On Jun 1, 2012, at 6:29 PM, Jorge wrote:
On May 30, 2012, at 12:57 AM, Tim Caswell wrote
On Tue, May 29, 2012 at 5:22 PM, Jorge jo...@jorgechamorro.com wrote:
FYI: setImmediate === nextTick and identical to a setTimeout(f,0) without
the clamping.
so I guess in a way setImmediate(fn)
Haha, nice. It's good to know that my outlook on things stays fairly
consistent. It's also interesting that everyone in that thread is pretty
much agreed that nextTick is a great idea and essential to node. But today
it's ZOMG nextTick is horribly broken!. I guess that's called progress :)
:Marco
IMHO, node needs as few non-IO browser non-compliant APIs as possible. If
something like setImmediate will suffice, use it instead of nextTick. If
something like thisTick is necessary, ideally it would be done in pure
JavaScript or in a browser-compliant way so any modules depending on it can
more
Sorry for the noise. That last point was mostly tangential, but felt it
should be mentioned.
On Fri, Jun 1, 2012 at 11:04 AM, Adam Crabtree atcrabt...@gmail.com wrote:
IMHO, node needs as few non-IO browser non-compliant APIs as possible. If
something like setImmediate will suffice, use it
Man Bruno, you almost had a really strong message. But then you ruined it :)
@isaac, @mikeal,
This is turning into a real joke!
Agreed
You tell us that you would like to change process.nextTick and you ask us
for feedback.
Several people respond that this change is problematic for
While I find it flattering to be referred to as a core guy and I rarely
correct people when they say it, you should know that I have less patches in to
core than I have fingers.
One of the patches was quite large, but there are many people who spend far
more time on core than I do.
What I
First of all, Mikeal Rogers is not a core guy who doesn't write
apps. Mikeal Rogers has written several apps with node, and has
authored very few patches to node core. He has a startup, and has
been building applications with Node since the 0.1 days.
And yes, I work on the node core project,
#3: Preventing event loop starvation
I sometimes find myself in a place where I think if this gets called
now, it'll go into a closed loop so I use process.nextTick to break
out of that, but still retain the speed and order of my calls.
setTimeout(fn, 0) is too scary for orderly calls, even if
Isaac,
Stop saying that the child process API is the answer to our problem: you
don't know what our problem is and you are making up a solution that would
not work for us.
process.nextTick is a non deteministic call and you want to turn it into a
deterministic one because its non-determinism
On Fri, Jun 1, 2012 at 12:19 PM, Isaac Schlueter i...@izs.me wrote:
First of all, Mikeal Rogers is not a core guy who doesn't write
apps. Mikeal Rogers has written several apps with node, and has
authored very few patches to node core. He has a startup, and has
been building applications
On Jun 1, 2012, at June 1, 20121:09 PM, Marco Rogers wrote:
On Fri, Jun 1, 2012 at 12:19 PM, Isaac Schlueter i...@izs.me wrote:
First of all, Mikeal Rogers is not a core guy who doesn't write
apps. Mikeal Rogers has written several apps with node, and has
authored very few patches to
On Fri, Jun 1, 2012 at 1:09 PM, Marco Rogers marco.rog...@gmail.com wrote:
Being a core guy means you have a tremendous amount of
influence on the way things go. If Mikeal did not agree with this change, we
would likely not be talking about it. Not saying you wouldn't want to fix
the core
Found the post with the clarification:
https://groups.google.com/d/msg/nodejs/NBSZRyBl11E/KbBS1r6VSBQJ
For the record, at the time:
1) It was very clear that events could interleave with next ticks.
2) I talked about using nextTick to break long computations into chunks in
the same thread and
So, you're saying that there are, today, in the real world,
applications where you must iterate over a work queue that is:
1. Impossible to serialize in a child process.
2. Arbitrarily large.
3. Small enough to fit in memory.
4. Well-partitioned enough to split up into multiple passes.
5.
On Fri, Jun 1, 2012 at 2:54 PM, Marco Rogers marco.rog...@gmail.com wrote:
nextTick works perfectly well for
large n. It doesn't work well when it is used to defer execution outside the
current stack, AND n is large, AND you still want to catch pending data
events.
