Node v0.5.5

177 views
Skip to first unread message

Ben Noordhuis

unread,
Aug 26, 2011, 7:23:15 PM8/26/11
to nod...@googlegroups.com
2011.08.26, Version 0.5.5 (unstable)

* typed arrays, implementation from Plesk

* fix IP multicast on SunOS

* fix DNS lookup order: IPv4 first, IPv6 second (--use-uv only)

* remove support for UNIX datagram sockets (--use-uv only)

* UDP support for Windows (Bert Belder)

* #1572 improve tab completion for objects in the REPL (Nathan Rajlich)

* #1563 fix buffer overflow in child_process module (reported by Dean McNamee)

* #1546 fix performance regression in http module (reported by Brian Geffon)

* #1491 add PBKDF2 crypto support (Glen Low)

* #1447 remove deprecated http.cat() function (Mikeal Rogers)

* #1140 fix incorrect dispatch of vm.runInContext's filename argument
(Antranig Basman)

* #1140 document vm.runInContext() and vm.createContext() (Antranig Basman)

* #1428 fix os.freemem() on 64 bits freebsd (Artem Zaytsev)

* #1164 make all DNS lookups async, fixes uncatchable exceptions
(Koichi Kobayashi)

* fix incorrect ssl shutdown check (Tom Hughes)

* various cmake fixes (Tom Hughes)

* improved documentation (Koichi Kobayashi, Logan Smyth, Fedor Indutny,
Mikeal Rogers, Maciej Małecki, Antranig Basman, Mickaël Delahaye)

* upgrade libuv to commit 835782a

* upgrade V8 to 3.5.8

Download: http://nodejs.org/dist/node-v0.5.5.tar.gz

Windows Executable: http://nodejs.org/dist/v0.5.5/node.exe

Website: http://nodejs.org/docs/v0.5.5/

Documentation: http://nodejs.org/docs/v0.5.5/api/

Benjamin Gudehus

unread,
Aug 26, 2011, 7:28:16 PM8/26/11
to nod...@googlegroups.com
Typed arrays... wow! :)

Benjamin

2011/8/27 Ben Noordhuis <in...@bnoordhuis.nl>
--
Job Board: http://jobs.nodejs.org/
Posting guidelines: https://github.com/joyent/node/wiki/Mailing-List-Posting-Guidelines
You received this message because you are subscribed to the Google
Groups "nodejs" group.
To post to this group, send email to nod...@googlegroups.com
To unsubscribe from this group, send email to
nodejs+un...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Benjamin Gudehus

unread,
Aug 26, 2011, 7:44:44 PM8/26/11
to nod...@googlegroups.com
Hi!

The node.exe for 0.5.5 is about half the size smaller than previous releases.
Was this executeble compressed with a runtime packer?

Greetings
Benjamin

2011/8/27 Benjamin Gudehus <hast...@googlemail.com>

Floby

unread,
Aug 27, 2011, 8:32:15 AM8/27/11
to nodejs
maybe stripped ?

On Aug 27, 1:44 am, Benjamin Gudehus <hasteb...@googlemail.com> wrote:
> Hi!
>
> The node.exe for 0.5.5 is about half the size smaller than previous
> releases.
> Was this executeble compressed with a runtime packer?
>
> Greetings
> Benjamin
>
> 2011/8/27 Benjamin Gudehus <hasteb...@googlemail.com>
>
>
>
>
>
>
>
> > Typed arrays... wow! :)
>
> > Benjamin
>
> > 2011/8/27 Ben Noordhuis <i...@bnoordhuis.nl>

Paul Spaulding

unread,
Aug 27, 2011, 2:51:01 PM8/27/11
to nod...@googlegroups.com
node.exe ->

Entry Point Not Found
The procedure entry point CreateSymbolicLinkA could not be located in the dynamic link library KERNEL32.dll.

This is new; I never had this error on the previous versions of node.exe.

