontext.h on linux
* modify gplc: --c-compiler also sets linker and --linker added
* port to x86_64/bsd - many thanks to:
David Holland <dholland@netbsd.org>
* fix problem using ebx as global reg (bug in gcc 4.3.2)
* fix a bug in is/2 with [X] (X should only be an integer)
* fix a bug with atoms '/*' '*/' and '%' (were not quoted)
* increase maximum number of atoms to 1048576 (2^20)
* increase default stack sizes (16Mb for heap, 8Mb for others)
* fix stack alignment for x86_64/Solaris
* include patch from Paul Eggert <eggert@cs.ucla.edu> for sparc/solaris8
* port to x86_64/Solaris - many thanks to:
Scott L. Burson <Scott@coral8.com>
* fix a bug in arithmetics (mod)
* fix a bug in the FD solver (under 64 bits machines)
|
|
Changes in Yap-6.0.5:
- Very large programs: indexing very large DBs requires very large
intermediate data-structures. malloc is changed to avoid
memory fragmentation.
- Heap data-structures: there was some confusion about what should be
going through save/restore and atom-gc, and what is local to a
run. Fixed that by two files describing the two types: global
variables and code-space data structures.
- SWI emulation cleanup: some builti-ins (like predsort/3, plus/3) are
now in YAP proper, the file was cleaned up and made to rely more on
Changes in Yap-6.0.4:
Overhaul/cleanup of the attributed variable and coroutining code:
- attributed variables now don't have a separate stack; instead,
they are allocated as global variables.
- Prolog code for attributed variables cleaned up: layers now
are engine, hprolog/SWI attvars, freeze and SICStus emulation
libraries
Changes in Yap-6.0.3:
6.0.3 is another bug fix release:
- a bug in indexing large integers
- a choice-point left over in consulting
- fix error handling in comparisons
- garbage collection of functor/3 could sometimes break
- fixes to chr port
- exec_prefix
- bad error reporting in some arithmetic conditions
- erf function
- win32 install fixes
- operators need to be exported in module list, some weren't.
- tabling stuff
Changes in Yap-6.0.2:
This is a bug fix release:
- fix current_op/3 in WIN32 (P Moura)
- fix nb_current/2 (J Santos) and add SWI user:exception/3
(Bernd and Jose), add fix to nb_ in distributed ProbLog.
- fix goal_expansion
- fix clp(fd) all_distinct/1 and circuit/1 (M Triska)
- fix seekable file detection in WIN32
- add yap.pdf and yap.html to win32 bundle (M Ferreira).
- check for badly typed calls to load_files and friends.
- threads: fix bad locking while spying (P Moura).
Changes in Yap-6.0.1:
- engine improvements;
- many bug fixes;
- new implementation of arithmetic;
- better ISO compatibility;
- improvements in tabling;
- SWI-compatibility, YAP now includes support for most of the
SWI foreign interface, and many built-ins, ports of SWI
packages include jpl, chr, clpfd, clpr, sgml, prolog_xref,
swi-minisat interface, and plunit;
- yap-6 also includes a really cool new language called ProbLog.
Changes in Yap-5.1.4:
- FIXED: bad locking in p_signal (obs from Paulo Moura).
- FIXED: DESTDIR and chr.
- FIXED: make term_variables a builtin. (SWI compatibility).
- NEW: X is random(Int) (SWI compatibility).
- NEW: seletchk/3.
- FIXED: do meta-expansion from undefp.
- FIXED: handle correctly flatten([_,[_]],L).
- FIXED: bad syntax in config.h (patch from Keri Harris).
- NEW: format over atom/1.
- FIXED: clean up apply_macros in swi mode.
- FIXED: clean up meta-expansion.
- FIXED: do meta-expansion from undefined call.
- NEW: selectchk/3, nth1/3 and nth1/4.
- FIXED: YapOpcodes has wrong formats (obs from Bart Demoen).
- FIXED: improve format message (obs from Bart Demoen).
- NEW: more versions of maplist (obs from Bart Demoen).