Not working for defer
Issac,
Do you have a count of cases where the current behavior is causing
problems? Can you estimate the % of high-demand sites which have
problems?
You *cannot* discover the range and severity of problems a proposed
change may cause by describing it on a mailing list! The only way to
find out
On Fri, Jun 1, 2012 at 8:02 PM, Isaac Schlueter i...@izs.me wrote:
It's not just deciding to not take the advice. Show me a real world
use case where you need the current nextTick semantics. Like, a
production application that will be broken by this change. I'm aware
of several production
Thanks for fix. Now I can compare it with my lib.
2012/6/1 Henri Gourvest hgourv...@gmail.com
Le 31/05/12 20:20, Denys Khanzhiyev a écrit :
I have created issue on GitHub with simple test.
thank you,it is fixed
--
Job Board: http://jobs.nodejs.org/
Posting guidelines:
Can you provide example of how not to fetch entire result?
It is not always efficient to accumulate response - it is better to return
it clients as it comes in chunks.
2012/6/1 Denys Khanzhiyev xden...@gmail.com
Thanks for fix. Now I can compare it with my lib.
2012/6/1 Henri Gourvest
I'm not too much into the node specifics of streams, but I say.
If its an expected end, call close.
If its an unexpected end, call error then close,
if someone calls write after close, throw an error. Since this should
never happen in a correctly written application, if its okay it
terminates if
Still trying and cannot get it worked. Can I just download some
web-socket-sever suit from github and replace npm server folder in my
computer. That's the last resort I can Think of.
--
Job Board: http://jobs.nodejs.org/
Posting guidelines:
Please post the errors npm writes to the console.
On Fri, Jun 1, 2012 at 9:59 AM, Ket kettin...@gmail.com wrote:
Still trying and cannot get it worked. Can I just download some
web-socket-sever suit from github and replace npm server folder in my
computer. That's the last resort I can Think
On Jun 1, 3:59 am, Ket kettin...@gmail.com wrote:
Still trying and cannot get it worked. Can I just download some
web-socket-sever suit from github and replace npm server folder in my
computer. That's the last resort I can Think of.
`ws.js` package !== `ws` package in npm.
--
Job Board:
On Jun 1, 4:57 am, mscdex msc...@gmail.com wrote:
On Jun 1, 3:59 am, Ket kettin...@gmail.com wrote:
Still trying and cannot get it worked. Can I just download some
web-socket-sever suit from github and replace npm server folder in my
computer. That's the last resort I can Think of.
On Jun 1, 5:04 am, mscdex msc...@gmail.com wrote:
On Jun 1, 4:57 am, mscdex msc...@gmail.com wrote:
On Jun 1, 3:59 am, Ket kettin...@gmail.com wrote:
Still trying and cannot get it worked. Can I just download some
web-socket-sever suit from github and replace npm server folder in my
that sounds very reasonable
I pretty much never call write directly. I always let pipe do that for
me, and pipe cleans up after close, so that would work. and write will
never be called after 'close'.
I've been extremely enthusiastic about the Stream api for nearly a
year now, but have only just
On Wed, May 30, 2012 at 11:15 AM, Tim Caswell t...@creationix.com wrote:
On Wed, May 30, 2012 at 7:15 AM, Oliver Leics oliver.le...@gmail.com
wrote:
is dustjs[1] really the only template engine that implements rt
streamed render output and supports asynchronous calls within
Hello there!
I had just ported php.js (http://phpjs.org) for Node.js.
My motivation was to available new functions that could help programmers
to accomplish simple tasks and yup, I came from PHP and I was really
missing some basic functions :)
A list of functions is available at
I have a question regarding these requirement bundlers above mentioned. Is
there one of these that will allow me to include the separate modules
during development while using require(), and then just concat-minify all
of the modules like browserify does for release?
On Saturday, April 21,
I have to scrape thousands of different websites, as fast as possible.
On a single node process I was able to fetch 10 urls per second.
Though if I fork the task to 10 worker processes, I can reach 64 reqs/
sec.
Why is so?
Why I am limited to 10 reqs/sec on a single process and have to spawn
That's ok, I had already separated some sticks and gasoline to help
people to burn me down because of this package...