Ted Young

unread,
Aug 27, 2011, 2:55:46 PM8/27/11
to nod...@googlegroups.com
I'm getting a build error on OSX 10.5

DEST_OS: darwin
DEST_CPU: ia32
Parallel Jobs: 1
Product type: program
[ 5/38] libv8.a: deps/v8/SConstruct -> out/Release/libv8.a
/System/Library/Frameworks/Python.framework/Versions/2.5/Resources/
Python.app/Contents/MacOS/Python "/Users/devel/Sites/node-v0.5.5/tools/
scons/scons.py" -j 1 -C "/Users/devel/Sites/node-v0.5.5/out/Release/" -
Y "/Users/devel/Sites/node-v0.5.5/deps/v8" visibility=default
mode=release arch=ia32 toolchain=gcc library=static snapshot=on
scons: Reading SConscript files ...

scons: warning: Ignoring missing SConscript 'obj/test/release/
SConscript'
File "/Users/devel/Sites/node-v0.5.5/deps/v8/SConstruct", line 1476,
in BuildSpecific
scons: done reading SConscript files.
scons: Building targets ...
/usr/bin/g++ -o obj/release/elements.o -c -fno-rtti -fno-exceptions -
Wall -Werror -W -Wno-unused-parameter -Wnon-virtual-dtor -pedantic -
m32 -O3 -fomit-frame-pointer -fdata-sections -ffunction-sections -ansi
-mmacosx-version-min=10.4 -DV8_TARGET_ARCH_IA32 -
DENABLE_DEBUGGER_SUPPORT -I/Users/devel/Sites/node-v0.5.5/deps/v8/src /
Users/devel/Sites/node-v0.5.5/deps/v8/src/elements.cc
cc1plus: warnings being treated as errors
/Users/devel/Sites/node-v0.5.5/deps/v8/src/elements.cc: In member
function 'uint32_t
v8::internal::ElementsAccessorBase<ElementsAccessorSubclass,
BackingStoreClass>::GetCapacity(v8::internal::FixedArrayBase*) [with
ElementsAccessorSubclass =
v8::internal::NonStrictArgumentsElementsAccessor, BackingStoreClass =
v8::internal::FixedArray]':
/Users/devel/Sites/node-v0.5.5/deps/v8/src/elements.cc:531: warning:
control may reach end of non-void function 'static
v8::internal::ElementsAccessor*
v8
::internal::ElementsAccessor::ForArray(v8::internal::FixedArrayBase*)'
being inlined
scons: *** [obj/release/elements.o] Error 1
scons: building terminated because of errors.
Waf: Leaving directory `/Users/devel/Sites/node-v0.5.5/out'
Build failed: -> task failed (err #2):
{task: libv8.a SConstruct -> libv8.a}
make: *** [program] Error 1


Ted

mscdex

unread,
Aug 27, 2011, 3:11:40 PM8/27/11
to nodejs
This was just fixed. See: https://github.com/joyent/node/issues/1592

Ben Noordhuis

unread,
Aug 27, 2011, 8:22:44 PM8/27/11
to nod...@googlegroups.com
On Sat, Aug 27, 2011 at 20:55, Ted Young <t...@radicaldesigns.org> wrote:
> I'm getting a build error on OSX 10.5
>
> DEST_OS: darwin
> DEST_CPU: ia32
> Parallel Jobs: 1
> Product type: program
> [ 5/38] libv8.a: deps/v8/SConstruct -> out/Release/libv8.a
> /System/Library/Frameworks/Python.framework/Versions/2.5/Resources/Python.app/Contents/MacOS/Python
> "/Users/devel/Sites/node-v0.5.5/tools/scons/scons.py" -j 1 -C
> "/Users/devel/Sites/node-v0.5.5/out/Release/" -Y
> "/Users/devel/Sites/node-v0.5.5/deps/v8" visibility=default mode=release
> arch=ia32 toolchain=gcc library=static snapshot=on
> scons: Reading SConscript files ...
>
> scons: warning: Ignoring missing SConscript 'obj/test/release/SConscript'
> File "/Users/devel/Sites/node-v0.5.5/deps/v8/SConstruct", line 1476, in
> BuildSpecific
> scons: done reading SConscript files.
> scons: Building targets ...
> /usr/bin/g++ -o obj/release/elements.o -c -fno-rtti -fno-exceptions -Wall
> -Werror -W -Wno-unused-parameter -Wnon-virtual-dtor -pedantic -m32 -O3

