A bug in the Technorati's events exporting tool

Hi,

I've hCal'd a website that contains with this kind of dtend
:

<abbr class="dtend"
title="2006-09-24T02:00:00+02:00">...</abbr&
gt;

and feeds.technorati.com/events/ convert it to

DTEND:20060924T0000000Z

with se7en 0 instead of six.

Test it from there : 
http://feeds.technorati.com/events/ht
tp://www.bikinitest.ch/agenda/2006/09/

Cheers,

--
Yoan Blanc
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

Good catch! i had a look at the issue and the problem was
that all our
test cases "added" the time from the timezone up
to 240000, in your
case you were subtracting to get '0' which was not being
checked for
so things were going padded then that '0' was being added so
we got
'7' instead of '6'. This has been corrected[1]. I have
committed the
changes to HG, so anyone at technorati can pull them down
(you might
need a few files - i'm not sure which version technorati is
running).
I have updated the code at http://suda.co.uk/pro
jects/X2V/ so you can
test it there as well. I will also add a 'subtraction'
instance into
the test cases as well.

-brian

[1] - http://hg.microformats.org/x2v?fd=753da7c7d63c;f
ile=datetime.xsl

On 9/25/06, Yoan BLANC <greut.listsdosimple.ch> wrote:
> Hi,
>
> I've hCal'd a website that contains with this kind of
dtend :
>
> <abbr class="dtend"
title="2006-09-24T02:00:00+02:00">...</abbr&
gt;
>
> and feeds.technorati.com/events/ convert it to
>
> DTEND:20060924T0000000Z
>
> with se7en 0 instead of six.
>
> Test it from there :
> http://feeds.technorati.com/events/ht
tp://www.bikinitest.ch/agenda/2006/09/
>
> Cheers,
>
> --
> Yoan Blanc
> _______________________________________________
> microformats-discuss mailing list
> microformats-discussmicroformats.org
> http://microformats.org/mailman/listinfo/microforma
ts-discuss
>


-- 
brian suda
http://suda.co.uk
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

Brian Suda wrote:
> Good catch! i had a look at the issue and the problem
was that all our
> test cases "added" the time from the timezone
up to 240000, in your
> case you were subtracting to get '0' which was not
being checked for
> so things were going padded then that '0' was being
added so we got
> '7' instead of '6'. This has been corrected[1]. I have
committed the
> changes to HG, so anyone at technorati can pull them
down (you might
> need a few files - i'm not sure which version
technorati is running).
> I have updated the code at http://suda.co.uk/pro
jects/X2V/ so you can
> test it there as well. I will also add a 'subtraction'
instance into
> the test cases as well.
x2v add serious bug with utf-8 (that technorati doesn't
have) and now 
the hours are false, not buggy but simply false.

this:

<abbr class="dtend"
title="2006-09-30T02:00:00+02:00">...</abbr&
gt;

becomes:

DTEND:20060930T020000Z

It's obviously false because:

Z is UTC
+02:00 is CEST

so 2am (CEST) can't be 2am (Z)

http://suda.co.uk/projec
ts/X2V/get-vcal.php?uri=http://www.bikinitest.ch/agenda/2006
/09/

Thanks for you work

--
Yoan
>
> -brian
>
> [1] - http://hg.microformats.org/x2v?fd=753da7c7d63c;f
ile=datetime.xsl
>
> On 9/25/06, Yoan BLANC <greut.listsdosimple.ch> wrote:
>> Hi,
>>
>> I've hCal'd a website that contains with this kind
of dtend :
>>
>> <abbr class="dtend"
title="2006-09-24T02:00:00+02:00">...</abbr&
gt;
>>
>> and feeds.technorati.com/events/ convert it to
>>
>> DTEND:20060924T0000000Z
>>
>> with se7en 0 instead of six.
>>
>> Test it from there :
>> http://feeds.technorati.com/events/ht
tp://www.bikinitest.ch/agenda/2006/09/ 
>>
>>
>> Cheers,
>>
>> -- 
>> Yoan Blanc
>> _______________________________________________
>> microformats-discuss mailing list
>> microformats-discussmicroformats.org
>> http://microformats.org/mailman/listinfo/microforma
ts-discuss
>>
>
>

_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

hmm...
1) I'll look into the UTF-8 issue, Your server is NOT
sending any
language encoding:
HTTP/1.1 200 OK
Date: Tue, 26 Sep 2006 12:02:13 GMT
Server: Apache
Transfer-Encoding: chunked
Content-Type: text/html

so i might be assuming it as UTF-8? when i run your code
through the
W3C tidy, they are NOT finding any encoding so it defaults
to
<?xml version="1.0"
encoding="iso-8859-1"?>

That might be why there are characters incorrectly encoded.
Do you
expect UTF-8 or ISO-8859-1?

