Re: [nodejs] require('non-loadable module') does not return detailed error information

In the meantime, you can also use gdb to break in that lib loading error clause and manually call dlerror. Am 13.02.2012 22:29 schrieb "Brian Fundakowski Feldman" < brianfundakowskifeld...@gmail.com>: I did this, neglecting to actually try to format the error message properly, but it was good eno

Re: [nodejs] Fork and patch npm modules

Yes, there was an issue with that - I think someone recently patched it, not sure. Am 14.02.2012 02:50 schrieb "Phoscur" : Well it seems it's still somewhat unstable as the module I'm using does not work as expected (and not as without symlink). Could that be a problem in path.dirname or path. n

[nodejs] Re: Async Error Handling and Domains

Interesting - benching it was my plan asap (as well as writing unit tests). I was kind of hoping that v8 would be making a simple copy of the stack trace it records anyway, and cheaply store that into the Error object somehow. In any case, any form of pre-processing is way too intrusive to me, and

[nodejs] Re: Node CPU use analysis

A meaningful thing to do might be to run Node in the v8 profiler mode, and then analyze the log file with the v8 tick-analyzer script. E.g. cd /yourdir node --prof --prof_auto yourscript.js cd /v8-source/out/native ../../tools/-tick-processor /yourdir/v8.log I could not get it to work without

Re: [nodejs] node_modules and DRYS

On Mon, Feb 13, 2012 at 9:00 PM, C. Mundi wrote: > > Hi.  I have another naive question.  Nathan has been helping me out == a lot > == with node-gyp and node-bindings and I want to say publicly that these > tools are going to be ++huge for those of us forced work in > platform-heterogeneous networ

[nodejs] Re: Node CPU use analysis

I'd be really curious about that patch and the benchmarks as well, and I'm sure it will make it to Node's core. The furthest I had time to get so far is: https://gist.github.com/1350901 I don't have the benchmarks but it was a lot faster than serving the file in a classic way, though less than 50%

[nodejs] Re: Resumable MD5 Hash Algorithm

but how to import the hash-till-pause back to the Hash Coder? I have not found an API in Node for that. On 2月13日, 下午8时02分, Damianos Mylonakis wrote: > Save the hash-till-pause in some sort of database that every process > talks to. > > On 02/13/2012 05:48 AM, Rong Shen wrote: > > > > > > > > >

[nodejs] node_modules and DRYS

Hi. I have another naive question. Nathan has been helping me out == a lot == with node-gyp and node-bindings and I want to say publicly that these tools are going to be ++huge for those of us forced work in platform-heterogeneous networks. One thing I notice is that node-bindings seems designed

Re: [nodejs] Node CPU use analysis

On Mon, Feb 13, 2012 at 8:27 PM, Tim Caswell wrote: > I've benchmarked it before, but it should be obvious too. Node is > doing less work than nginx in this case. (writing an already buffered > response to the socket is much easier than reading a file from disk > and then sending it to the sock

[nodejs] Re: User Groups in South East Asia -- Singapore, Malaysia, Indonesia, Philippines

There's a Facebook group in Taiwan. http://www.facebook.com/NodeJS.tw Not sure whether they meet up. On Feb 14, 12:39 pm, Ruben Tan wrote: > In Malaysia there is nodehack organized by my company 40 square aka > flightoffice pty ltd, and a general user group called node.js.my > > Ruben Tan > >

Re: [nodejs] User Groups in South East Asia -- Singapore, Malaysia, Indonesia, Philippines

In Malaysia there is nodehack organized by my company 40 square aka flightoffice pty ltd, and a general user group called node.js.my Ruben Tan On 14 Feb 2012, at 12:23, Kenneth Shaw wrote: > I was wondering if there were any user groups or meetings for Node.js > going on in the SEA area. Would

[nodejs] User Groups in South East Asia -- Singapore, Malaysia, Indonesia, Philippines

