Re: [nodejs] Is nodejs good to be a video live broadcast server?

[jason . 桂林]

Hi darren, what should I know for streams.

What is  "Microsoft MMSH video streams", What should I know for flash video
stream and html video stream up/down, and how IOS/android works for video
stream.

2012/6/13 darren 

> You might find this interesting: https://github.com/orospakr/stream-pump/
>
>
> On Tuesday, June 12, 2012 11:33:50 PM UTC-4, Jason.桂林(Gui Lin) wrote:
>>
>> I want to make a video live broadcast server, I don't know how to make
>> one, and I don't if nodejs is good to do this thing, I hope guys could give
>> me some suggestion.
>>
>> --
>> Best regards,
>>
>> 桂林 (Gui Lin)
>>
>> guileen@twitter 
>> 桂糊涂@weibo 
>> guileen@github 
>>
>>  --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en
>



-- 
Best regards,

桂林 (Gui Lin)

guileen@twitter 
桂糊涂@weibo 
guileen@github 

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

[nodejs] Re: starting a nodejs app during boot

[mscdex]

I use upstart which can additionally function as an alternative to
forever.

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Matt]

On Tue, Jun 12, 2012 at 10:47 PM, Stewart Mckinney wrote:

> A key thing to understand about v8 is that it JIT compiles your Javascript
> straight into machine language ( via the meta-lang pipeline: Crankshaft->
> Helium-> Lithium ). This makes context switches far, far more costly,
> because v8 does not actually deal directly ( much ) with its abstractions.
> It does use them in formulation but then discards them when no longer
> needed. This means if you go a-fetching v8 has to do some searching and
> constructing before it can make your switch, unless you've asked v8 to
> 'hold onto' something ( which in turn, slows down the VM slightly via the
> GC and forces some parts of the VM to not optimize because your essentially
> telling the interpreter "this is volatile" ). This is part of the reason
> why v8 encourages you to use Local Handles whenever possible.
>
> The smackdown here isn't that hard  - your looking at about 0.00032ms or
> so ( off the top of my head ) - but you take an array of say, a million
> elements, and there you go. Suddenly your looking at 1/3 seconds time just
> to iterate over a massive JS array in C++.
>

Sure, but IIRC neither NumPy nor PDL concepts of Arrays/Vectors/whatever
are strictly language native. So it's basically the same as a Buffer.

(sorry I guess this should be another thread).

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] starting a nodejs app during boot

[Hack Sparrow]

Hi Angelo,

This post will help you - 
http://www.hacksparrow.com/make-forever-reboot-proof-with-cron.html


Captain



Angelo Chen wrote:


Hi,
I use forever to keep nodejs app running, this works quite well if I
started it from command line:

usr/local/bin/forever start -c /usr/local/bin/node -a -l /home/photos/
logs/resizer.log -o /home/photos/logs/resizer_out.log -e /home/photos/
logs/resizer_err.log /home/photos/resizer.js

but if I start this during boot in rc.local, it does not start, got a
message ：

/usr/bin/env: node: No such file or directory

？


--
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Is nodejs good to be a video live broadcast server?

[darren]

You might find this interesting: https://github.com/orospakr/stream-pump/

On Tuesday, June 12, 2012 11:33:50 PM UTC-4, Jason.桂林(Gui Lin) wrote:
>
> I want to make a video live broadcast server, I don't know how to make 
> one, and I don't if nodejs is good to do this thing, I hope guys could give 
> me some suggestion.
>
> -- 
> Best regards,
>
> 桂林 (Gui Lin)
>
> guileen@twitter 
> 桂糊涂@weibo 
> guileen@github 
>
>

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

[nodejs] starting a nodejs app during boot

[Angelo Chen]

Hi,
I use forever to keep nodejs app running, this works quite well if I
started it from command line:

usr/local/bin/forever start -c /usr/local/bin/node -a -l /home/photos/
logs/resizer.log -o /home/photos/logs/resizer_out.log -e /home/photos/
logs/resizer_err.log  /home/photos/resizer.js

but if I start this during boot in rc.local, it does not start, got a
message ：

/usr/bin/env: node: No such file or directory

？

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

[nodejs] Is nodejs good to be a video live broadcast server?

[jason . 桂林]

I want to make a video live broadcast server, I don't know how to make one,
and I don't if nodejs is good to do this thing, I hope guys could give me
some suggestion.

-- 
Best regards,

桂林 (Gui Lin)

guileen@twitter 
桂糊涂@weibo 
guileen@github 

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Stewart Mckinney]

It's important to understand that the overhead of context switching in V8
is pretty painful when compared to Python. I'm doing a bit of work myself
embedding v8 in a fairly intensive environment, and it's been a hell of a
learning experience. There's a *metric s*** ton* of difference between the
standard Python interpreter and v8 and how they are interpreted.

A key thing to understand about v8 is that it JIT compiles your Javascript
straight into machine language ( via the meta-lang pipeline: Crankshaft->
Helium-> Lithium ). This makes context switches far, far more costly,
because v8 does not actually deal directly ( much ) with its abstractions.
It does use them in formulation but then discards them when no longer
needed. This means if you go a-fetching v8 has to do some searching and
constructing before it can make your switch, unless you've asked v8 to
'hold onto' something ( which in turn, slows down the VM slightly via the
GC and forces some parts of the VM to not optimize because your essentially
telling the interpreter "this is volatile" ). This is part of the reason
why v8 encourages you to use Local Handles whenever possible.

The smackdown here isn't that hard  - your looking at about 0.00032ms or so
( off the top of my head ) - but you take an array of say, a million
elements, and there you go. Suddenly your looking at 1/3 seconds time just
to iterate over a massive JS array in C++.

Compare this to Python, or Ruby, or Perl, most interpreters of which keep
their abstractions available ( in memory ) and deal with them directly.
Context switches are not as painful because Python/Ruby/Perl do not have to
do any building up on "fetch" or "switch", they simply grab the
abstraction's pointer and go. So its really best to limit those as much as
possible.

If you do proceed, some things to keep in mind:

-As I mentioned, big arrays are painful if you need to access them in C. I
would recommend following Node's lead and using "buffer" patterns where
appropriate. In other words, try to make your switches as 'meaty' as
possible.

-It seems to be more painful to access JS from C++ ( via ->Get() ) then it
does to callback C++ from JS ( via FunctionTemplates ).

-... but it doesn't seem that painful to call JS functions, or to create
one-use JS Objects for your arguments in C++ ( via Function->Call() and
v8::Object::New ).

These are just my (not particularly scientific) observations.

Best of luck!