As for the incorrect dtend: i couldn't find your example in
your HTML,
so i can't verify the error. Can you point to an example you
know is
incorrect so i can check it?

The one on the 24th:
<abbr class="dtend"
title="2006-09-24T02:00:00+02:00">02:00</abb
r>
converts to:
DTEND:20060924T000000Z
which is correct.

Thanks,
-brian

> x2v add serious bug with utf-8 (that technorati doesn't
have) and now
> the hours are false, not buggy but simply false.
>
> this:
>
> <abbr class="dtend"
title="2006-09-30T02:00:00+02:00">...</abbr&
gt;
>
> becomes:
>
> DTEND:20060930T020000Z
>
> It's obviously false because:
>
> Z is UTC
> +02:00 is CEST
>
> so 2am (CEST) can't be 2am (Z)
>
> http://suda.co.uk/projec
ts/X2V/get-vcal.php?uri=http://www.bikinitest.ch/agenda/2006
/09/
>
> Thanks for you work
>
> --
> Yoan
> >
> > -brian
> >
> > [1] - http://hg.microformats.org/x2v?fd=753da7c7d63c;f
ile=datetime.xsl
> >
> > On 9/25/06, Yoan BLANC <greut.listsdosimple.ch> wrote:
> >> Hi,
> >>
> >> I've hCal'd a website that contains with this
kind of dtend :
> >>
> >> <abbr class="dtend"
title="2006-09-24T02:00:00+02:00">...</abbr&
gt;
> >>
> >> and feeds.technorati.com/events/ convert it to
> >>
> >> DTEND:20060924T0000000Z
> >>
> >> with se7en 0 instead of six.
> >>
> >> Test it from there :
> >> http://feeds.technorati.com/events/ht
tp://www.bikinitest.ch/agenda/2006/09/
> >>
> >>
> >> Cheers,
> >>
> >> --
> >> Yoan Blanc
> >>
_______________________________________________
> >> microformats-discuss mailing list
> >> microformats-discussmicroformats.org
> >> http://microformats.org/mailman/listinfo/microforma
ts-discuss
> >>
> >
> >
>
> _______________________________________________
> microformats-discuss mailing list
> microformats-discussmicroformats.org
> http://microformats.org/mailman/listinfo/microforma
ts-discuss
>


-- 
brian suda
http://suda.co.uk
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

Hi,

Brian Suda wrote:
> hmm...
> 1) I'll look into the UTF-8 issue, Your server is NOT
sending any
> language encoding:
> HTTP/1.1 200 OK
> Date: Tue, 26 Sep 2006 12:02:13 GMT
> Server: Apache
> Transfer-Encoding: chunked
> Content-Type: text/html
if my memory is good.

text/html is by default: iso-8859-1
text/xml is by default: utf-8

Do you prefer if I send a Content-Type: text/html;
charset=iso-8859-1 
It's much a matter of your tool. Technorati deals with this
page 
correctly (except with the dates) so why x2v cannot, strange
no?
>
> so i might be assuming it as UTF-8? when i run your
code through the
> W3C tidy, they are NOT finding any encoding so it
defaults to
> <?xml version="1.0"
encoding="iso-8859-1"?>
>
> That might be why there are characters incorrectly
encoded. Do you
> expect UTF-8 or ISO-8859-1?
DESCRIPTION;LANGUAGE=fr;CHARSET=utf-8:M�langez The Sonics,
...

it tells utf-8 but gives a strangly encoded utf-8 string
('�' should 
be an 'é').'�' looks like BOM, isn't it ?
>
> As for the incorrect dtend: i couldn't find your
example in your HTML,
> so i can't verify the error. Can you point to an
example you know is
> incorrect so i can check it?
>
> The one on the 24th:
> <abbr class="dtend"
title="2006-09-24T02:00:00+02:00">02:00</abb
r>
> converts to:
> DTEND:20060924T000000Z
> which is correct.
It's correct, I'm sorry. My fault.