- FIXED: use import mechanism for SWI's flatten/2 (obs from Bart Demoen).
- FIXED: steps towards typed rbtrees plus some cases where one should not copy the null node (Tom Schrijvers and Bart
Demoen).
- FIXED: map_tree in trees library (Bart Demoen).
- FIXED: bad call to splay_tree (Bart Demoen).
- FIXED: bad type for write_x_var in add_info (Bart Demoen).
- FIXED: exec/3 should flush streams.
- FIXED: load_files/2 compilation_mode to allow :- source and replace compile by compact.
- FIXED: list_concat/2 (fix from Bart Demoen).
- NEW: dgraph_reachable/3 and friends.
|
|
|
|
Notable changes in SWI-Prolog 5.11.0:
* ODBC now properly handles locale. This means you can
use it with Unicode if you use a UTF-8 based locale.
* Recently reported manifest problems around xpce should be
fixed with help from Degski.
* Message queues can now be destroyed while they have suspended
writers, making pipe-like cooperation between threads much easier.
* Using set_prolog_flag(qcompile, auto), you can ask the system
to automatically maintain quick-load versions of your source-files.
This can greatly improve development comfort for really large
applications.
* Jeff Rosenwald contributed an interface for Google protocol buffers
Notable changes in SWI-Prolog 5.9.10:
* A lot of stuff has been added to enhance the compatibility to SICStus.
This involves some changes to the kernel, some additions to the libraries
and starting and populating a SICStus emulation layer. All this was
initiated by the with to make the Alpino NLP parser suite
(http://www.let.rug.nl/vannoord/alp/Alpino/) available for SWI-Prolog.
Thanks to Gertjan van Noord, this is now almost reality (there
are some remaining tweaks, notably with the Tcl/Tk interface).
* Also various enhancements and fixes to the tabbed-editing support for
the built-in editor.
|
|
|
|
Submitted by Peter Bex on IRC.
Changes in 4.5.0
- internal fixes of handling of alternative installation-prefix
in setup-api
- certain compiler-warnings that are in really just notes
and don't indicate a possible error (like reimport of
identifiers) are only shown with -S or in verbose mode
- fixed handling of VARDIR in `chicken-install' (thanks to
Davide Puricelli)
- `chicken-install -test' doesn't runs tests for dependencies
- when a non-else clause follows an else-clause in `cond',
`case' or `select' a warning (note) is shown in verbose
mode
- removed the deprecated `define-extension' and
`define-compiled-syntax'
- `chicken-uninstall' now always asks before removing
extensions, unless `-force' is given
- improved performance of keyword-argument processing slightly
- `export' outside of a module definition has no effect
- `number->string' now accepts arbitrary bases between 2 and 16
(thanks to Christian Kellermann)
- fixed `standard-extension' in `setup-api' module
- literal constants keep their identity, even when inlined
- Unit library: added `fxodd?' and `fxeven?'
- All hardcoded special forms have been replaced with
syntax definitions that expand into internal forms, this
allows redefinition and shadowing of all Scheme core forms
- faster implementations of `get' and `put!'