Why everyone get scared when they read PHP in the name of the package?
I should have name it node-bunch-of-functions :)
There's no PHP inside this package, just similarly named
I should have name it node-bunch-of-functions :)
You exactly right! And really, in this case you can skip some phpjs
fucntions from there,
especially aliases like explode/implode and, IMO, unefective bcmath port.
There is node-bigint for node.js that, i think, will better fit in node.js
On Fri, Jun 1, 2012 at 7:07 AM, Vitaly Puzrin vit...@rcdesign.ru wrote:
A there any known expectation about node 0.8 release? Or about 0.7.x
stability.
We need 3-bytes unicode chars support, and that's impossible in 0.6 branch.
Just more fresh v8 required.
It's ok in node v0.7, but i'm not
Thanks. A bit tricky. And how not to fetch blobs? Blobs are better to
stream.
2012/6/1 Henri Gourvest hgourv...@gmail.com
Le 01/06/12 08:57, Denys Khanzhiyev a écrit :
Can you provide example of how not to fetch entire result?
i have commited something, see test5 in test.js
--
Job
On Fri, Jun 1, 2012 at 2:00 PM, Mick mickodan...@gmail.com wrote:
I have to scrape thousands of different websites, as fast as possible.
On a single node process I was able to fetch 10 urls per second.
Though if I fork the task to 10 worker processes, I can reach 64 reqs/
sec.
Why is so?
On Fri, Jun 1, 2012 at 7:38 PM, catshow ga...@dynafocus.com wrote:
My scenario that I am looking at doing is this.
Mac OS/X app xpc_message_send 'start' --- xpc_module with node.js
embedded library initialized
Mac OX/X app xpc_message_send 'do work' -- xpc_module receive 'do
work' --
I don't think posting the whole code here would benefit the question.
The process is very simple: 1. start 10 workers with
child_process.fork; 2. read 400 random urls from db; 3. pass each url
to random worker; 4. calculate number of responses per second.
I've increased settings for
Le 01/06/12 16:23, Denys Khanzhiyev a écrit :
Thanks. A bit tricky.
how would you do it more simply ?
everything is async!
And how not to fetch blobs? Blobs are better to
stream.
if you do not give a transaction to the fetch method, blobs are not fetched.
--
Job Board:
Thanks for the quick feedback.
The XPC Module would be receiving messages in a Grand Central Dispatch
serial queue that I create, so it would NOT be receiving them in the
same thread as node/v8 is running.
So I guess that I would have to use the uv_async_send(). Is that well
documented? I
2012/6/1 Henri Gourvest hgourv...@gmail.com
Le 01/06/12 16:23, Denys Khanzhiyev a écrit :
Thanks. A bit tricky.
how would you do it more simply ?
everything is async!
It is not obvious what and where is going on. Here is my proposal, while it
is a bit tricky too :-)
Have not tested if it
path.existsSync('/my-file/.')
returns *true* even on a file.
is it expected? (not by me ^_^)
--
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
If the file exists it should return true. The path can be to anything.
On Fri, Jun 1, 2012 at 10:57 AM, rejetto reje...@gmail.com wrote:
path.existsSync('/my-file/.')
returns *true* even on a file.
is it expected? (not by me ^_^)
--
Job Board: http://jobs.nodejs.org/
Posting guidelines:
FYI, we're refactoring this in 0.9.
If I remember correctly it goes something like.
No more destroy, destorySoon and close. Only destroy.
I can't remember if we decided that:
- error will always call destroy
or
- a stream is guaranteed to eventually emit either error or destroy
maybe isaac
Hi all,
Thanks for those that tried the prototype, and share your comments.
I move it to the Amazon EC2 server:
http://ec2-72-44-39-177.compute-1.amazonaws.com/
Here is the guide which is very useful:
i'm sorry, i could have been clearer.
Consider /my-file to be an existing file.
But i added 2 characters to the argument string: a slash and a dot, for a
resulting /my-file/.
This would totally make sense if /my-file was actually a directory (with a
misleading name).
--
Job Board:
Ben, thanks for info. And thanks to all node.js team for your work.
Vitaly.