I was wondering if there were any user groups or meetings for Node.js going on in the SEA area. Would be interested in attending a meeting if there was (and I was in the country at the time). Conversely, I'd be interested in starting a group if anyone else was up for it. -Ken -- Job Board: http

Re: [nodejs] Fork and patch npm modules

Well it seems it's still somewhat unstable as the module I'm using does not work as expected (and not as without symlink). Could that be a problem in path.dirname or path. normalize? Am 14.02.2012 00:08, schrieb Phoscur: > Just updated my nodejs to be sure, from 6.8 to 6.10, now it works.. > Could

Re: [nodejs] Node CPU use analysis

On Mon, Feb 13, 2012 at 5:18 PM, Matt wrote: > On Mon, Feb 13, 2012 at 4:28 PM, Tim Caswell wrote: >> >> But node wins over apache and nginx by having the right >> event model. > > > Maybe over Apache, but you need to read more about nginx. Sorry, that period was a typo. I know that nginx is ev

Re: [nodejs] trademarked logos in tarball

Ok. I'll just remove them from the dist tarball builds. (Makes the download smaller anyhow, which is a nice plus.) They'll still be in the git repo so we can easily track their changes. v0.7.4 will be the first build with this change. On Mon, Feb 13, 2012 at 11:01, Jérémy Lal wrote: > In the

[nodejs] Re: Node CPU use analysis

from tests i have done using the low level node.js bindings and sendfile on linux (which requires a small patch to the node.js c++ source), node.js can get pretty close to nginx (within 10-20%) for serving static files. the larger the files, the smaller the difference as most of the work is then be

Re: [nodejs] Node CPU use analysis

On Mon, Feb 13, 2012 at 4:28 PM, Tim Caswell wrote: > But node wins over apache and nginx by having the right > event model. > Maybe over Apache, but you need to read more about nginx. > Node serving a static file from > memory is faster than nginx serving the same file from disk. > Benchmar

Re: [nodejs] Fork and patch npm modules

Just updated my nodejs to be sure, from 6.8 to 6.10, now it works.. Couldn't find an entry of this in the changelog.. Am 14.02.2012 00:00, schrieb Kevin Swiber: > On Mon, Feb 13, 2012 at 5:49 PM, Phoscur > wrote: > > The object (even in the REPL) I get is empty({

Re: [nodejs] Fork and patch npm modules

On Mon, Feb 13, 2012 at 5:49 PM, Phoscur wrote: > The object (even in the REPL) I get is empty({}), it should be a function. > It seems node does not follow windows symlinks. > Hmm. Which version are you running? It works for me using v0.6.3. See below. Note: The required module in this exam

Re: [nodejs] Fork and patch npm modules

The object (even in the REPL) I get is empty({}), it should be a function. It seems node does not follow windows symlinks. Am 13.02.2012 23:21, schrieb Kevin Swiber: > > On Mon, Feb 13, 2012 at 5:05 PM, Phoscur > wrote: > > > "npm link" > >> How do I accompli

Re: [nodejs] Everyauth and access token

If you do not want your Facebook access_token to expire, you need to include offline_access in the scope when requesting it. -- Dick On Feb 13, 2012, at 1:12 PM, Tiziano Rossi wrote: > Hi to all! > > I use everyauth in my app for permit facebook authentication... > After that i want to use tha

Re: [nodejs] Fork and patch npm modules

On Mon, Feb 13, 2012 at 5:05 PM, Phoscur wrote: > > "npm link" > >> How do I accomplish this on windows? > > You're screwed, I guess. > Oh man, but I want this!! > > Windows 7 is able to make symbolic links too "mklink", but it seems all > I get back from a symlinked folder is an empty object..?