On Tue, Jun 12, 2012 at 6:29 PM, Joshua Holbrook wrote:

> > people said that at first about numerical calculations in Python before
> NumPy existed
>
> I doubt it. Scientists *love* having a scripting language on top of
> their fortrans. It's a very old idea.
>
> --Josh
>
> On Tue, Jun 12, 2012 at 3:15 PM, Matt  wrote:
> > On Tue, Jun 12, 2012 at 6:09 PM, Joshua Holbrook <
> josh.holbr...@gmail.com>
> > wrote:
> >>
> >> > it's not something that can't be done, it just hasn't been done yet,
> >> > IMHO.
> >>
> >> Right. It's doable, it's just not actually that great of an idea.
> >
> >
> > Hah, well people said that at first about numerical calculations in
> Python
> > before NumPy existed, and Perl before PDL existed - after all that's what
> > Mathematica and FORTRAN were for ;-)
> >
> > --
> > 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 nodejs@googlegroups.com
> > To unsubscribe from this group, send email to
> > nodejs+unsubscr...@googlegroups.com
> > For more options, visit this group at
> > http://groups.google.com/group/nodejs?hl=en?hl=en
>
>
>
> --
> Joshua Holbrook
> Engineer
> Nodejitsu Inc.
> j...@nodejitsu.com
>
> --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en
>

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Re: Isolates removed

[Stewart Mckinney]

+1

Please start a new post and explain why you need 10k apps ( thats a huge
effing number in that context ); for a particular application, or just for
research?

On Tue, Jun 12, 2012 at 9:24 PM, Marak Squires wrote:

> I'd recommend starting a new mailing list post explaining the problem you
> are attempting to solve. Someone should be able to help guide you towards
> the correct architecture.
>
> Better to figure out a general solution before trying to figure out the
> specific technical limitations of a potential implementation.
>
> On Tue, Jun 12, 2012 at 6:15 PM, Andrew Finnell 
> wrote:
>
>> Thank you for the reply,
>>
>> I did come across your product. So if you put any work into getting
>> Google to bring up your product, it worked.
>>
>> I cannot use any application management system until I solve the problem
>> of being able to run 10,000 instances of an "app". Ultimately I have
>> control over how these apps are written so I can create a framework in
>> nodejs that was essentially a nodejs fragment. But I like the idea of
>> creating a system that allows one to run the app outside of the system as
>> well as inside of it. In all reality, without writing an app framework,
>> I'll have this same issue in Erland and Java. Spawning 10,000 threads in
>> Java is also unreasonable and smells of a bad design.
>>
>> Clustering multiple of these "apps" in a single nodejs isn't out of the
>> issue. But they would have to be agnostic as to which apps are running and
>> I would need to control how many apps run in X number of nodes. I saw
>> threads a gogo and thought this would help but after reviewing it, it
>> solves another issue I was having (CPU intensive calculations).
>>
>> I think I may have a design issue to work out first that has nothing to
>> do with NodeJS. I have looked at Erlang before, mostly because that is what
>> RabbitMQ was written in. I really like what nodejs would give me. The
>> ability to write short, event driven I/O applications. That is exactly what
>> I need. It's just a matter of finding a way to run thousands of these.
>>
>> The requirements are:
>>
>> Apps should be code complete. All the code needed to run the app is there
>> (no fragments)
>> Apps cannot bring down other apps if they crash
>>
>> --
>> 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 nodejs@googlegroups.com
>> To unsubscribe from this group, send email to
>> nodejs+unsubscr...@googlegroups.com
>> For more options, visit this group at
>> http://groups.google.com/group/nodejs?hl=en?hl=en
>>
>
>
>
> --
> --
> Marak Squires
> Co-founder and Chief Evangelist
> Nodejitsu, Inc.
> marak.squi...@gmail.com
>
>  --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en
>

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Re: Isolates removed

[Marak Squires]

I'd recommend starting a new mailing list post explaining the problem you
are attempting to solve. Someone should be able to help guide you towards
the correct architecture.

Better to figure out a general solution before trying to figure out the
specific technical limitations of a potential implementation.

On Tue, Jun 12, 2012 at 6:15 PM, Andrew Finnell wrote:

> Thank you for the reply,
>
> I did come across your product. So if you put any work into getting Google
> to bring up your product, it worked.
>
> I cannot use any application management system until I solve the problem
> of being able to run 10,000 instances of an "app". Ultimately I have
> control over how these apps are written so I can create a framework in
> nodejs that was essentially a nodejs fragment. But I like the idea of
> creating a system that allows one to run the app outside of the system as
> well as inside of it. In all reality, without writing an app framework,
> I'll have this same issue in Erland and Java. Spawning 10,000 threads in
> Java is also unreasonable and smells of a bad design.
>
> Clustering multiple of these "apps" in a single nodejs isn't out of the
> issue. But they would have to be agnostic as to which apps are running and
> I would need to control how many apps run in X number of nodes. I saw
> threads a gogo and thought this would help but after reviewing it, it
> solves another issue I was having (CPU intensive calculations).
>
> I think I may have a design issue to work out first that has nothing to do
> with NodeJS. I have looked at Erlang before, mostly because that is what
> RabbitMQ was written in. I really like what nodejs would give me. The
> ability to write short, event driven I/O applications. That is exactly what
> I need. It's just a matter of finding a way to run thousands of these.
>
> The requirements are:
>
> Apps should be code complete. All the code needed to run the app is there
> (no fragments)
> Apps cannot bring down other apps if they crash
>
> --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en
>



-- 
-- 
Marak Squires
Co-founder and Chief Evangelist
Nodejitsu, Inc.
marak.squi...@gmail.com

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Re: Isolates removed

[Andrew Finnell]

Thank you for the reply,

I did come across your product. So if you put any work into getting Google 
to bring up your product, it worked.

I cannot use any application management system until I solve the problem of 
being able to run 10,000 instances of an "app". Ultimately I have control 
over how these apps are written so I can create a framework in nodejs that 
was essentially a nodejs fragment. But I like the idea of creating a system 
that allows one to run the app outside of the system as well as inside of 
it. In all reality, without writing an app framework, I'll have this same 
issue in Erland and Java. Spawning 10,000 threads in Java is also 
unreasonable and smells of a bad design. 

Clustering multiple of these "apps" in a single nodejs isn't out of the 
issue. But they would have to be agnostic as to which apps are running and 
I would need to control how many apps run in X number of nodes. I saw 
threads a gogo and thought this would help but after reviewing it, it 
solves another issue I was having (CPU intensive calculations). 

I think I may have a design issue to work out first that has nothing to do 
with NodeJS. I have looked at Erlang before, mostly because that is what 
RabbitMQ was written in. I really like what nodejs would give me. The 
ability to write short, event driven I/O applications. That is exactly what 
I need. It's just a matter of finding a way to run thousands of these.  

