Yep that seems more likely than webkit1 vs webkit2.
On 14 February 2014 14:05, Gonzalo Odiard wrote:
> Can be a difference between the webkit version in F18 and F19 too.
>
> Gonzalo
>
>
> On Fri, Feb 14, 2014 at 10:01 AM, Daniel Narvaez wrote:
>
>> Did we even prove this is webkit1 only? webkit
Can be a difference between the webkit version in F18 and F19 too.
Gonzalo
On Fri, Feb 14, 2014 at 10:01 AM, Daniel Narvaez wrote:
> Did we even prove this is webkit1 only? webkit2 is basically just an API
> layer on top of webkit1 so it seems to be unlikely to cause something like
> this, I wa
Did we even prove this is webkit1 only? webkit2 is basically just an API
layer on top of webkit1 so it seems to be unlikely to cause something like
this, I was more suspecting our own webkit1 layer, that's pretty different
from webkit2 because it uses http etc.
On 5 February 2014 02:11, Gonzalo O
Maybe webkit is smarter in version 2, and don't try to connect
to external resources in the svg...
Gonzalo
On Tue, Feb 4, 2014 at 8:50 PM, James Cameron wrote:
> I'm not confident that it relates to the server part.
>
> When I looked at the tcpdump it wasn't clear that it _was_ the server
> pa
I'm not confident that it relates to the server part.
When I looked at the tcpdump it wasn't clear that it _was_ the server
part at fault; the client was keeping several connections open, it was
the client that initiated the connection closure after one minute, and
the server properly responded at
Now we know the problem, I can't understand why is related to webkit1
(at least to the server part)
Gonzalo
On Tue, Feb 4, 2014 at 8:15 PM, Manuel Quiñones wrote:
> 2014-02-04 Gonzalo Odiard :
> > Probably would be good discuss with the web developers, manuq or
> dnarvaez?
>
> add a "WebKit1 k
2014-02-04 Gonzalo Odiard :
> Probably would be good discuss with the web developers, manuq or dnarvaez?
add a "WebKit1 known issues" section somewhere around
http://developer.sugarlabs.org/#sugar%20web ?
we can also make clear that WebKit1 is used as fallback when WebKit2
is not available, and t
Probably would be good discuss with the web developers, manuq or dnarvaez?
Gonzalo
On Tue, Feb 4, 2014 at 5:11 PM, Lionel Laské wrote:
>
> Sure. Where do you think I could document it ?
>
> Lionel.
>
>
>
> 2014-02-04 Gonzalo Odiard :
>
> Great. Would be good have this issue docume
Sure. Where do you think I could document it ?
Lionel.
2014-02-04 Gonzalo Odiard :
> Great. Would be good have this issue documented.
>
> Gonzalo
>
>
>
___
Sugar-devel mailing list
Sugar-devel@lists.sugarlabs.org
http://lists.sugarlabs.o
Great. Would be good have this issue documented.
Gonzalo
On Mon, Feb 3, 2014 at 6:45 PM, Lionel Laské wrote:
> Got it!
> It come from the SVG files referenced by the toolbar and used for FR and
> EN flag.
>
> #main-toolbar #en-button {
> margin-left: 100px;
> background-image: url(images/e
Got it!
It come from the SVG files referenced by the toolbar and used for FR and EN
flag.
#main-toolbar #en-button {
margin-left: 100px;
background-image: url(images/en.svg);
}
#main-toolbar #fr-button {
margin-left: 5px;
background-image: url(images/fr.svg);
}
It seems that these files
Nice !
Great Gonzalo, I'm going to inspect that.
Lionel.
2014-02-03 James Cameron :
> On Mon, Feb 03, 2014 at 05:11:30PM -0200, Gonzalo Odiard wrote:
> > The lines with doctype and body are not related with the problem,
> > just commenting the reerence to styles.css solves the problem.
On Mon, Feb 03, 2014 at 05:11:30PM -0200, Gonzalo Odiard wrote:
> The lines with doctype and body are not related with the problem,
> just commenting the reerence to styles.css solves the problem.
>
> I have been looking at styles.css, but didn't find what could be the
> problem. I hope this is o
Hi Lionel,
I have found something.
Attached is a modified index.html, who makes the activity start fast.