[nodejs] Re: Node CPU use analysis

On Feb 13, 5:55 pm, Cosmere Infahm wrote: > I analysed node-0.6.10/windows2008r2 running my clustered http server > script in VTune with a view to understanding why node took 50% more CPU > than nginx/apache for a given scenario. The results indicate that > GetQueuedCompletionStatusEx takes 40% of

Re: [nodejs] Fork and patch npm modules

> "npm link" >> How do I accomplish this on windows? > You're screwed, I guess. Oh man, but I want this!! Windows 7 is able to make symbolic links too "mklink", but it seems all I get back from a symlinked folder is an empty object..? Am 13.02.2012 21:11, schrieb Jann Horn: > 2012/2/13 Phoscur :

Re: [nodejs] Fork and patch npm modules

On Mon, Feb 13, 2012 at 2:57 PM, Phoscur wrote: > Hi, > After reading this: > > http://www.debuggable.com/posts/how-to-fork-patch-npm-modules:4e2eb9f3-e584-44be-b1a9-3db7cbdd56cb > I see there is a nice way to replace broken modules with your own > patched ones. But this always involves git push

[nodejs] Re: require('non-loadable module') does not return detailed error information

On Feb 13, 4:02 pm, Brian Fundakowski Feldman wrote: > I did this, neglecting to actually try to format the error message > properly, but it was good enough to figure out where I went wrong with > linking my module.  It would be nice if something like this were > integrated to assist module writer

Re: [nodejs] require('non-loadable module') does not return detailed error information

On Mon, Feb 13, 2012 at 22:02, Brian Fundakowski Feldman wrote: > I did this, neglecting to actually try to format the error message > properly, but it was good enough to figure out where I went wrong with > linking my module.  It would be nice if something like this were > integrated to assist mo

Re: [nodejs] Node CPU use analysis

It may be written in JavaScript and thus has the overhead of a language VM. This especially hurts in terms of memory overhead per process. But node wins over apache and nginx by having the right event model. Good low-level primitives in libuv and the ability to script-in assumptions. For exampl

[nodejs] require('non-loadable module') does not return detailed error information

I did this, neglecting to actually try to format the error message properly, but it was good enough to figure out where I went wrong with linking my module. It would be nice if something like this were integrated to assist module writers. --- src/node.cc.orig2012-02-02 19:56:42.0 -050

Re: [nodejs] Node CPU use analysis

It's written in Javascript. This isn't rocket science :) On Mon, Feb 13, 2012 at 3:50 PM, Cosmere Infahm wrote: > Looking at the Node I did not see a particular reason for the higher CPU -- Job Board: http://jobs.nodejs.org/ Posting guidelines: https://github.com/joyent/node/wiki/Mailing-List-

[nodejs] Re: Async Error Handling and Domains

On Feb 13, 1:56 pm, Jann Horn wrote: > Doesn't that mean that plain async methods might be missing in the > generated trace? Yes, unfortunately. It won't happen with low level async calls like fs.readFile, db calls, etc. but it may happen if you have intermediate async functions to which you p

Re: [nodejs] Node CPU use analysis

> > Why are people expecting Node to compare well with nginx or Apache for > serving static files? Node is faster than PHP or Ruby, but comparing it to > something specifically designed for high performance static file serving is > always going to be a losing proposition. > > Use Node for your dyna

[nodejs] Everyauth and access token

Hi to all! I use everyauth in my app for permit facebook authentication... After that i want to use tha accessToken to performe some graph api call... (maybe client side for performance reason). What is the correct way to do this? I got the accessToken returned from findOrCreateUser, but this can

Re: [nodejs] Fork and patch npm modules

2012/2/13 Phoscur : > Hi, > After reading this: > http://www.debuggable.com/posts/how-to-fork-patch-npm-modules:4e2eb9f3-e584-44be-b1a9-3db7cbdd56cb > I see there is a nice way to replace broken modules with your own > patched ones. But this always involves git push and npm update commands > when y

Re: [nodejs] Fork and patch npm modules

There is the npm link command that does just this. I don't know if it works on windows though since it uses symlinks. If it doesn't work, then yes nesting works just fine. Git is fairly good at ignoring nested repos automatically. On Mon, Feb 13, 2012 at 1:57 PM, Phoscur wrote: > Hi, > After r