пятница, 1 июня 2012 г., 18:22:23 UTC+4 пользователь Ben Noordhuis написал:
On Fri, Jun 1, 2012 at 7:07 AM, Vitaly Puzrin vit...@rcdesign.ru wrote:
A there any known expectation about node 0.8 release? Or about
The path is normalized and get back to /my-file, check the source
пятница, 1 июня 2012 г., 21:13:33 UTC+3 пользователь rejetto написал:
i'm sorry, i could have been clearer.
Consider /my-file to be an existing file.
But i added 2 characters to the argument string: a slash and a dot, for a
On 6/1/12 3:16 AM, Oliver Leics wrote:
On Fri, Jun 1, 2012 at 7:50 AM, Alan Gutierreza...@prettyrobots.com wrote:
Never confused me. The (err, result) signature means one thing. This means
another. Not very confusing at all. Note: I'm people.
Good for you :) BTW: You are not people, you are
On 6/1/12 8:41 AM, Gabriel Farrell wrote:
On Wed, May 30, 2012 at 11:15 AM, Tim Caswellt...@creationix.com wrote:
On Wed, May 30, 2012 at 7:15 AM, Oliver Leicsoliver.le...@gmail.com
wrote:
is dustjs[1] really the only template engine that implements rt
streamed render output and supports
I am running into an issue with building the MSI right at the end:
Generating code
Finished generating code
node.vcxproj - C:\Users\mspisars\dev\node\Release\node.exe
SignTool Error: No certificates were found that met all the given criteria.
C:\Users\mspisars\dev\node\npm.wxs(380):
Due the great expectation around node.js, libs and functions grow up
without control.
Nowadays we have a lot of requires and we need to memorize a lot of
names (fs, path, vm, ...); with time, we'll spend more and more time
reading documentation :-(
My proposal is incorporate the mechanism
Interesting... but
A key property of a require is the isolation of its product, something
like namespaces in other languages/technologies.
So myrequire.abs is clearly different from yourrequire.abs
Paraphrasing David Hilbert (about Cantor)
No one shall expel us from the paradise that
clear enough, thank you!
i consider it an unwanted side-effect of the normalization.
--
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
You want `existsSync` to tell you if a path is a directory or a file. But
`existsSync` doesn't do that. You would need to make that a feature
request.
On Fri, Jun 1, 2012 at 11:51 AM, Massimo Melina reje...@gmail.com wrote:
clear enough, thank you!
i consider it an unwanted side-effect of
tl;dr
On Fri, Jun 1, 2012 at 8:41 PM, Alan Gutierrez a...@prettyrobots.com wrote:
On 6/1/12 3:16 AM, Oliver Leics wrote:
On Fri, Jun 1, 2012 at 7:50 AM, Alan Gutierreza...@prettyrobots.com
wrote:
Never confused me. The (err, result) signature means one thing. This
means
another. Not very
yes Mark, that's what i was trying to do.
I know the common way of doing this is by fs.stat()
Nonetheless the correct value for exists(my-file/.) is false, because
that doesn't exist.
The dot is a real entry in the file system, but exists only within
directories.
If you ask the operating system if
Since i am testing i don't need to have it signed.
Passing in nosign to .\vcbuild.bat release nosign msi
gets rid of the issue.
Michael
On Friday, June 1, 2012 1:01:23 PM UTC-4, Michael Pisarski wrote:
I am running into an issue with building the MSI right at the end:
Generating code
+1 to the idea.
Btw, don't mention PHP or you will get in troubles here...
--
Att,
Alan Hoffmeister
2012/6/1 Angel Java Lopez ajlopez2...@gmail.com:
Interesting... but
A key property of a require is the isolation of its product, something
like namespaces in other
I must admit your lib is faster than mine
here is code
https://gist.github.com/2854642
here is ab results
node-firebird-libfbclient:
Concurrency Level: 5
Time taken for tests: 3.346934 seconds
Complete requests: 1000
Failed requests:0
Write errors: 0
Total
-1 as I like my modules exactly as they are.