The only changes are:
-
Food Chain
@@ -8,9 +7,10 @@
-
+
+
After start, the images are not well positioned,
but that is another problem.
The lines wi
On Thu, Jan 30, 2014 at 9:41 AM, Daniel Narvaez wrote:
> This seems something we should apply in any case.
>
>
At least looks right :)
I am waiting a little to see if we can find the solution to the problem
Lionel found,
but anyway I will send the patch.
Gonzalo
>
> On 28 January 2014 21:06,
I have tried in sugar-build, with and without the env variable, and can't
reproduce the issue.
Gonzalo
On Thu, Jan 30, 2014 at 9:41 AM, Daniel Narvaez wrote:
> In sugar-build you was not seeing the issue right?
>
> What I was suggesting was to enable webkit1 usage in sugar-build using
> that e
This seems something we should apply in any case.
On 28 January 2014 21:06, Gonzalo Odiard wrote:
> Lionel,
> Can you try this patch?
>
> diff --git a/src/sugar3/activity/webkit1.py
> b/src/sugar3/activity/webkit1.py
> index deaaa12..dfd0a50 100644
> --- a/src/sugar3/activity/webkit1.py
> +++ b
In sugar-build you was not seeing the issue right?
What I was suggesting was to enable webkit1 usage in sugar-build using that
env variable, and see if that made the issue appear. It wouldn't
immediately solve your issue, but being able to blame it on webkit1 (or
not) would be helpful in further d
Hi Lionel,
Lionel Laské writes:
> However the tool seems to be a bit paranoiac, it consider for example
> this file as NON-FREE: http://llaske.github.io/Sugarizer/js/
> constant.js !
It's just a matter of declaring licenses at the right place.
A bit boring I agree, but not difficult. I've been
Hi Bastien,
I've just tried. Don't pass :-(
However the tool seems to be a bit paranoiac, it consider for example this
file as NON-FREE: http://llaske.github.io/Sugarizer/js/constant.js !
Lionel
2014-01-29 Bastien
> Hi Lionel,
>
> would be nice to see if the sugarizer passes the lib
Hi Lionel,
would be nice to see if the sugarizer passes the librejs tests:
see http://www.gnu.org/software/librejs/
Thanks for your work on this!
--
Bastien
___
Sugar-devel mailing list
Sugar-devel@lists.sugarlabs.org
http://lists.sugarlabs.org/listi
Maybe you can compare with other web activities like ClockWeb [1]
and see if there are any other difference?
Gonzalo
[1] http://activities.sugarlabs.org/en-US/sugar/addon/4691
On Tue, Jan 28, 2014 at 5:56 PM, Lionel Laské wrote:
> Thanks for your help.
> Sorry to say that it's something I've t
Thanks for your help.
Sorry to say that it's something I've tested.
I've just done the test again with the patched version. No more result with
or without connection :-(
Lionel.
2014-01-28 Gonzalo Odiard
> No more ideas here, and I don't have a XO to test right now.
> In sugar-build
No more ideas here, and I don't have a XO to test right now.
In sugar-build in my desktop is not doing this.
Is your XO connected to a access point or not?
Can you test with both situations?
Gonzalo
On Tue, Jan 28, 2014 at 5:23 PM, Lionel Laské wrote:
>
> Done. But no more luck :-(
>
>
Done. But no more luck :-(
Lionel.
2014-01-28 Gonzalo Odiard
> Lionel,
> Can you try this patch?
>
> diff --git a/src/sugar3/activity/webkit1.py
> b/src/sugar3/activity/webkit1.py
> index deaaa12..dfd0a50 100644
> --- a/src/sugar3/activity/webkit1.py
> +++ b/src/sugar3/activity/webk
You're right.
favicon.ico was not in the original package. I've added it during my test
because I've seen the error message like you and thought that the timeout
come from it. But it's not the case, the error is the same with favicon.ico
and some other Sugar Web activities without favicon.ico don't
Lionel,
Can you try this patch?
diff --git a/src/sugar3/activity/webkit1.py b/src/sugar3/activity/webkit1.py
index deaaa12..dfd0a50 100644
--- a/src/sugar3/activity/webkit1.py
+++ b/src/sugar3/activity/webkit1.py
@@ -46,6 +46,8 @@ class LocalRequestHandler(BaseHTTPRequestHandler):
new_pat
There are something strange in this log:
This line:
1390662262.276506 ERROR root: Using port 50662
is definitively WEBKIT1 stuff (is the port where the server is running)
and the server show every file served,
but when I run it here, I get a line:
file /home/gonzalo/Activities/FoodChain.activity//
2014-01-28 James Cameron :
> I agree, no host lookups.