[nodejs] Fork and patch npm modules

Hi, After reading this: http://www.debuggable.com/posts/how-to-fork-patch-npm-modules:4e2eb9f3-e584-44be-b1a9-3db7cbdd56cb I see there is a nice way to replace broken modules with your own patched ones. But this always involves git push and npm update commands when you change something. Isn't there

[nodejs] Re: non blocking JSON.stringify

here is my version i use: a fun little json database: https://gist.github.com/1819395 On Feb 12, 7:49 pm, "Kilian C." wrote: > I'm noticing that most of you (also on other similar threads) are focusing > on parsing objects. > I'm focusing instead on stringify the output of a big json. > > My goa

Re: [nodejs] non blocking JSON.stringify

2012/2/13 Nuno Job : > Some discussion has done into how such API would look like: > > https://github.com/thejh/node-jsos/issues/2 > > If you go on and implement anything like this I would be super excited. I'm > also pretty sure you would have some suggestion to how the API should look > like. > >

Re: [nodejs] trademarked logos in tarball

In the dist tarballs. A short sentence about those given permissions in e.g. the LICENSE file might be enough. If paths are mentionned, beware that doc/api/api/assets/ contains the green twitter logo too. Removing the website but letting doc/api in the dist tarball would probably be a good thing

Re: [nodejs] trademarked logos in tarball

I believe that all the companies whose logos were used gave us explicit permission to do so. But yes, they're kind of big, and it's not all that necessary to be able to render the website from the tarball. Removing them from the git repo would be a bit more annoying. Is it causing you legal hass

Re: [nodejs] Re: how to write for-each loop in JavaScript properly

I still dislike forEach() same as some(); since the semantics are just more complicated. break doesn work (ok return true), but named break is gone, continue is now called return, and named continue is gone, and you cannot return from the parent function, which often I do to shortcut. You lose so m

Re: [nodejs] non blocking JSON.stringify

Not to forget you also did jsonparse, which for convenience reasons might be simpler to use until such high level abstraction exists on top of clarinet. :) Some discussion has done into how such API would look like: https://github.com/thejh/node-jsos/issues/2 If you go on and implement anything

[nodejs] Re: how to write for-each loop in JavaScript properly

My post mainly about loops, not about prototypes. To not use prototypes when you write a plugin script to be integrated into 3rd parity website is a very good idea. Generaly Node.js is very controlled-settings environment. I like to add some agile convenience by using prototypes, I add them to to

Re: [nodejs] Re: how to write for-each loop in JavaScript properly

On Mon, Feb 13, 2012 at 11:21 AM, Axel Kittenberger wrote: > some() is only for Arrays tough. Same with forEach. I had forgotten about Array.prototype.some(). If you want the short-circuit of some, use my code from above, but call it "some" to avoid confusion. > On Mon, Feb 13, 2012 at 5:47 PM

Re: [nodejs] non blocking JSON.stringify

Yajl does both parsing and encoding. Clarinet is basically a pure js tool that does much the same thing yajl can do. It's not hard to create JSON.parse and JSON.stringify APIs on top of yajl (though, at that point, you're buffering the final object or need the full json string buffered to parse i

[nodejs] Re: Watcher/debugger as to why a process is exiting? - or how to understand the event loop?

This has got to be a memory leak of some kind. Anyone know how to track it down? All I need is some debugging tool that can show me what is holding onto the references... On Feb 13, 4:55 pm, deitch wrote: > K, better answer out there? > > On Feb 13, 4:50 pm, Dan Milon wrote: > > > > > > > > >ht

Re: [nodejs] Node CPU use analysis

Why are people expecting Node to compare well with nginx or Apache for serving static files? Node is faster than PHP or Ruby, but comparing it to something specifically designed for high performance static file serving is always going to be a losing proposition. Use Node for your dynamic content.

