Node v0.2.5

[Ryan Dahl]

2010.11.16, Version 0.2.5

* Add ref to buffer during fs.write and fs.read. Sometimes buffers
would be GCed before making it to the thread pool.

* Fix http buffer pushing bug:
http://groups.google.com/group/nodejs/browse_thread/thread/f66cd3c960406919

* Gracefully handle EMFILE and server.maxConnections

* "node --eval" evals in the global scope. (Jorge Chamorro Bieling)

* Handle cyclic links smarter in fs.realpath (isaacs, Issue #167)

* Rewrite libeio After callback to use req->result instead of req->errorno
for error checking (Micheil Smith)

* Removed range read optimization as it doesn't work with libeio. (Chandra
Sekar S)

* Make writes to process.env update the real environment (Ben Noordhuis)

* Set FD_CLOEXEC flag on stdio FDs before spawning. (Guillaume Tuton)

* Upgrade http-parser

* Readline: Switch \n with \r\n for all strings printed out.
Simplify displayPrompt

* Fix segfault on test-crypto

* Add --profile flag to configure script, enables gprof profiling. (Ben
Noordhuis)

* writeFileSync could exhibit pathological behavior when a buffer could
not be written to the file in a single write() call. (isaacs)

* Fix OS::GetExecutablePath for platform_none shouldn't return garbage.

Download: http://nodejs.org/dist/node-v0.2.5.tar.gz

Website: http://nodejs.org/docs/v0.2.5/

Documentation: http://nodejs.org/docs/v0.2.5/api.html

[John Taber]

is vers 3.1 less stable than 2.5 ? or is there just some abi breakage
in the 3.x series ?

[Oleg "Sannis" Efimov]

> * writeFileSync could exhibit pathological behavior when a buffer could
> not be written to the file in a single write() call. (isaacs)

Hm, as I understand, this is a bug? Please, mybe someone can explain
me this in more simple way :-)

On 17 ноя, 08:56, Ryan Dahl <r...@tinyclouds.org> wrote:
> 2010.11.16, Version 0.2.5
>
> * Add ref to buffer during fs.write and fs.read. Sometimes buffers
>   would be GCed before making it to the thread pool.
>
> * Fix http buffer pushing bug:

>  http://groups.google.com/group/nodejs/browse_thread/thread/f66cd3c960...

[Ben Noordhuis]

On Wed, Nov 17, 2010 at 20:42, Oleg "Sannis" Efimov <efim...@gmail.com> wrote:
>> * writeFileSync could exhibit pathological behavior when a buffer could
>>   not be written to the file in a single write() call. (isaacs)
>
> Hm, as I understand, this is a bug? Please, mybe someone can explain
> me this in more simple way :-)

Node would go into a busy loop (100% CPU usage). It eventually also
died with an out-of-memory error for me but I don't know if that was
reproducible behaviour.