The requirements are:

Apps should be code complete. All the code needed to run the app is there 
(no fragments)
Apps cannot bring down other apps if they crash

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Re: Isolates removed

[Marak Squires]

If you are looking to spawn 10k worker nodes all on the same machine, you
are probably doing it wrong. Try scaling out, instead of up. Maybe node
isn't the right choice for what you are trying to do.

As for spawning up several hundred instances of a node.js application,
we've put a lot of work into our node application server haibu (
https://github.com/nodejitsu/haibu )


On Tue, Jun 12, 2012 at 5:06 PM, Andrew Finnell wrote:

> Isolates was my last saving grace while doing NodeJS research. I am
> attempting to do a proof of concept for a product I am working on.
>
> Scenario:
>
> 10,000 distinct'ish workers on a single machine. As it stands I'd need to
> spawn 10,000 separate nodejs instances which amounts to about 300 gigs of
> memory for a very simple task. I was hoping that Isolates or something
> similar that could share common memory in some fashion would of alleviated
> this problem. I could maybe have 200 nodejs clusters that each run 50
> internal instances. I do not need to share variables etc across threads or
> processes. I am very appreciate and understanding of the mailbox/message
> model for communication between processes. I just need to reduce the
> handle/memory requirements to something more manageable.
>
> Of all the workers, the idea is that a single one can go down and I still
> have all the other ones working. In a typical Java implementation one might
> create a cluster of a few VM nodes 3-5. Even in this scenario 1 node going
> down would bring down a substantial number of workers and the memory
> overhead of running a JVM is too high.
>
> In a little more detail, lets say I wanted to create a management
> framework that allowed developers to write their own nodejs applications
> that could run in the management framework. I might need to spawn 400
> instances of App A, 100 of App B, 90 or App C, etc. The separate instances
> of the Apps do not share state nor need to. The key issue I am facing is
> that these app's are written independently of each other, thus it makes it
> difficult to run them all in the same event loop. Too many bad things can
> happen with differently written apps.
>
>
> Any suggestions? I have been looking into Vert.x as an alternative.
>
> -
> Andrew
>
> On Friday, February 3, 2012 4:28:54 PM UTC-5, Isaac Schlueter wrote:
>>
>> The Isolates feature was intended to make it possible to run
>> child_process.fork() in a thread, rather than a full process.  The
>> justification was to make it cheaper to spin up new child node
>> instances, as well as allowing for fast message-passing using shared
>> memory in binary addons, while retaining the semantics of node's
>> child_process implementation by keeping them in completely isolated v8
>> instances.
>>
>> It was a very informative experiment, but has ultimately turned out to
>> cause too much instability in node's internal functionality to justify
>> continuing with it at this time.  It requires a lot of complexity to
>> be added to libuv and node, and isn't likely to yield enough gains to
>> be worth the investment.
>>
>> We're going to roll back the changes in master that were added to
>> support Isolates, and instead focus on Domains and other things that
>> promise to increase stability and make debugging easier.  This change
>> will land in 0.7.3.  It's entirely possible that we may decide to use
>> v8 isolates in some future version of node, but they will not be in
>> 0.8.
>>
>> If you were eagerly looking forward to using this feature, and find
>> yourself shocked or frustrated by this decision, please contact me
>> directly.  It's not our intention to leave anyone stuck, and I do
>> understand that this feature was promised for some time.  If this
>> causes hardship for you, let's figure out another way to get your use
>> cases handled.
>>
>> It's never easy to back-pedal, but doing experimental things means
>> sometimes finding out that you were headed in the wrong direction.
>> The sooner we make this change, the easier it will be.
>>
>> Thanks.
>>
>>  --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en
>



-- 
-- 
Marak Squires
Co-founder and Chief Evangelist
Nodejitsu, Inc.
marak.squi...@gmail.com

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
ht

Re: [nodejs] Re: Isolates removed

[Ben Noordhuis]

On Wed, Jun 13, 2012 at 2:06 AM, Andrew Finnell
 wrote:
> Isolates was my last saving grace while doing NodeJS research. I am
> attempting to do a proof of concept for a product I am working on.
>
> Scenario:
>
> 10,000 distinct'ish workers on a single machine. As it stands I'd need to
> spawn 10,000 separate nodejs instances which amounts to about 300 gigs of
> memory for a very simple task. I was hoping that Isolates or something
> similar that could share common memory in some fashion would of alleviated
> this problem. I could maybe have 200 nodejs clusters that each run 50
> internal instances. I do not need to share variables etc across threads or
> processes. I am very appreciate and understanding of the mailbox/message
> model for communication between processes. I just need to reduce the
> handle/memory requirements to something more manageable.
>
> Of all the workers, the idea is that a single one can go down and I still
> have all the other ones working. In a typical Java implementation one might
> create a cluster of a few VM nodes 3-5. Even in this scenario 1 node going
> down would bring down a substantial number of workers and the memory
> overhead of running a JVM is too high.
>
> In a little more detail, lets say I wanted to create a management framework
> that allowed developers to write their own nodejs applications that could
> run in the management framework. I might need to spawn 400 instances of App
> A, 100 of App B, 90 or App C, etc. The separate instances of the Apps do not
> share state nor need to. The key issue I am facing is that these app's are
> written independently of each other, thus it makes it difficult to run them
> all in the same event loop. Too many bad things can happen with differently
> written apps.
>
>
> Any suggestions? I have been looking into Vert.x as an alternative.

Isolates wouldn't have helped you. Each isolate has a 5-10 MB memory
footprint so 10,000 isolates need at least 50 GB of memory. Never mind
that all the context switching will likely kill you.

Java or Erlang may work better for you. It's certainly possible to
fire up 10K Java threads but you probably need to tweak the JVM a
little (decrease stack size, that sort of thing.)

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

[nodejs] Re: Isolates removed

[Andrew Finnell]

Isolates was my last saving grace while doing NodeJS research. I am 
attempting to do a proof of concept for a product I am working on.

Scenario:

10,000 distinct'ish workers on a single machine. As it stands I'd need to 
spawn 10,000 separate nodejs instances which amounts to about 300 gigs of 
memory for a very simple task. I was hoping that Isolates or something 
similar that could share common memory in some fashion would of alleviated 
this problem. I could maybe have 200 nodejs clusters that each run 50 
internal instances. I do not need to share variables etc across threads or 
processes. I am very appreciate and understanding of the mailbox/message 
model for communication between processes. I just need to reduce the 
handle/memory requirements to something more manageable. 

Of all the workers, the idea is that a single one can go down and I still 
have all the other ones working. In a typical Java implementation one might 
create a cluster of a few VM nodes 3-5. Even in this scenario 1 node going 
down would bring down a substantial number of workers and the memory 
overhead of running a JVM is too high. 