Re: [nodejs] Re: how to write for-each loop in JavaScript properly

some() is only for Arrays tough. On Mon, Feb 13, 2012 at 5:47 PM, substack wrote: > On Feb 13, 7:01 am, Axel Kittenberger wrote: >> Why they didn't make the standard forEach in that returning false >> terminates the loop is beyond me... > > You can just abuse the short-circuiting of [].some: > >

Re: [nodejs] Re: node v0.6.6 crashes on http.get calls. Error: connect Unknown system errno 10060

On Mon, Feb 13, 2012 at 09:34, Jeroen Janssen wrote: > Hi, > > I don't see a patch/fix for this yet, is it as simple as adding: > > case WSAETIMEDOUT : return UV_ETIMEDOUT; > > to uv_translate_sys_error in libuv src/win/error.c? Yes. Which I've done in libuv@c1cea70. :-) -- Job Board: http://jo

[nodejs] Node CPU use analysis

I analysed node-0.6.10/windows2008r2 running my clustered http server script in VTune with a view to understanding why node took 50% more CPU than nginx/apache for a given scenario. The results indicate that GetQueuedCompletionStatusEx takes 40% of CPU. Does this make sense? I was wondering if the

Re: [nodejs] how to write for-each loop in JavaScript properly

Since this is the nodejs list, I'm going to assume you're working in node. In that case, Object.keys is built-in and it's really fast in V8 (much faster than for .. in). Also if you want to share code with the browser and it's not a mobile app (meaning it needs to support old ES3 browsers), then

[nodejs] Re: how to write for-each loop in JavaScript properly

On Feb 13, 7:01 am, Axel Kittenberger wrote: > Why they didn't make the standard forEach in that returning false > terminates the loop is beyond me... You can just abuse the short-circuiting of [].some: > [ 1, 2, 3, 4, 5, 6, 7 ].some(function (x) { console.log(x); if (x > 4) return > true }) 1

[nodejs] node.js meetup in Cape Town, South Africa

We're starting a node.js-centered meetup in Cape Town. There will be an initial meeting for drinks on Thursday after work just to gauge interest, then regular meetings with informal presentations starting soon after that. The details are on the website: http://nodecpt.github.com/ Or follow us on

Re: [nodejs] how to write for-each loop in JavaScript properly

> You could do :  myobj.hasOwnProperty(key) But then if myobj is NULL it will > throw an error. Unless nil suddendly has keys, it wouldn't go into loop anyway. Personally I don't like forEach too much, since I cannot break from it, when I decide the rest is uninteresting. Or even return from the

[nodejs] Re: Watcher/debugger as to why a process is exiting? - or how to understand the event loop?

K, better answer out there? On Feb 13, 4:50 pm, Dan Milon wrote: > https://github.com/dannycoates/v8-profiler > "authored a year ago" > > The project is abandoned and has several issues with v0.4++ > > On 02/13/2012 04:35 PM, deitch wrote: > > > > > > > > > Hmm, require('v8-profiler') after npm i

Re: [nodejs] Re: Watcher/debugger as to why a process is exiting? - or how to understand the event loop?

https://github.com/dannycoates/v8-profiler "authored a year ago" The project is abandoned and has several issues with v0.4++ On 02/13/2012 04:35 PM, deitch wrote: Hmm, require('v8-profiler') after npm install, then launch, run node- inspector, then enable profiling in Safari Profiles tab, click

[nodejs] Re: how to write for-each loop in JavaScript properly

Well, in < ES4 that is. On Feb 13, 9:43 am, Thom Blake wrote: > > Or if you do make them DontEnum. > > DontEnum is not a property that can be set by the user. > > On Feb 12, 3:09 pm, Marcel Laverdet wrote: > > > > > > > > > This is one of those old Crockfordisms that I definitely don't agree wit

[nodejs] Re: how to write for-each loop in JavaScript properly

> Or if you do make them DontEnum. DontEnum is not a property that can be set by the user. On Feb 12, 3:09 pm, Marcel Laverdet wrote: > This is one of those old Crockfordisms that I definitely don't agree with. > How about instead you just don't add garbage to Object.prototype? Or if you > do ma

[nodejs] Re: Watcher/debugger as to why a process is exiting? - or how to understand the event loop?