--Josh
On Fri, Jun 1, 2012 at 12:32 PM, Angel Java Lopez ajlopez2...@gmail.com wrote:
neither to add some sync function at core ;-)
On Fri, Jun 1, 2012 at 4:31 PM, Alan Hoffmeister alanhoffmeis...@gmail.com
wrote:
+1 to the idea.
Btw, don't
-1 as well. Thank you Josh.
On Fri, Jun 1, 2012 at 12:39 PM, Joshua Holbrook josh.holbr...@gmail.comwrote:
-1 as I like my modules exactly as they are.
--Josh
On Fri, Jun 1, 2012 at 12:32 PM, Angel Java Lopez ajlopez2...@gmail.com
wrote:
neither to add some sync function at core ;-)
On Fri, Jun 01, 2012 at 12:42:42PM -0700, Martin Wawrusch wrote:
On Fri, Jun 1, 2012 at 2:26 PM, Seiji Sam Lee seijisam...@gmail.com
wrote:
Due the great expectation around node.js, libs and functions grow up
without control.
Nowadays we have a lot of requires and we
No, we're not going to do this. The module system is finished. It
won't be changing except for critical bugs.
On Fri, Jun 1, 2012 at 12:57 PM, Alan Gutierrez a...@prettyrobots.com wrote:
On Fri, Jun 01, 2012 at 12:42:42PM -0700, Martin Wawrusch wrote:
On Fri, Jun 1, 2012 at 2:26 PM, Seiji
On Fri, Jun 1, 2012 at 8:44 PM, Alan Gutierrez a...@prettyrobots.com wrote:
A big issue is that if there is an error in your template, there's no way to
report it if you've already started streaming 200 OK.
If there is an parse error in the template, you can not start
streaming and you send a
Oliver,
Node has two very different API styles:
1) The callback style where the callback is called exactly once and
receives (err, result) arguments
2) The event style where the callback may be called multiple times with
event data and where errors are dispatched as a separate error event. The
Uh, doesn't it already work this way?
mymodule.js:
Math.abs = function() ...
require('util').isset = function() ...
That being said, don't do this, modules/namespaces are there for very
good reason.
On Jun 1, 12:59 pm, Isaac Schlueter i...@izs.me wrote:
No, we're not going to do this.
Bruno,
On Fri, Jun 1, 2012 at 11:35 PM, Bruno Jouhier bjouh...@gmail.com wrote:
Node has two very different API styles:
[...]
I know that. I knew before I wrote into this thread.
Looks like you want to align 2) on 1) by adding an error parameter to event
handlers. Sounds wrong as errors
Le 01/06/12 21:39, Denys Khanzhiyev a écrit :
I must admit your lib is faster than mine
Well it is not a competition.
I think the problem is firebird client library still connect
synchronously to firebird server. If you run your tests again with a
remote firebird server, you might have more
For example, run as service?
--
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
On 6/1/12 4:36 PM, Oliver Leics wrote:
On Fri, Jun 1, 2012 at 8:44 PM, Alan Gutierreza...@prettyrobots.com wrote:
A big issue is that if there is an error in your template, there's no way to
report it if you've already started streaming 200 OK.
If there is no parse error and the error
On Fri, Jun 1, 2012 at 9:26 PM, catshow ga...@dynafocus.com wrote:
Thanks for the quick feedback.
The XPC Module would be receiving messages in a Grand Central Dispatch
serial queue that I create, so it would NOT be receiving them in the
same thread as node/v8 is running.
So I guess that I
I think the Todo app is today's version of the guest book tutorials. It
shows off post requests, DB connectivity and web sockets.
I did notice that you have a Authentication and Authorization script, which
would be nice to see too.
On Thursday, May 31, 2012 12:55:03 PM UTC-5, Shogun wrote:
The main benefit in kernel isn't streaming to save memory. In fact as I
mentioned in my first comment here, kernel doesn't support
streaming currently because I didn't find it useful enough. Where kernel
shines if the fact that you no longer have to gather your data before
passing instructions
On Friday, June 1, 2012 2:32:50 AM UTC-4, Diego Varese wrote:
I have a question regarding these requirement bundlers above mentioned. Is
there one of these that will allow me to include the separate modules
during development while using require(), and then just concat-minify all
of the
74 matches
Mail list logo