In a little more detail, lets say I wanted to create a management framework 
that allowed developers to write their own nodejs applications that could 
run in the management framework. I might need to spawn 400 instances of App 
A, 100 of App B, 90 or App C, etc. The separate instances of the Apps do 
not share state nor need to. The key issue I am facing is that these app's 
are written independently of each other, thus it makes it difficult to run 
them all in the same event loop. Too many bad things can happen with 
differently written apps. 


Any suggestions? I have been looking into Vert.x as an alternative. 

-
Andrew

On Friday, February 3, 2012 4:28:54 PM UTC-5, Isaac Schlueter wrote:
>
> The Isolates feature was intended to make it possible to run
> child_process.fork() in a thread, rather than a full process.  The
> justification was to make it cheaper to spin up new child node
> instances, as well as allowing for fast message-passing using shared
> memory in binary addons, while retaining the semantics of node's
> child_process implementation by keeping them in completely isolated v8
> instances.
>
> It was a very informative experiment, but has ultimately turned out to
> cause too much instability in node's internal functionality to justify
> continuing with it at this time.  It requires a lot of complexity to
> be added to libuv and node, and isn't likely to yield enough gains to
> be worth the investment.
>
> We're going to roll back the changes in master that were added to
> support Isolates, and instead focus on Domains and other things that
> promise to increase stability and make debugging easier.  This change
> will land in 0.7.3.  It's entirely possible that we may decide to use
> v8 isolates in some future version of node, but they will not be in
> 0.8.
>
> If you were eagerly looking forward to using this feature, and find
> yourself shocked or frustrated by this decision, please contact me
> directly.  It's not our intention to leave anyone stuck, and I do
> understand that this feature was promised for some time.  If this
> causes hardship for you, let's figure out another way to get your use
> cases handled.
>
> It's never easy to back-pedal, but doing experimental things means
> sometimes finding out that you were headed in the wrong direction.
> The sooner we make this change, the easier it will be.
>
> Thanks.
>
>

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Nodejs bi-dimensional array [][] table

[Mark Hahn]

The question assumes arrays are pre-allocated.  They aren't.  They are only
added dynamically.

So every array starts as one-dimension and then later can be looked at as
two-dimensional if more arrays are added inside.  Even then the analogy to
two-dimensional arrays can break down if you add something like an integer
aside the other internal arrays.

On Tue, Jun 12, 2012 at 3:28 PM, Ben Noordhuis  wrote:

> On Wed, Jun 13, 2012 at 12:21 AM, vitoa  wrote:
> > Hi, I want to create a empty bidimensinal array (table) in node js but
> > i'm not going in good way
> >
> > For one dimension array its simple like  var array=[]
> > But for two dimensional array I can not use var table = [][]
> >
> > How can I do this?
>
> This question would have been more appropriate for
> comp.lang.javascript. The syntax you're looking for is `var a = [[]]`.
>
> --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en
>

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Joshua Holbrook]

> people said that at first about numerical calculations in Python before NumPy 
> existed

I doubt it. Scientists *love* having a scripting language on top of
their fortrans. It's a very old idea.

--Josh

On Tue, Jun 12, 2012 at 3:15 PM, Matt  wrote:
> On Tue, Jun 12, 2012 at 6:09 PM, Joshua Holbrook 
> wrote:
>>
>> > it's not something that can't be done, it just hasn't been done yet,
>> > IMHO.
>>
>> Right. It's doable, it's just not actually that great of an idea.
>
>
> Hah, well people said that at first about numerical calculations in Python
> before NumPy existed, and Perl before PDL existed - after all that's what
> Mathematica and FORTRAN were for ;-)
>
> --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en



-- 
Joshua Holbrook
Engineer
Nodejitsu Inc.
j...@nodejitsu.com

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Nodejs bi-dimensional array [][] table

[Ben Noordhuis]

On Wed, Jun 13, 2012 at 12:21 AM, vitoa  wrote:
> Hi, I want to create a empty bidimensinal array (table) in node js but
> i'm not going in good way
>
> For one dimension array its simple like  var array=[]
> But for two dimensional array I can not use var table = [][]
>
> How can I do this?

This question would have been more appropriate for
comp.lang.javascript. The syntax you're looking for is `var a = [[]]`.

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

[nodejs] Nodejs bi-dimensional array [][] table

[vitoa]

Hi, I want to create a empty bidimensinal array (table) in node js but
i'm not going in good way

For one dimension array its simple like  var array=[]
But for two dimensional array I can not use var table = [][]

How can I do this?

Regards

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Matt]

On Tue, Jun 12, 2012 at 6:09 PM, Joshua Holbrook wrote:

> > it's not something that can't be done, it just hasn't been done yet,
> IMHO.
>
> Right. It's doable, it's just not actually that great of an idea.
>

Hah, well people said that at first about numerical calculations in Python
before NumPy existed, and Perl before PDL existed - after all that's what
Mathematica and FORTRAN were for ;-)

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Joshua Holbrook]

> it's not something that can't be done, it just hasn't been done yet, IMHO.

Right. It's doable, it's just not actually that great of an idea.

--Josh

On Tue, Jun 12, 2012 at 3:08 PM, Matt  wrote:
> On Tue, Jun 12, 2012 at 5:34 PM, Joshua Holbrook 
> wrote:
>>
>> Python has a python/c barrier yes, but python's runtime is better
>> equipped to handle crossing the boundary with low overhead because it
>> was written with that use case in mind. v8, on the other hand, is
>> optimized for running javascript code and so making the jump back and
>> forth is considerably more effort for the vm.
>>
>> I'm sure someone with more experience with node C bindings (or python
>> C bindings for that matter) could give a more technical explanation.
>> I'm more of an end-user and not as much a developer when it comes to
>> such bindings.
>
>
> I'm familiar with Node's C++ bindings and I'm very familiar with Perl's C
> interface (XS) and the overheads of passing objects back and forth there.
> And Perl was designed very similarly to Python in that sense. The answer is
> there's not that much difference between Node and Perl in that respect, and
> I haven't seen any reason the equivalent to NumPy (or PDL which is the Perl
> equivalent) couldn't be built in Node from my knowledge of the internals.
>
> Not that this exists now, or is even in the scope of a masters thesis, but
> I'm just saying it's not something that can't be done, it just hasn't been
> done yet, IMHO.
>
> Matt.
>
> --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en



-- 
Joshua Holbrook
Engineer
Nodejitsu Inc.
j...@nodejitsu.com

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Matt]

On Tue, Jun 12, 2012 at 5:34 PM, Joshua Holbrook wrote:

> Python has a python/c barrier yes, but python's runtime is better
> equipped to handle crossing the boundary with low overhead because it
> was written with that use case in mind. v8, on the other hand, is
> optimized for running javascript code and so making the jump back and
> forth is considerably more effort for the vm.
>
> I'm sure someone with more experience with node C bindings (or python
> C bindings for that matter) could give a more technical explanation.
> I'm more of an end-user and not as much a developer when it comes to
> such bindings.
>

I'm familiar with Node's C++ bindings and I'm very familiar with Perl's C
interface (XS) and the overheads of passing objects back and forth there.
And Perl was designed very similarly to Python in that sense. The answer is
there's not that much difference between Node and Perl in that respect, and
I haven't seen any reason the equivalent to NumPy (or PDL which is the Perl
equivalent) couldn't be built in Node from my knowledge of the internals.