Hmm, require('v8-profiler') after npm install, then launch, run node- inspector, then enable profiling in Safari Profiles tab, click the "eye" icon to take a heap snapshot and the program crashes? On Feb 13, 4:08 pm, deitch wrote: > node-inspector I know; v8-profiler, on the other hand, I hav

[nodejs] Re: Watcher/debugger as to why a process is exiting? - or how to understand the event loop?

node-inspector I know; v8-profiler, on the other hand, I have never used for this. Good guide anywhere? On Feb 13, 3:16 pm, Ben Noordhuis wrote: > On Mon, Feb 13, 2012 at 13:29, deitch wrote: > > Running more, also posted on the zombie forum. > > > Looks like my heap is growing by ~4MB with each

Re: [nodejs] Re: Watcher/debugger as to why a process is exiting? - or how to understand the event loop?

On Mon, Feb 13, 2012 at 13:29, deitch wrote: > Running more, also posted on the zombie forum. > > Looks like my heap is growing by ~4MB with each test... *after* GC. It > exits without warning when the heap is around 160-170MB, which seems > low for collapsing, but very high for this program. Each

Re: [nodejs] Make errors: Linux ia32, for node-v0.6.6 and node-v0.4.8

On Mon, Feb 13, 2012 at 04:24, steve johnston wrote: > Hello Ben, I applied the patch. It does create the 'out' dir. But that's > about it: > >>make > tools/gyp_node -f make > output_dir: /usr/local/src/joyent-node-d4ccdea/out > > > make -C out BUILDTYPE=Release > make[1]: Entering directory `/usr

Re: [nodejs] Re: Async Error Handling and Domains

Doesn't that mean that plain async methods might be missing in the generated trace? Am 13.02.2012 09:25 schrieb "Bruno Jouhier" : I had to handle this in streamline and I also did it similarly with a small callback wrapper that does the try/catch. On the other hand, I found that allocating an Er

Re: [nodejs] Re: Resumable MD5 Hash Algorithm

Save the hash-till-pause in some sort of database that every process talks to. On 02/13/2012 05:48 AM, Rong Shen wrote: You can not expect the next part of file will go to the same Node process before when the file uploading working in resumable way. That is the reason why h.update(data) does

[nodejs] Re: Watcher/debugger as to why a process is exiting? - or how to understand the event loop?

Running more, also posted on the zombie forum. Looks like my heap is growing by ~4MB with each test... *after* GC. It exits without warning when the heap is around 160-170MB, which seems low for collapsing, but very high for this program. Each test should cause the heap to increase by the resource

Re: [nodejs] Printing straight to default printer without showing print dialogue in CakePHP application

I thank you for all your help. I'm considering using java fx shell as a browser to accomplish this. Also yes this is a software that has user/password login creds. I'm building an application for my company with branches all over the place. On 13 February 2012 04:29, Skyler Brungardt wrote: > Ag

[nodejs] Re: node v0.6.6 crashes on http.get calls. Error: connect Unknown system errno 10060

Hi, I don't see a patch/fix for this yet, is it as simple as adding: case WSAETIMEDOUT : return UV_ETIMEDOUT; to uv_translate_sys_error in libuv src/win/error.c? Best regards, Jeroen Janssen On 6 feb, 00:29, Bert Belder wrote: > On Feb 5, 11:45 pm, Ben Noordhuis wrote: > > > On Sun, Feb 5,

[nodejs] Re: Async Error Handling and Domains

I had to handle this in streamline and I also did it similarly with a small callback wrapper that does the try/catch. On the other hand, I found that allocating an Error object to keep track of the source stack is very costly (a new Error call costs 400 empty object {} allocations) so I chose to g