- faster implementation of `assq' in unsafe mode
- the `-sx' option prefixes each output line with `;'
- slightly better expansion performance
- more documentation of the C API (thanks to Peter Bex)
- `module' supports a shorthand form that refers directly
to a file to be included as the module body
- added runtime option `-:G' to force GUI mode (on platforms
that distinguish between GUI and non-GUI applications)
- removed the unsafe runtime library (`libuchicken'), this
simplifies and speeds up the build and reduces the risk
of executables loading library units from different
variants of the runtime system
- removed the `-unsafe-libraries' option from `chicken'
and `csc'
- removed bootstrapping target and bootstrapping files from
development repository; to bootstrap the system, either
use a release or development-snapshot tarball or fetch
a statically linked precompiled `chicken' binary from
http://chicken.wiki.br/chicken-projects/bootstrap/
- Jim Ursetto provided some fixes for building universal
binaries on Mac OS X
- `csc' now compiles and links Windows resource (.rc) files
when given on the command line
- `chicken-install' and `chicken-uninstall' have an embedded
manifest that suppresses the elevation dialog on Windows
Vista and later when UAC is activated (Thanks to Thomas Chust)
- the `install' program is not used in the build on mingw
and mingw/MSYS platforms, since this is broken on older
mingw versions
- line-number-information is now properly handled (in the
few places where it is used) correctly for included files;
the source file is given in trace-output in addition to
the line number
- removed compiler warning for shared objects compiled in
unsafe mode
- unboxing is now only done in unsafe mode
- in unsafe mode, pointer-accessors from the `lolevel' unit
are now handled intrinsically by the compiler
- `chicken-install' accepts now relative pathnames for the
`-prefix' option
- `define-record-type' now optionally allows using SRFI-17
setters as record-field modifiers
- `integer?' returns `#f' for NaN and infinite numbers
- `csc' now has an `-no-elevation' option for embedding a
manifest that prevents the elevation dialog on Windows
Vista and later when IAC is activated
- the `,d' csi command displays qualified symbols properly
- symbols starting with the `#\#' character trigger an
error when encountered by the reader
- Unit posix: `glob->regexp' now always returns a regular
expression object or optionally an SRE
- Unit posix: `terminal-port?' and `terminal-size' have been
implemented for Windows, the latter always returns `0 0',
though (thanks to Jim Ursetto)
- Unit regex: `regexp' now accepts a regular expression
object as argument
- Unit regex: removed `glob?'
- fixed bug in `chicken-install'/`chicken-uninstall' and
`chicken-status' that prevented collapsed command-line
options to be handled correctly.
- disabled runpath-fix for deployed applications for netbsd
(but resurrected providing a runpath at all, thanks to
Peter Bex)
- Peter Bex provided documentation for the `C_closure' C API
function
|
|
|
|
|
|
The [1]GNU project and the GCC developers are pleased to announce the
release of GCC 4.4.4.
This release is a bug-fix release, containing fixes for regressions in
GCC 4.4.3 relative to previous releases of GCC.
|
|
|
|
http://lists.gnu.org/archive/html/libtool/2009-11/msg00065.html
|
|
PR pkg/43237 by NAKAJIMA Yoshihiro.
This isn't too useful yet due to an incompatibility (apparently
a bug in gcc-4.1) which will be fixed soon.
bump PKGREVISION
|
|
|
|
What's New in Python 2.6.5?
===========================
*Release date: 2010-03-18*
What's New in Python 2.6.5 rc 2?
================================
*Release date: 2010-03-09*
Core and Builtins
-----------------
- Issue #8089: a OS X framework build with --with-universal-archs=3-way|intel
had no way to select a 32-bit executable.
- Issue #8084: fixes build issues on OSX 10.6 when targetting OSX 10.4.
Library
-------
- Reverting the changes made in r78432. Discussed in the tracker issue #7540.
- Issue #8107: Fixed test_distutils so it doesn't crash when the source
directory cannot be found.
Extension Modules
-----------------
- Issue #7670: sqlite3: Fixed crashes when operating on closed connections.
- Issue #8053: logic was inverted on which platforms to run a test on.
caused test_thread to fail on Windows.
What's New in Python 2.6.5 rc 1?
================================
*Release date: 2010-03-01*
Core and Builtins
-----------------
- Issue #7309: Fix unchecked attribute access when converting
UnicodeEncodeError, UnicodeDecodeError, and UnicodeTranslateError to
strings.
- Issue #7649: "u'%c' % char" now behaves like "u'%s' % char" and raises a
UnicodeDecodeError if 'char' is a byte string that can't be decoded using
the default encoding.
- Issue #5677: Explicitly forbid write operations on read-only file objects,
and read operations on write-only file objects. On Windows, the system C
library would return a bogus result; on Solaris, it was possible to crash
the interpreter. Patch by Stefan Krah.
- Issue #4978: Passing keyword arguments as unicode strings is now allowed.
- Issue #7819: Check sys.call_tracing() arguments types.
- Issue #7788: Fix an interpreter crash produced by deleting a list
slice with very large step value.
- Issue #7561: Operations on empty bytearrays (such as `int(bytearray())`)
could crash in many places because of the PyByteArray_AS_STRING() macro
returning NULL. The macro now returns a statically allocated empty
string instead.
- Issue #7604: Deleting an unset slotted attribute did not raise an
AttributeError.
- Issue #7413: Passing '\0' as the separator to datetime.datetime.isoformat()
used to drop the time part of the result.
- Issue #6108: unicode(exception) and str(exception) should return the same
message when only __str__ (and not __unicode__) is overridden in the
subclass.
- Issue #7491: Metaclass's __cmp__ method was ignored.
- Add Py3k warnings for parameter names in parenthesis.
- Issue #7362: Give a proper error message for def f((x)=3): pass.
- Issue #7085: Fix crash when importing some extensions in a thread
on MacOSX 10.6.
- Issue #7070: Fix round bug for large odd integer arguments.
- Issue #7078: Set struct.__doc__ from _struct.__doc__.
- Issue #1722344: threading._shutdown() is now called in Py_Finalize(), which
fixes the problem of some exceptions being thrown at shutdown when the
interpreter is killed. Patch by Adam Olsen.
- Issue #7084: Fix a (very unlikely) crash when printing a list from one
thread, and mutating it from another one. Patch by Scott Dial.
- Issue #1747858: Fix lchown & fchown to work with large uid's and gid's on
64-bit platforms.
Library
-------
- Issue #7250: Fix info leak of os.environ across multi-run uses of
wsgiref.handlers.CGIHandler.
- Issue #1729305: Fix doctest to handle encode error with "backslashreplace".
- Issue #691291: codecs.open() should not convert end of lines on reading and
writing.
- Issue #7975: correct regression in dict methods supported by bsddb.dbshelve.
- Issue #7959: ctypes callback functions are now registered correctly
with the cycle garbage collector.
- Issue #6243: curses.getkey() can segfault when used with ungetch.
Fix by Trundle and Jerry Chen.
- Issue #7597: curses.use_env() can now be called before initscr().
Noted by Kan-Ru Chen.
- Issue #7970: email.Generator.flatten now correctly flattens message/rfc822
messages parsed by email.Parser.HeaderParser.
- Issue #3426: ``os.path.abspath`` now returns unicode when its arg is unicode.
- Issue #7835: shelve should no longer produce mysterious warnings during
interpreter shutdown.
- Issue #4772: Raise a ValueError when an unknown Bluetooth protocol is
specified, rather than fall through to AF_PACKET (in the `socket` module).
Also, raise ValueError rather than TypeError when an unknown TIPC address
type is specified. Patch by Brian Curtin.
- Issue #6939: Fix file I/O objects in the `io` module to keep the original
file position when calling `truncate()`. It would previously change the
file position to the given argument, which goes against the tradition of
ftruncate() and other truncation APIs. Patch by Pascal Chambon.
- Issue #7773: Fix an UnboundLocalError in platform.linux_distribution() when
the release file is empty.
- Issue #7748: Since unicode values are supported for some metadata options
in Distutils, the DistributionMetadata get_* methods will now return an utf-8
encoded string for them. This ensure that the upload and register commands
send the right values to PyPI without any error.
- Issue #1670765: Prevent email.generator.Generator from re-wrapping
headers in multipart/signed MIME parts, which fixes one of the sources of
invalid modifications to such parts by Generator.
- Issue #7701: Fix crash in binascii.b2a_uu() in debug mode when given a
1-byte argument. Patch by Victor Stinner.
- Issue #3299: Fix possible crash in te _sre module when given bad
argument values in debug mode. Patch by Victor Stinner.
- Issue #5827: Make sure that normpath preserves unicode. Initial patch
by Matt Giuca.
- Issue #5372: Drop the reuse of .o files in Distutils' ccompiler (since
ing the .c
file). Initial patch by Collin Winter.