Not that this exists now, or is even in the scope of a masters thesis, but
I'm just saying it's not something that can't be done, it just hasn't been
done yet, IMHO.

Matt.

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Dean Mao]

If you currently do your math using matlab, I would suggest numpy.  There's
some documents on the scipy site that will help you transition your matlab
knowledge into numpy & python syntax.  There's no good reason for you to do
it in nodejs unless you want to make a significant math library
contribution.


On Tue, Jun 12, 2012 at 2:43 PM, Igor Skoric  wrote:

> On Tuesday, June 12, 2012 9:20:40 PM UTC+2, Ben Noordhuis wrote:
>>
>> On Tue, Jun 12, 2012 at 3:08 PM, Igor Skoric  wrote:
>> > It will be a suite of tools for a geological lab. I plan to create a
>> simple
>> > customer/project management suite and plug in a handful of applications
>> for
>> > different processes. If you would like to recommend a different (but
>> cutting
>> > edge) technology for this I would also be grateful.
>>
>> Igor, see my previous email. node.js may or may not work for you. If
>> you're already familiar with, say, flask or django, I'd go with that.
>> The extra learning curve of a new language and framework will just get
>> in the way of getting things done.
>>
>
> Unfortunately I'm not familiar with either. It's going to be a learning
> experience either way. :)
>
> --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en
>

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Dean Mao]

Well for one, scipy & numpy already exist.  If you want to do something
like linear regression, you'll have to build all kinds of base math
libraries into node before you can start working on anything.  It's great
if you want to create an equivalent scipy for node, but probably not in the
scope of a master's thesis.


On Tue, Jun 12, 2012 at 2:18 PM, Matt  wrote:

> On Tue, Jun 12, 2012 at 3:29 PM, Joshua Holbrook 
> wrote:
>
>> To elaborate on the math aspect: Because of the significant js/C
>> barrier and the need for optimized code, node is a poor choice for
>> significant computation. I would recommend python and numpy/scipy for
>> this kinda stuff.
>>
>
> Python has the same barrier. What makes it different? i.e. why couldn't
> someone created a NumNode/SciNode?
>
> Matt.
>
> --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en
>

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Igor Skoric]

On Tuesday, June 12, 2012 9:20:40 PM UTC+2, Ben Noordhuis wrote:
>
> On Tue, Jun 12, 2012 at 3:08 PM, Igor Skoric  wrote: 
> > It will be a suite of tools for a geological lab. I plan to create a 
> simple 
> > customer/project management suite and plug in a handful of applications 
> for 
> > different processes. If you would like to recommend a different (but 
> cutting 
> > edge) technology for this I would also be grateful. 
>
> Igor, see my previous email. node.js may or may not work for you. If 
> you're already familiar with, say, flask or django, I'd go with that. 
> The extra learning curve of a new language and framework will just get 
> in the way of getting things done. 
>

Unfortunately I'm not familiar with either. It's going to be a learning 
experience either way. :) 

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Igor Skoric]

On Tuesday, June 12, 2012 9:29:10 PM UTC+2, jesusabdullah wrote:
>
> > things like hard real-time, complex math, etc. 
>
> To elaborate on the math aspect: Because of the significant js/C 
> barrier and the need for optimized code, node is a poor choice for 
> significant computation. I would recommend python and numpy/scipy for 
> this kinda stuff. 
>

The computations are not very complex. Computing performance is not 
important and the computation time is insignificant. But I still find your 
recommendations for that case (should i have it come up) interesting.
 

> > The extra learning curve of a new language and framework will just get 
> in the way of getting things done. 
>
> There's definitely a truth to this. When I did my masters, the biggest 
> lesson was that your time is worth *way* more than both money and 
> cool-factor. Flask might be boring but it's worth it if you're able to 
> ship faster in the short term (and with the timescale of a masters you 
> definitely want to be thinking short-term). Just something to think 
> about. 
>

Right now I'm more concerned about the right topic for my thesis.
I have not worked with django or flask before. I have coded some pure 
php/js and i don't like PHP very much.

The planned scope is about 6-12 months.

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Joshua Holbrook]

Python has a python/c barrier yes, but python's runtime is better
equipped to handle crossing the boundary with low overhead because it
was written with that use case in mind. v8, on the other hand, is
optimized for running javascript code and so making the jump back and
forth is considerably more effort for the vm.

I'm sure someone with more experience with node C bindings (or python
C bindings for that matter) could give a more technical explanation.
I'm more of an end-user and not as much a developer when it comes to
such bindings.

To further elaborate: You *can* do basic math in javascript just fine,
and at a reasonable rate. For heavy lifting though, ya gotta go
lapack/atlas/etc.

--Josh

On Tue, Jun 12, 2012 at 2:18 PM, Matt  wrote:
> On Tue, Jun 12, 2012 at 3:29 PM, Joshua Holbrook 
> wrote:
>>
>> To elaborate on the math aspect: Because of the significant js/C
>> barrier and the need for optimized code, node is a poor choice for
>> significant computation. I would recommend python and numpy/scipy for
>> this kinda stuff.
>
>
> Python has the same barrier. What makes it different? i.e. why couldn't
> someone created a NumNode/SciNode?
>
> Matt.
>
> --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en



-- 
Joshua Holbrook
Engineer
Nodejitsu Inc.
j...@nodejitsu.com

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Matt]

On Tue, Jun 12, 2012 at 3:29 PM, Joshua Holbrook wrote:

> To elaborate on the math aspect: Because of the significant js/C
> barrier and the need for optimized code, node is a poor choice for
> significant computation. I would recommend python and numpy/scipy for
> this kinda stuff.
>

Python has the same barrier. What makes it different? i.e. why couldn't
someone created a NumNode/SciNode?

Matt.

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Joshua Holbrook]

> things like hard real-time, complex math, etc.

To elaborate on the math aspect: Because of the significant js/C
barrier and the need for optimized code, node is a poor choice for
significant computation. I would recommend python and numpy/scipy for
this kinda stuff.

One approach you may consider is to use node to manage basic web ui
stuff, and then use the child_process module to shell out to workers
written in more appropriate languages as necessary. For a scientific
application this would probably even be a good fit.

> The extra learning curve of a new language and framework will just get
in the way of getting things done.