> -fomit-frame-pointer -fdata-sections -ffunction-sections -ansi
> -mmacosx-version-min=10.4 -DV8_TARGET_ARCH_IA32 -DENABLE_DEBUGGER_SUPPORT
> -I/Users/devel/Sites/node-v0.5.5/deps/v8/src
> /Users/devel/Sites/node-v0.5.5/deps/v8/src/elements.cc
> cc1plus: warnings being treated as errors
> /Users/devel/Sites/node-v0.5.5/deps/v8/src/elements.cc: In member function
> 'uint32_t v8::internal::ElementsAccessorBase<ElementsAccessorSubclass,
> BackingStoreClass>::GetCapacity(v8::internal::FixedArrayBase*) [with
> ElementsAccessorSubclass = v8::internal::NonStrictArgumentsElementsAccessor,
> BackingStoreClass = v8::internal::FixedArray]':
> /Users/devel/Sites/node-v0.5.5/deps/v8/src/elements.cc:531: warning: control
> may reach end of non-void function 'static v8::internal::ElementsAccessor*
> v8::internal::ElementsAccessor::ForArray(v8::internal::FixedArrayBase*)'
> being inlined
> scons: *** [obj/release/elements.o] Error 1
> scons: building terminated because of errors.
> Waf: Leaving directory `/Users/devel/Sites/node-v0.5.5/out'
> Build failed:  -> task failed (err #2):
>        {task: libv8.a SConstruct -> libv8.a}
> make: *** [program] Error 1

Upgrade your gcc or strip the -Werror directives from
deps/v8/SConstruct. It's a gcc bug that's been fixed in 4.1.

Ryan Schmidt

unread,
Aug 28, 2011, 4:15:03 PM8/28/11
to nod...@googlegroups.com
On Aug 27, 2011, at 19:22, Ben Noordhuis wrote:

> On Sat, Aug 27, 2011 at 20:55, Ted Young wrote:
>> I'm getting a build error on OSX 10.5

>> /usr/bin/g++ -o obj/release/elements.o -c -fno-rtti -fno-exceptions -Wall


>> -Werror -W -Wno-unused-parameter -Wnon-virtual-dtor -pedantic -m32 -O3
>> -fomit-frame-pointer -fdata-sections -ffunction-sections -ansi
>> -mmacosx-version-min=10.4 -DV8_TARGET_ARCH_IA32 -DENABLE_DEBUGGER_SUPPORT
>> -I/Users/devel/Sites/node-v0.5.5/deps/v8/src
>> /Users/devel/Sites/node-v0.5.5/deps/v8/src/elements.cc
>> cc1plus: warnings being treated as errors
>

> Upgrade your gcc or strip the -Werror directives from
> deps/v8/SConstruct. It's a gcc bug that's been fixed in 4.1.

On Mac OS X 10.5, /usr/bin/g++ is version 4.0, though 4.2 is available as /usr/bin/g++-4.2. On Mac OS X 10.4, /usr/bin/g++ is version 4.0 and no newer version is available from Apple.

Is the above error a problem that will be fixed in the node source code, or is gcc >= 4.1 a requirement? If the latter, we'll have to make some changes for the nodejs port in MacPorts to force the use of a newer compiler.


Ben Noordhuis

unread,
Aug 28, 2011, 5:11:53 PM8/28/11
to nod...@googlegroups.com

We normally don't patch the bundled V8 - maintenance hassle and all
that - but I talked to one of the V8 guys and he was willing to land a
fix upstream. Can you or someone else double-check the attached patch?

0001-v8-suppress-gcc-4.0-diagnostic-error.patch

Thomas Shinnick

unread,
Aug 29, 2011, 3:38:56 AM8/29/11
to nod...@googlegroups.com
On Saturday, August 27, 2011 7:22:44 PM UTC-5, Ben Noordhuis wrote:
On Sat, Aug 27, 2011 at 20:55, Ted Young <t...@radicaldesigns.org> wrote:
[snip]

Upgrade your gcc or strip the -Werror directives from
deps/v8/SConstruct. It's a gcc bug that's been fixed in 4.1.