- Issue #7617: Make sure distutils.unixccompiler.UnixCCompiler recognizes
gcc when it has a fully qualified configuration prefix. Initial patch
by Arfrever.
- Issue #7071: byte-compilation in Distue.
- Issue #7092: Remove py3k warning when importing cPickle. 2to3 handles
renaming of `cPickle` to `pickle`. The warning was annoying since there's
no alternative to cPickle if you care about performance. Patch by Florent
Xicluna.
- Issue #745tch by
Victor Stinner.
- Issue #6511: ZipFile now raises BadZipfile (instead of an IOError) when
opening an empty or very small file.
- Issue #7552: Removed line feed in the base64 Authorization header in
the Distutils upload command to avoid an ers on long passwords. Initial patch by JP St. Pierre.
- Issue #7231: urllib2 cannot handle https with proxy requiring auth. Patch by
Tatsuhiro Tsujikawa.
- Issue #7348: StringIO.StringIO.readline(-1) now acts as if it got no argument
like other file objects.
- Issue #5949: fixed IMAP4_SSL hang when the IMAP server response is
missing proper end-of-line termination.
- Fix variations of extending deques: d.extend(d) d.extendleft(d) d+=d
- Issue #1923: Fixed the removal of meaningful spaces when PKG-INFO is
generated in Distutils. Patch by Stephen Emslie.
- Issue #4120: Drop reference to CRT from manifest when building extensions
with msvc9compiler.
- Issue #7410: deepcopy of itertools.count() erroneously reset the count.
- Issue #7403: logging: Fixed possible race condition in lock creation.
- Issue #7341: Close the internal file object in the TarFile constructor in
case of an error.
- Issue #7328: pydoc no longer corrupts sys.path when run with the '-m' switch
- Issue #7318: multiprocessing now uses a timeout when it fails to establish
a connection with another process, rather than looping endlessly. The
default timeout is 20 seconds, which should be amply sufficient for
local connections.
- Issue #7282: Fix a memory leak when an RLock was used in a thread other
than those started through `threading.Thread` (for example, using
`thread.start_new_thread()`.
- Issue #7264: Fix a possible deadlock when deallocating thread-local objects
which are part of a reference cycle.
- Issue #7249: Methods of io.BytesIO now allow `long` as well as `int`
arguments.
- Issue #6665: Fix fnmatch to properly match filenames with newlines in them.
- Issue #1008086: Fixed socket.inet_aton() to always return 4 bytes even on
LP64 platforms (most 64-bit Linux, bsd, unix systems).
- Issue #7246 & Issue #7208: getpass now properly flushes input before
reading from stdin so that existing input does not confuse it and
lead to incorrect entry or an IOError. It also properly flushes it
afterwards to avoid the terminal echoing the input afterwards on
OSes such as Solaris.
- Issue #7244: itertools.izip_longest() no longer ignores exceptions
raised during the formation of an output tuple.
- Issue #7233: Fix a number of two-argument Decimal methods to make
sure that they accept an int or long as the second argument. Also
fix buggy handling of large arguments (those with coefficient longer
than the current precision) in shift and rotate.
- Issue #7082: When falling back to the MIME 'name' parameter, the
correct place to look for it is the Content-Type header.
- Issue #7099: Decimal.is_normal now returns True for numbers with exponent
larger than emax.
- Issue #7205: Fix a possible deadlock when using a BZ2File object from
several threads at once.
- Issue #7048: Force Decimal.logb to round its result when that result
is too large to fit in the current precision.
- Issue #1488943: difflib.Differ() doesn't always add hints for tab characters
- Issue #5037: Proxy the __unicode__ special method to __unicode__ instead of
__str__.
- Issue #7481: When a threading.Thread failed to start it would leave the
instance stuck in initial state and present in threading.enumerate().
- Issue #1068268: The subprocess module now handles EINTR in internal
os.waitpid and os.read system calls where appropriate.
Extension Modules
-----------------
- Issue #7808: Fix reference leaks in _bsddb and related tests.