There's definitely a truth to this. When I did my masters, the biggest
lesson was that your time is worth *way* more than both money and
cool-factor. Flask might be boring but it's worth it if you're able to
ship faster in the short term (and with the timescale of a masters you
definitely want to be thinking short-term). Just something to think
about.

--Josh


On Tue, Jun 12, 2012 at 12:20 PM, Ben Noordhuis  wrote:
> On Tue, Jun 12, 2012 at 3:08 PM, Igor Skoric  wrote:
>> It will be a suite of tools for a geological lab. I plan to create a simple
>> customer/project management suite and plug in a handful of applications for
>> different processes. If you would like to recommend a different (but cutting
>> edge) technology for this I would also be grateful.
>
> Igor, see my previous email. node.js may or may not work for you. If
> you're already familiar with, say, flask or django, I'd go with that.
> The extra learning curve of a new language and framework will just get
> in the way of getting things done.
>
> --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en



-- 
Joshua Holbrook
Engineer
Nodejitsu Inc.
j...@nodejitsu.com

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Re: # Fatal error in ../deps/v8/src/objects-inl.h, line 2131 # CHECK(object->IsString()) failed

[Ben Noordhuis]

On Tue, Jun 12, 2012 at 3:14 PM, Debbio  wrote:
> A look in node/deps/v8/Changelog reports that v8 version is  3.9.24

Can you open an issue? I'm not able to reproduce it but maybe others can.

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Ben Noordhuis]

On Tue, Jun 12, 2012 at 3:08 PM, Igor Skoric  wrote:
> It will be a suite of tools for a geological lab. I plan to create a simple
> customer/project management suite and plug in a handful of applications for
> different processes. If you would like to recommend a different (but cutting
> edge) technology for this I would also be grateful.

Igor, see my previous email. node.js may or may not work for you. If
you're already familiar with, say, flask or django, I'd go with that.
The extra learning curve of a new language and framework will just get
in the way of getting things done.

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Ben Noordhuis]

On Tue, Jun 12, 2012 at 3:14 PM, Paul Spaulding  wrote:
> Ben, I am interested in what you consider to be poor holes for a node.js
> peg.

There are some things you cannot reasonably do in node.js no matter
how hard you try, things like hard real-time, complex math, etc.
They're fundamentally at odds with the language and the environment it
runs in.

Then there is the broad category of things you /can/ but not
necessarily /should/ do. CRUD web apps come to mind: you /can/ write
your new blog engine in node.js but it doesn't have a killer edge over
other languages or frameworks (except for client/server code sharing,
maybe.)

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Re: { [Error: socket hang up] code: 'ECONNRESET' }

[Faysal Banna]

   i use 302 just to tell the browser that the file is temporary moved 
to another location which is the location that i set after removing the 
range=xxx-yyy parameter in the query,
previously i didn't use 302 just used whatever the FwRes.statusCode 
gave and as well the same issue


   another thing i noticed as well is that the memory always increasing 
and never drops down as if the data fetched never gets free after being 
served


much regards


On 06/09/2012 04:32 PM, Soren wrote:

Why are you using 302 ?

On Jun 9, 2:19 am, degreane  wrote:

var AdyRequestHandler = function(req,res){


..

 var FwReq = http.get(myheaders,function(FwRes){
 res.writeHead(302,FwRes['headers']);
 FwRes.on('data',function(chunk){
 res.write(chunk);
 });
 FwRes.on('end',function(){
 res.end();
 });
 });


--
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Patterns for dealing with command+response type connections, specifically TL1.

[Elliot]

Alternatively:  https://github.com/andris9/rai

On Tue, Jun 12, 2012 at 7:34 AM, Matt  wrote:

> Yes you need a state machine. This is basically how SMTP works, so you
> should look into one of the SMTP modules like Haraka.
>
>
> On Tue, Jun 12, 2012 at 10:14 AM, Andrew Finnell  > wrote:
>
>> I have been investigating NodeJS for quite some time. I see the benefits
>> of the asynchronous nature and the pool of talent available to write in
>> JavaScript.
>>
>> I have having a little difficulty understanding how I might write a
>> system to deal with TL1 type devices. For those that do not know TL1 is a
>> command based protocol that can run over TCP/IP. You can do things like:
>>
>> > CONNECT (id) username/password
>> < Success (id)
>> RETRIEVE-DATA (id) parameter1
>> < data 1 (id)
>> < data 2 (id)
>> < asynchronous response (out-of-band) (id)
>> < asynchronous response (out-of-band) (someone elses id)
>>
>> I made up the commands to simplify the idea, but most of the time the
>> flow fits a synchronous model.
>>
>> I thought about creating a simple FSM around the current states of the
>> system such as connecting, asking for data etc, but this seems overly
>> complex for doing something so simple that can be done with TELNET and a
>> bash script file.
>>
>> Does anyone have patterns that they have successfully used to solve these
>> kinds of problems?
>>
>>
>>  --
>> 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 nodejs@googlegroups.com
>> To unsubscribe from this group, send email to
>> nodejs+unsubscr...@googlegroups.com
>> For more options, visit this group at
>> http://groups.google.com/group/nodejs?hl=en?hl=en
>>
>
>  --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en
>

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

[nodejs] Re: [node-dev] Re: Version 0.7.10 (unstable, but please read anyway)

[Isaac Schlueter]

It does not require admin rights (Junctions behave a lot more like
unix symlinks in node now.)

I'm not 100% sure offhand which versions it works in.  I've only
tested it in Windows 7, but I think Junctions haven't really changed
since XP.  Please try it out and report your findings ;)


