Hi Pavel,

I just tried on a machine of mine.

I changed testNetwork.st to

Transcript show: (ZnClient new logToTranscript; get: 
'http://zn.stfx.eu/zn/numbers.txt'); cr.

That way you get a bit more progress info. As I suspected the request is 
written and the failure is while reading the response. I tried connecting to a 
local Zinc server which confirmed this as well: the request comes in on the 
server and it writes the response.

This is the output now:

root@stfx:~# ./pharo PharoKernel.image testNetwork.st 

2014-02-17 10:00:59 463573 I Wrote a ZnRequest(GET /random)
2014-02-17 10:00:59 463573 D Sent headers

User-Agent: Zinc HTTP Components 1.0

Accept: */*

Host: localhost:1701


===============================================================================
Notice: Errors in script loaded from /root/testNetwork.st
===============================================================================
Errors in script loaded from /root/testNetwork.st
==== Startup Error: ConnectionClosed: Connection closed while waiting for data.
Socket>>waitForDataFor: in Block: [ ConnectionClosed signal: 'Connection closed 
whil...etc...
Socket>>waitForDataFor:ifClosed:ifTimedOut:
Socket>>waitForDataFor:
Socket>>receiveDataSignallingTimeout:into:startingAt:
SocketStream>>receiveData
SocketStream>>next
ZnLineReader>>processNext
ZnLineReader>>nextLine
ZnStatusLine>>readFrom:
ZnStatusLine class>>readFrom:
ZnResponse>>readHeaderFrom:
ZnResponse(ZnMessage)>>readFrom:
ZnResponse class(ZnMessage class)>>readFrom:
ZnClient>>executeRequestResponse
ZnClient>>getConnectionAndExecute in Block: [ self executeRequestResponse ]
BlockClosure>>ensure:
ZnClient>>getConnectionAndExecute
ZnClient>>executeWithRedirectsRemaining:
ZnClient>>executeWithRetriesRemaining: in Block: [ self 
executeWithRedirectsRemaining: self maxNumb...etc...
BlockClosure>>on:do:
ZnClient>>executeWithRetriesRemaining:
ZnClient>>executeWithTimeout in Block: [ self executeWithRetriesRemaining: self 
numberOfR...etc...
BlockClosure>>on:do:
ZnClient>>executeWithTimeout in Block: [ ...
ZnClient>>withTimeoutDo: in Block: [ ^ block value ]
ZnConnectionTimeout(DynamicVariable)>>value:during: in Block:

You are using regular socket streams, which is good. For now I can't think of a 
reasons why the reading would fail, if I think of something, I will let you 
know.

It is pretty hard to debug this way ;-)

Sven

On 17 Feb 2014, at 10:22, Pavel Krivanek <pavel.kriva...@gmail.com> wrote:

> (NetNameResolver addressForName: 'www.esug.org') returns good value...
> 
> -- Pavel
> 
> 
> 2014-02-17 9:32 GMT+01:00 Pavel Krivanek <pavel.kriva...@gmail.com>:
> Hi,
> 
> thank you for the tip. It seems that it is not directly Zinc related. But if 
> you want something to play with, do: 
> 
> - load the bootstrapped image:
> https://ci.inria.fr/pharo-contribution/view/Pharo-Kernel-3.0/job/PharoKernel3.0-Bootstrap/lastSuccessfulBuild/artifact/PharoKernel/results/PharoKernel.zip
> 
> ./pharo PharoKernel.image fix.st
> ./pharo PharoKernel.image loadNetwork.st 
> ./pharo PharoKernel.image testNetwork.st 
> 
> The CI slaves revolt now so I cannot make it more comfortable :-)
> 
> Cheers,
> -- Pavel
> 
> 
> 
> 
> 2014-02-16 22:58 GMT+01:00 Sven Van Caekenberghe <s...@stfx.eu>:
> 
> Hi Pavel,
> 
> On 16 Feb 2014, at 22:23, Pavel Krivanek <pavel.kriva...@gmail.com> wrote:
> 
> > Hi Sven,
> >
> > we can successfully load network packages to the bootstrapped image but 
> > when we try to test it using:
> >
> > ZnClient new get: 'http://zn.stfx.eu/zn/numbers.txt'.
> >
> > we get "ConnectionClosed: Connection closed while waiting for data" error 
> > after few seconds. Do you have some idea, what may be wrong? Probably 
> > something is not initialized well but I have no clue what.
> >
> > Cheers,
> > -- Pavel
> 
> Hard to say. I would guess that you are already on the point of waiting for a 
> server response, so you already successfully initialised networking, resolved 
> DNS, opened a TCP socket and wrote the request over it. On the other hand, 
> one of those steps might not have been done right, since you get no response.
> 
> Any chance of a stack trace ? Or an image ? But that would be a headless one 
> I guess ;-)
> 
> Sven
> 
> 
> 
> 
> 


Reply via email to