- Stop providing crtassem.h symbols when compiling with Visual Studio 2010, as
msvcr100.dll is not a platform assembly anymore.
- Issue #6877: Make it possible to link the readline extension to libedit
on OSX.
- Expat: Fix DoS via XML document with malformed UTF-8 sequences
(CVE_2009_3560).
- Issue #7242: On Solaris 9 and earlier calling os.fork() from within a
thread could raise an incorrect RuntimeError about not holding the import
lock. The import lock is now reinitialized after fork.
- Issue #7999: os.setreuid() and os.setregid() would refuse to accept a -1
parameter on some platforms such as OS X.
Build
-----
- Issue #3920, #7903: Define _BSD_SOURCE on OpenBSD 4.4 through 4.9.
- Issue #7661: Allow ctypes to be built from a non-ASCII directory path.
Patch by Florent Xicluna.
- Issue #7589: Only build the nis module when the correct header files are
found.
- Switch to OpenSSL 0.9.8l on Windows.
- Issue #6603: Change READ_TIMESTAMP macro in ceval.c so that it
compiles correctly under gcc on x86-64. This fixes a reported
problem with the --with-tsc build on x86-64.
- Ensure that it possible to build extensions for the default
binary distribution on OSX 10.6 even when the user does not
have the 10.4u SDK installed.
- Issue #7541: when using ``python-config`` with a framework install the
compiler might use the wrong library.
Documentation
-------------
- Updating `Using Python` documentation to include description of CPython's
-J, -U and -X options.
- Update python manual page (options -B, -O0, -s, environment variables
PYTHONDONTWRITEBYTECODE, PYTHONNOUSERSITE).
Tests
-----
- issue #7728: test_timeout was changed to use test_support.bind_port
instead of a hard coded port.
- Issue #7498: test_multiprocessing now uses test_support.find_unused_port
instead of a hardcoded port number in test_rapid_restart.
- Issue #7431: use TESTFN in test_linecache instead of trying to create a
file in the Lib/test directory, which might be read-only for the
user running the tests.
- Issue #7324: add a sanity check to regrtest argument parsing to
catch the case of an option with no handler.
- Issue #7295: Do not use a hardcoded file name in test_tarfile.
- Issue #7270: Add some dedicated unit tests for multi-thread synchronization
primitives such as Lock, RLock, Condition, Event and Semaphore.
- Issue #7055: test___all__ now greedily detects all modules which have an
__all__ attribute, rather than using a hardcoded and incomplete list.
|
|
0.9.9
-----
New features:
- C++ struct declarations, 'new' operator, and del for C++ delete.
- As well as the 'not None' modifier for extension type arguments, there
is now 'or None' to explicitly allow passing None. It is planned to make
'not None' the default in a future version. For now, a warning is issued
(once per Pyrex run) if you do not specify one or the other.
- Extension types may have a 'nogc' option to suppress GC support.
Modifications:
- Exceptions caught by an except clause are no longer put into the thread
state and cannot be retrieved using sys.exc_info(). To access the caught
exception, it must be bound to a name in the except clause. A third name
can be supplied to capture the traceback.
- PyString_InternFromString is now exposed under the name 'cintern' because
it is not a complete substitute for 'intern' (it can't handle strings
containing null bytes). [John Arbash Meinel]
- Disabled size check for cimported types because it was generating too
many false positives in the field.
- Added __fastcall calling convention. Also no longer assuming that an
unspecified calling convention is the same as __cdecl.
- Operations between signed and unsigned ints of the same size now have
an unsigned result.
- Py_ssize_t now ranked between long and long long.
- Declaration of __new__ in an extension type is now an error, in
preparation for introducing new semantics.
- Added size_t type, and made sizeof() return it.
Bug fixes:
- When casting the result of a function call to a Python type, the
function could be called twice. [David Martinez]
- __Pyx_GetException was not taking account of the fact that the traceback
could be NULL.
- sizeof(module.typename) did not work. [Daniele Pianu]
|