On Tue, Jun 12, 2012 at 12:04 AM, Jan Kolkmeier  wrote:
>> Enable npm link on windows
>
> Awesome - all windows versions, or just 7? I assume it requires admin
> rights?
>
>
> On Monday, June 11, 2012 5:59:52 PM UTC+2, Isaac Schlueter wrote:
>>
>> 2012.06.11, Version 0.7.10 (unstable)
>>
>> This is the second-to-last release on the 0.7 branch.  Version 0.8.0
>> will be released some time next week.  As other even-numbered Node
>> releases before it, the v0.8.x releases will maintain API and binary
>> compatibility.
>>
>> The major changes are detailed in
>> https://github.com/joyent/node/wiki/API-changes-between-v0.6-and-v0.8
>>
>> Please try out this release.  There will be very few changes between
>> this and the v0.8.x release family.  This is the last chance to comment
>> on the API before it is locked down for stability.
>>
>>
>> * Roll V8 back to 3.9.24.31
>>
>> * build: x64 target should always pass -m64 (Robert Mustacchi)
>>
>> * add NODE_EXTERN to node::Start (Joel Brandt)
>>
>> * repl: Warn about running npm commands (isaacs)
>>
>> * slab_allocator: fix crash in dtor if V8 is dead (Ben Noordhuis)
>>
>> * slab_allocator: fix leak of Persistent handles (Shigeki Ohtsu)
>>
>> * windows/msi: add node.js prompt to startmenu (Jeroen Janssen)
>>
>> * windows/msi: fix adding node to PATH (Jeroen Janssen)
>>
>> * windows/msi: add start menu links when installing (Jeroen Janssen)
>>
>> * windows: don't install x64 version into the 'program files (x86)'
>> folder (Matt Gollob)
>>
>> * domain: Fix #3379 domain.intercept no longer passes error arg to cb
>> (Marc Harter)
>>
>> * fs: make callbacks run in global context (Ben Noordhuis)
>>
>> * fs: enable fs.realpath on windows (isaacs)
>>
>> * child_process: expose UV_PROCESS_DETACHED as options.detached
>> (Charlie McConnell)
>>
>> * child_process: new stdio API for .spawn() method (Fedor Indutny)
>>
>> * child_process: spawn().ref() and spawn().unref() (Fedor Indutny)
>>
>> * Upgrade npm to 1.1.25
>> - Enable npm link on windows
>> - Properly remove sh-shim on Windows
>> - Abstract out registry client and logger
>>
>>
>> Source Code: http://nodejs.org/dist/v0.7.10/node-v0.7.10.tar.gz
>>
>> Windows Installer: http://nodejs.org/dist/v0.7.10/node-v0.7.10.msi
>>
>> Windows x64 Files: http://nodejs.org/dist/v0.7.10/x64/
>>
>> Macintosh Installer (Universal):
>> http://nodejs.org/dist/v0.7.10/node-v0.7.10.pkg
>>
>> Other release files: http://nodejs.org/dist/v0.7.10/
>>
>> Website: http://nodejs.org/docs/v0.7.10/
>>
>> Documentation: http://nodejs.org/docs/v0.7.10/api/

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Patterns for dealing with command+response type connections, specifically TL1.

[Matt]

Yes you need a state machine. This is basically how SMTP works, so you
should look into one of the SMTP modules like Haraka.

On Tue, Jun 12, 2012 at 10:14 AM, Andrew Finnell
wrote:

> I have been investigating NodeJS for quite some time. I see the benefits
> of the asynchronous nature and the pool of talent available to write in
> JavaScript.
>
> I have having a little difficulty understanding how I might write a system
> to deal with TL1 type devices. For those that do not know TL1 is a command
> based protocol that can run over TCP/IP. You can do things like:
>
> > CONNECT (id) username/password
> < Success (id)
> RETRIEVE-DATA (id) parameter1
> < data 1 (id)
> < data 2 (id)
> < asynchronous response (out-of-band) (id)
> < asynchronous response (out-of-band) (someone elses id)
>
> I made up the commands to simplify the idea, but most of the time the flow
> fits a synchronous model.
>
> I thought about creating a simple FSM around the current states of the
> system such as connecting, asking for data etc, but this seems overly
> complex for doing something so simple that can be done with TELNET and a
> bash script file.
>
> Does anyone have patterns that they have successfully used to solve these
> kinds of problems?
>
>
>  --
> 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 nodejs@googlegroups.com
> To unsubscribe from this group, send email to
> nodejs+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/nodejs?hl=en?hl=en
>

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

[nodejs] Patterns for dealing with command+response type connections, specifically TL1.

[Andrew Finnell]

I have been investigating NodeJS for quite some time. I see the benefits of 
the asynchronous nature and the pool of talent available to write in 
JavaScript.

I have having a little difficulty understanding how I might write a system 
to deal with TL1 type devices. For those that do not know TL1 is a command 
based protocol that can run over TCP/IP. You can do things like:

> CONNECT (id) username/password
< Success (id)
RETRIEVE-DATA (id) parameter1 
< data 1 (id)
< data 2 (id)
< asynchronous response (out-of-band) (id)
< asynchronous response (out-of-band) (someone elses id)

I made up the commands to simplify the idea, but most of the time the flow 
fits a synchronous model.

I thought about creating a simple FSM around the current states of the 
system such as connecting, asking for data etc, but this seems overly 
complex for doing something so simple that can be done with TELNET and a 
bash script file.

Does anyone have patterns that they have successfully used to solve these 
kinds of problems?


-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Paul Spaulding]

Ben, I am interested in what you consider to be poor holes for a node.js 
peg.

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

[nodejs] Re: # Fatal error in ../deps/v8/src/objects-inl.h, line 2131 # CHECK(object->IsString()) failed

[Debbio]

A look in node/deps/v8/Changelog reports that v8 version is  3.9.24

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Igor Skoric]

It will be a suite of tools for a geological lab. I plan to create a simple 
customer/project management suite and plug in a handful of applications for 
different processes. If you would like to recommend a different (but 
cutting edge) technology for this I would also be grateful.

On Tuesday, June 12, 2012 2:59:24 PM UTC+2, Ben Noordhuis wrote:
>
> On Tue, Jun 12, 2012 at 2:27 PM, Igor Skoric  wrote: 
> > I'm developing a web application which aims to migrate at least two VB3 
> > 16-bit windows applications and a collection of other tools to a modern 
> > platform for the University of Technology Graz. 
> > Because I have a free pass to choose technologies and environments I aim 
> to 
> > contribute to the technologies that I use and document that contribution 
> and 
> > its implementation as part of my Master's Thesis in Software Development 
> and 
> > Business Management. 
> > 
> > I would appreciate your input - ideas that you would like to see become 
> > reality for node.js and also in general. 
> > I will choose one more of the ideas that fit and aim to make the sources 
> > available. 
> > 
> > Thank you for your help. 
> > Igor Skoric 
>
> Could you tell more about what you're trying to build / port? Node is 
> not necessarily the best tool for every project. 
>

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Ben Noordhuis]

On Tue, Jun 12, 2012 at 2:27 PM, Igor Skoric  wrote:
> I'm developing a web application which aims to migrate at least two VB3
> 16-bit windows applications and a collection of other tools to a modern
> platform for the University of Technology Graz.
> Because I have a free pass to choose technologies and environments I aim to
> contribute to the technologies that I use and document that contribution and
> its implementation as part of my Master's Thesis in Software Development and
> Business Management.
>
> I would appreciate your input - ideas that you would like to see become
> reality for node.js and also in general.
> I will choose one more of the ideas that fit and aim to make the sources
> available.
>
> Thank you for your help.
> Igor Skoric

Could you tell more about what you're trying to build / port? Node is
not necessarily the best tool for every project.

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Re: # Fatal error in ../deps/v8/src/objects-inl.h, line 2131 # CHECK(object->IsString()) failed

[Ben Noordhuis]

On Tue, Jun 12, 2012 at 10:13 AM, Debbio  wrote:
>
> Arch : ia32
>
> The stable version 0.6.19 compiles and seems to work fine on my
> system..

At what commit are you? We recently rolled back V8 to 3.9 because 3.10
and 3.11 had some serious stability issues.

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] # Fatal error in ../deps/v8/src/objects-inl.h, line 2131 # CHECK(object->IsString()) failed

[Ben Noordhuis]

On Tue, Jun 12, 2012 at 2:15 AM, Ryan Schmidt
 wrote:
>
> On Jun 11, 2012, at 18:02, Ben Noordhuis wrote:
>
>> What arch (x86, ppc) is this
>
> I was under the impression that v8 and nodejs did not support PowerPC 
> systems. Am I mistaken?

My bad, it doesn't.

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

[nodejs] Ideas for contribution to node.js in the context of a Master's Thesis

[Igor Skoric]

I'm developing a web application which aims to migrate at least two VB3 
16-bit windows applications and a collection of other tools to a modern 
platform for the University of Technology Graz.
Because I have a free pass to choose technologies and environments I aim to 
contribute to the technologies that I use and document that contribution 
and its implementation as part of my Master's Thesis in Software 
Development and Business Management.

I would appreciate your input - ideas that you would like to see become 
reality for node.js and also in general.
I will choose one more of the ideas that fit and aim to make the sources 
available.

Thank you for your help.
Igor Skoric

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

[nodejs] Re: # Fatal error in ../deps/v8/src/objects-inl.h, line 2131 # CHECK(object->IsString()) failed

[Debbio]

Arch : ia32

The stable version 0.6.19 compiles and seems to work fine on my
system..


On 12 Giu, 01:02, Ben Noordhuis  wrote:
> On Tue, Jun 12, 2012 at 12:40 AM, Debbio  wrote:
> > I'm trying to install nodejs :
>
> > Mac osx 10.5.8
> > nodejs v0.7.11-pre
> > gcc version 4.0.1 (Apple Inc. build 5493)
>
> > with commands :
>
> >  ./configure --debug --without-ssl --without-snapshot
> >  make
> >  sudo make install   == > I got bus error 
>
> > then :
>
> >  touch empty.js
> >  ./gdb --args ./node_g empty.js
>
> >  backtrace :
>
> > #0  0x94a70d52 in __kill ()
> > #1  0x94a70d44 in kill$UNIX2003 ()
> > #2  0x94ae3242 in raise ()
> > #3  0x94aef681 in abort ()
> > #4  0x0033f0cf in v8::internal::OS::Abort () at ../deps/v8/src/
> > platform-macos.cc:186
> > #5  0x00094dee in V8_Fatal (file=0x365ae0 "../deps/v8/src/objects-
> > inl.h", line=2131, format=0x365abc "CHECK(%s) failed") at ../deps/v8/
> > src/checks.cc:58
> > #6  0x0007092f in v8::internal::String::cast (object=0xdeadbeef) at
> > objects-inl.h:2131
> > #7  0x001bd6eb in v8::internal::Utf8SymbolKey::IsMatch
> > (this=0x94a70d52, string=0x0) at ../deps/v8/src/objects.cc:10632
> > #8  0x0018eaa1 in
> > v8::internal::HashTable > v8::internal::HashTableKey*>::FindEntry (this=0x2170809d, isolate=0x0,
> > key=0xbfffec08) at objects.h:2884
> > #9  0x0018eae1 in
> > v8::internal::HashTable > v8::internal::HashTableKey*>::FindEntry (this=0xdeadbeef,
> > key=0x94a70d52) at objects-inl.h:2070
> > #10 0x00198858 in v8::internal::SymbolTable::LookupKey
> > (this=0x2170809d, key=0xbfffec08, s=0xbfffec4c) at ../deps/v8/src/
> > objects.cc:11745
> > #11 0x001accfd in ~Utf8SymbolKey [inlined] () at ../deps/v8/src/
> > objects.cc:11718
> > #12 0x001accfd in v8::internal::SymbolTable::LookupSymbol
> > (this=0x2170809d, str={start_ = 0x385986 "undefined", length_ = 9},
> > s=0x0) at ../deps/v8/src/objects.cc:11718
> > #13 0x000ebc51 in v8::internal::Heap::LookupSymbol (this=0x840054,
> > string={start_ = 0x385986 "undefined", length_ = 9}) at ../deps/v8/src/
> > heap.cc:5159
> > #14 0x000fe49a in v8::internal::Heap::LookupAsciiSymbol (this=0x0,
> > str=0x0) at heap.h:997
> > #15 0x001a23d1 in v8::internal::Oddball::Initialize (this=0x4b108091,
> > to_string=0x94a70d52 "s\016?", to_number=0x422080a5, kind=5 '\005')
> > at ../deps/v8/src/objects.cc:7552
> > #16 0x000f5af1 in v8::internal::MaybeObject::ToObject () at objects.h:
> > 2453
> > #17 0x000f5af1 in v8::internal::Heap::CreateInitialObjects
> > (this=0x840054) at ../deps/v8/src/heap.cc:2444
> > #18 0x000f9422 in v8::internal::Heap::SetUp (this=0x840054,
> > create_heap_objects=true) at ../deps/v8/src/heap.cc:5964
> > #19 0x00160f4a in v8::internal::Isolate::Init (this=0x84, des=0x0)
> > at ../deps/v8/src/isolate.cc:1800
> > #20 0x00286ff2 in v8::internal::V8::Initialize (des=0x0) at ../deps/v8/
> > src/v8.cc:91
> > #21 0x00056a6d in InitializeHelper () at ../deps/v8/src/api.cc:295
> > #22 0x00057bcc in v8::V8::Initialize () at ../deps/v8/src/api.cc:4177
> > #23 0xad7a in node::Start (argc=2, argv=0xbfffeec8) at ../src/
> > node.cc:2870
> > #24 0x26ae in start ()
>
> > I didn't find nothing related to this error in google...
> > Thanks in advance.
> > Debbio
>
> It feels like a V8 bug.
>
> > v8::internal::HashTableKey*>::FindEntry (this=0xdeadbeef,
>
> The 0xdeadbeef value is a guard that the V8 garbage collector puts in
> place of reaped objects.
>
> What arch (x86, ppc) is this and does it also happen with the v8
> shell? (If you don't know what that is, let me know and I'll walk you
> through building it.)

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Re: [nodejs] Re: Microsoft Driver for Node.JS for SQL Server

[Graeme Foster]

Could you change "now" into "soon"? It's very early days and Windows only
for the moment!

G.
--
Graeme Foster


On 12 June 2012 00:36, Chad Retz  wrote:

> Fantastic. I guess I can retire https://github.com/cretz/node-tds now as
> I haven't had much time to work on it anyways.

-- 
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 nodejs@googlegroups.com
To unsubscribe from this group, send email to
nodejs+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en