>
> But at the time before the delay, there are several connections open to the
> server process, and then after the delay there is a sudden server side
> connection reset and a new connection issued by the client side.
>
> I wish I knew wher
s [nop,nop,TS val 23013 ecr 12819], length 0
>
>
>
>
>
> 2014-01-27
>
> Date: Mon, 27 Jan 2014 10:00:26 +1100
> From: James Cameron
> To: sugar-devel@lists.sugarlabs.org
> Subject: Re: [Sugar-devel] Sugar web activity wait 1 minute at startup
> Message-I
7,
options [nop,nop,TS val 23013 ecr 12819], length 0
2014-01-27
>
> Date: Mon, 27 Jan 2014 10:00:26 +1100
> From: James Cameron
> To: sugar-devel@lists.sugarlabs.org
> Subject: Re: [Sugar-devel] Sugar web activity wait 1 minute at startup
> Message-ID: <20140126230026.gc26
The image is already setting SUGAR_USE_WEBKIT1.
I am downloading the activity and will test it here.
Gonzalo
On Mon, Jan 27, 2014 at 6:20 PM, Lionel Laské wrote:
>
> Thanks Daniel.
>
> Same issue with SUGAR_USE_WEBKIT1 so I guess it's not related to it.
>
> Lionel.
>
>
> 2014-01
Thanks Daniel.
Same issue with SUGAR_USE_WEBKIT1 so I guess it's not related to it.
Lionel.
2014-01-27 Daniel Narvaez
> The webkit1 code does not seem to be sending a response if a file does not
> exist, which I suppose could cause what you are seeing. Though it's logging
> an
Hi Gonzalo,
I've tested on the last testing image here:
http://wiki.sugarlabs.org/go/0.100/Testing
Don't know if it's webkit1.
Lionel.
2014-01-27 Gonzalo Odiard
> Hi Lionel,
> What image are you using to test sugar 0.100 in the xo?
> Probably you are using the webkit1 implementa
Hi Lionel,
What image are you using to test sugar 0.100 in the xo?
Probably you are using the webkit1 implementation, but I didn't see what
you report yet.
Gonzalo
On Sun, Jan 26, 2014 at 8:30 PM, Daniel Narvaez wrote:
> The webkit1 code does not seem to be sending a response if a file does no
The webkit1 code does not seem to be sending a response if a file does not
exist, which I suppose could cause what you are seeing. Though it's logging
an error, which I don't see in your log. But the locale.ini does not exist
in your activity I guess?
It would be good to know for sure if it's webk
Delays of around 40 seconds are consistent with host name resolution
timeout. Either forward or reverse.
Use tcpdump to capture network traffic, on both internal and external
network interfaces, and look for host lookups.
--
James Cameron
http://quozl.linux.org.au/
_
Thanks for your help Emil.
It seems to be a logical explanation, more precisely the 40 seconds that
make sense with my symptom.
Unfortunately I've detailed in Chrome - outside Sugar - the network
exchanges when the page is loaded and I don't see any loading error :-( My
first thought was that it c
Since you've said that after deleting the file, the loading time does not
reduce, I'd say that the problem is not in the file.
Have you tried to not include the file in the HTML pages? Does the load
time change?
I've had a similar problem. If I include a non existent file, webkit will
still try to
I guess the webkit1 stuff is to blame here? It uses a different way to
serve files, so it seems consistent with what you are seeing.
If that's the issue, Gonzalo might be able to help.
On Sunday, 26 January 2014, Lionel Laské wrote:
>
> Hi all,
>
> I'm currently porting my FoodChain activity [1
Hi all,
I'm currently porting my FoodChain activity [1] as a full Sugar Web
activity. It's mainly a packaging work because FoodChain is already written
in HTML5/JavaScript but with a Python package around it.
The new Sugar Web activity works well but when I'm testing it on Sugar
0.100 on the XO,
41 matches
Mail list logo