Separate from the OSX build error, I'm seeing the same build error on Fedora 15 with GCC 4.6 that I saw on v0.5.4 :

/usr/lib/ccache/g++ -o obj/release/ia32/full-codegen-ia32.o -c -fno-rtti -fno-exceptions -Wall -Werror -W -Wno-unused-parameter -Wnon-virtual-dtor -pedantic -m32 -O3 -fomit-frame-pointer -fdata-sections -ffunction-sections -ansi -DV8_TARGET_ARCH_IA32 -DENABLE_DEBUGGER_SUPPORT -I/home/Tom/study/javascript/node.js/node-v0.5.5/deps/v8/src /home/Tom/study/javascript/node.js/node-v0.5.5/deps/v8/src/ia32/full-codegen-ia32.cc
/home/Tom/study/javascript/node.js/node-v0.5.5/deps/v8/src/ia32/full-codegen-ia32.cc: In member function 'virtual void v8::internal::FullCodeGenerator::VisitCompareOperation(v8::internal::CompareOperation*)':
/home/Tom/study/javascript/node.js/node-v0.5.5/deps/v8/src/ia32/full-codegen-ia32.cc:4207:12: error: variable 'strict' set but not used [-Werror=unused-but-set-variable]
cc1plus: all warnings being treated as errors

scons: *** [obj/release/ia32/full-codegen-ia32.o] Error 1

scons: building terminated because of errors.

Pursuant to your suggestion last time I tried the "strip the -Werror directives from deps/v8/SConstruct" thing:

[Tom@TLSF15A v8]$ diff SConstruct.v0.5.5.original SConstruct
291d290
<                        '-Werror',

And, after first doing a "make distclean" to ensure a valid test, the make now completed successfully.

Is this the same situation you mention elsewhere with regard to not wanting to patch v8, even though it means node v0.5.x doesn't build "out of the box" ?

Evgeny Bogdanov

unread,
Aug 29, 2011, 6:41:38 AM8/29/11
to nod...@googlegroups.com
Is this a stable release?

Should I switch from 0.4.10 to 0.5.5?

Phil Plante

unread,
Aug 29, 2011, 7:31:16 AM8/29/11
to nod...@googlegroups.com
0.5.x is unstable, 0.6.x will be the stable release.  Generally builds with even numbers 0.4.x or 0.6.0 are considered "stable", odd build numbers 0.5.x are considered unstable as they are where the current development is happening.

EllisGL

unread,
Aug 31, 2011, 6:23:15 PM8/31/11
to nodejs
Downloaded that fix, now http.listen doesn't exist.

On Aug 27, 2:11 pm, mscdex <msc...@gmail.com> wrote:
> On Aug 27, 2:51 pm, Paul Spaulding <pjs67...@gmail.com> wrote:
>
> > node.exe ->
>
> > Entry Point Not Found
> > The procedure entry pointCreateSymbolicLinkAcould not be located in the

EllisGL

unread,
Aug 31, 2011, 6:25:31 PM8/31/11
to nodejs
Using the code from the "hello world" front page example.

Server running at http://127.0.0.1:1337/

node.js:205
throw e; // process.nextTick error, or 'error' event on first
tick
^
Error: listen UNKNOWN
at errnoException (net_uv.js:564:11)
at Array.<anonymous> (net_uv.js:630:26)
at EventEmitter._tickCallback (node.js:197:26)

Ryan Schmidt

unread,
Sep 1, 2011, 11:44:56 PM9/1/11
to nod...@googlegroups.com

On Aug 28, 2011, at 16:11, Ben Noordhuis wrote:

I have now confirmed the problem on my Leopard machine, and the patch does not fix it or change the error message.


Ben Noordhuis

unread,
Sep 2, 2011, 9:43:16 AM9/2/11
to nod...@googlegroups.com
On Fri, Sep 2, 2011 at 05:44, Ryan Schmidt <googl...@ryandesign.com> wrote:
>> We normally don't patch the bundled V8 - maintenance hassle and all
>> that - but I talked to one of the V8 guys and he was willing to land a
>> fix upstream. Can you or someone else double-check the attached patch?
>
> I have now confirmed the problem on my Leopard machine, and the patch does not fix it or change the error message.

Okay, thanks. We're going to strip the -Werror directives from now on.

Reply all
Reply to author
Forward
0 new messages