Btw, I've tried this tool and Google Calendar doesn't deal
with Summer 
Time. Bad bad bad.

http://torrez.us/code/googlehcal/converthcal
endartogooglecal.user.js

Cheers,

Yoan
>
> Thanks,
> -brian
>
>> x2v add serious bug with utf-8 (that technorati
doesn't have) and now
>> the hours are false, not buggy but simply false.
>>
>> this:
>>
>> <abbr class="dtend"
title="2006-09-30T02:00:00+02:00">...</abbr&
gt;
>>
>> becomes:
>>
>> DTEND:20060930T020000Z
>>
>> It's obviously false because:
>>
>> Z is UTC
>> +02:00 is CEST
>>
>> so 2am (CEST) can't be 2am (Z)
>>
>> http://suda.co.uk/projec
ts/X2V/get-vcal.php?uri=http://www.bikinitest.ch/agenda/2006
/09/ 
>>
>>
>> Thanks for you work
>>
>> -- 
>> Yoan
>> >
>> > -brian
>> >
>> > [1] - http://hg.microformats.org/x2v?fd=753da7c7d63c;f
ile=datetime.xsl
>> >
>> > On 9/25/06, Yoan BLANC <greut.listsdosimple.ch> wrote:
>> >> Hi,
>> >>
>> >> I've hCal'd a website that contains with
this kind of dtend :
>> >>
>> >> <abbr class="dtend"
title="2006-09-24T02:00:00+02:00">...</abbr&
gt;
>> >>
>> >> and feeds.technorati.com/events/ convert
it to
>> >>
>> >> DTEND:20060924T0000000Z
>> >>
>> >> with se7en 0 instead of six.
>> >>
>> >> Test it from there :
>> >> 
>> http://feeds.technorati.com/events/ht
tp://www.bikinitest.ch/agenda/2006/09/ 
>>
>> >>
>> >>
>> >> Cheers,
>> >>
>> >> --
>> >> Yoan Blanc
>> >>
_______________________________________________
>> >> microformats-discuss mailing list
>> >> microformats-discussmicroformats.org
>> >> http://microformats.org/mailman/listinfo/microforma
ts-discuss
>> >>
>> >
>> >
>>
>> _______________________________________________
>> microformats-discuss mailing list
>> microformats-discussmicroformats.org
>> http://microformats.org/mailman/listinfo/microforma
ts-discuss
>>
>
>

_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

On 9/26/06, Yoan BLANC <greut.listsdosimple.ch> wrote:
> Do you prefer if I send a Content-Type: text/html;
charset=iso-8859-1
> It's much a matter of your tool. Technorati deals with
this page
> correctly (except with the dates) so why x2v cannot,
strange no?

I'll look into this now that i know you want it to be
"iso-8859-1".
The technorati service and my site are running the same
XSLT. To
actually execute the transform you need something else. In
my case, i
am using PHP, i don't know what technorati uses to call the
XML and
XSLT.

Since the XSLT can't detect the HTTP Header mimeType that
needs to be
passed in from the executing code (in my case PHP, where i
have it
wrong).

I won't go too much technical detail on the discuss list,
you can
email me off list for a complete explaination and how the
W3C defines
the order of where to look for language encodings, etc.

-brian

-- 
brian suda
http://suda.co.uk
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

Brian Suda wrote:
> On 9/26/06, Yoan BLANC <greut.listsdosimple.ch> wrote:
>> Do you prefer if I send a Content-Type: text/html;
charset=iso-8859-1
>> It's much a matter of your tool. Technorati deals
with this page
>> correctly (except with the dates) so why x2v
cannot, strange no?
>
> I'll look into this now that i know you want it to be
"iso-8859-1".
I just want it to be an usable .ics. Good utf-8 or good
iso-8859-1. I'll 
use the technorati service so if the date with a positive
timezone 
problem is fixed I don't bother.

Thanks for your work,

Yoan
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

On Sep 26, 2006, at 5:44 AM, Brian Suda wrote:
> On 9/26/06, Yoan BLANC <greut.listsdosimple.ch> wrote:
>> Do you prefer if I send a Content-Type: text/html;
charset=iso-8859-1
>> It's much a matter of your tool. Technorati deals
with this page
>> correctly (except with the dates) so why x2v
cannot, strange no?
>
> I'll look into this now that i know you want it to be
"iso-8859-1".
> The technorati service and my site are running the same
XSLT. To
> actually execute the transform you need something else.
In my case, i
> am using PHP, i don't know what technorati uses to call
the XML and
> XSLT.

We're using PHP, but its a different implementation than
yours, Brian.

-ryan
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

On Sep 25, 2006, at 6:48 AM, Brian Suda wrote:

> This has been corrected[1]. I have committed the
changes to HG, so  
> anyone at technorati can pull them down

done.

-ryan
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

On Sep 26, 2006, at 5:44 AM, Brian Suda wrote:
> I won't go too much technical detail on the discuss
list, you can
> email me off list for a complete explaination and how
the W3C defines
> the order of where to look for language encodings, etc.


Our own Mark Pilgrim wrote a good explanation of this:

http://www.feedparser.org/docs/character-encoding.html

_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss