Smalltalk › Frameworks & Tools › Seaside › Seaside General

[Q] WAKomEncoded equivalent thing for WAListener?

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

6 messages Options

Chun, Sungjin

[Q] WAKomEncoded equivalent thing for WAListener?

Hi,

I've used WAKomEncoded for my web app, and found that my application
does not work with WAListener if there's Korean character in page/response.

As I said I've used WAKomEncoded and with this, my application works well :-)

The problem is that with WAKomEncoded, I can assume all my instances of string
is encoded as Squeak Native internal encoding and if I make my string conforms to
this, everything is OK. (Yes, I have to convert encoding from/to Squeak Native/UTF-8
when using external resources like Database, which uses UTF-8).

So, my question is that what part should I change to make WAKomEncoded equivalent
thing for WAListener, for processing input encoding, I can make it referring WAKom and
WAKomEncoded, but I cannot make proper conversion for output. My response string
looks like this

" H T T P 2 0 0 O K..."

instead of

"HTTP 200 OK..."

if it contains korean(or any other multibyte encoding characters).

Thanks in advance.

_______________________________________________
Seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

Lukas Renggli

Re: [Q] WAKomEncoded equivalent thing for WAListener?

> I've used WAKomEncoded for my web app, and found that my application
> does not work with WAListener if there's Korean character in page/response.

Yes, this is a problem with the current (bad) design of the server
adaptors: automatic encoding is only supported in the WAKom hierarchy.
It would be probably better if we moved this encoding transformation
as a strategy into the request/response so that it can be reused with
any server adapter.

So far I was probably the only one using WAListener. As you maybe know
from earlier posts of mine, I don't trust the encoding of Squeak and
just keep all data in UTF-8 in my images.

Lukas

--
Lukas Renggli
http://www.lukas-renggli.ch
_______________________________________________
Seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

cbeler

Re: [Q] WAKomEncoded equivalent thing for WAListener?

Lukas Renggli a écrit :

>> I've used WAKomEncoded for my web app, and found that my application
>> does not work with WAListener if there's Korean character in page/response.
>>
>
> Yes, this is a problem with the current (bad) design of the server
> adaptors: automatic encoding is only supported in the WAKom hierarchy.
> It would be probably better if we moved this encoding transformation
> as a strategy into the request/response so that it can be reused with
> any server adapter.
>

Besides, this encoding problem, is WAListener a sort of classic WAKom
when not streaming and also a streaming adaptor when needed ? In other
word, could we use only WAListener (if no encoding stuff...) ? Is it
what you do ?

> So far I was probably the only one using WAListener. As you maybe know
> from earlier posts of mine, I don't trust the encoding of Squeak and
> just keep all data in UTF-8 in my images.
>
>
What's the drawback of having UTF-8 data in the image ? you cannot
"change" them in squeak nor inspect ?

Thanks

Cédrick
_______________________________________________
Seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

Lukas Renggli

Re: [Q] WAKomEncoded equivalent thing for WAListener?

> >> I've used WAKomEncoded for my web app, and found that my application
> >> does not work with WAListener if there's Korean character in page/response.
> >
> > Yes, this is a problem with the current (bad) design of the server
> > adaptors: automatic encoding is only supported in the WAKom hierarchy.
> > It would be probably better if we moved this encoding transformation
> > as a strategy into the request/response so that it can be reused with
> > any server adapter.
> >
> Besides, this encoding problem, is WAListener a sort of classic WAKom
> when not streaming and also a streaming adaptor when needed ?

WAListener implements a minimal web server on its own. It uses the
request parsing of Kom though.

> In other
> word, could we use only WAListener (if no encoding stuff...) ? Is it
> what you do ?

WAKom is equivalent to WAListener. There are no equivalents for
WAKomEncoded, WAKomEncoded39 and WAKomEncodedSuperPlus ;-)

> > So far I was probably the only one using WAListener. As you maybe know
> > from earlier posts of mine, I don't trust the encoding of Squeak and
> > just keep all data in UTF-8 in my images.
> >
> What's the drawback of having UTF-8 data in the image ? you cannot
> "change" them in squeak nor inspect ?

In my experience this is usually not a problem. For example, #size
might not answer what you expect: it answers the bytes the string
consists of, not the number of actual characters. All the stream
methods, and #copyFrom:to:, #indexOf:, #findSubString:, #, etc. work
fine (for example all that is needed in the Pier parser). Data might
look scrambled in the inspectors, but you can easily convert it using
#utf8ToIso and #isoToUtf8.

Cheers,
Lukas

--
Lukas Renggli
http://www.lukas-renggli.ch
_______________________________________________
Seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

Philippe Marschall

Re: [Q] WAKomEncoded equivalent thing for WAListener?

2007/9/9, Lukas Renggli <[hidden email]>:

> > >> I've used WAKomEncoded for my web app, and found that my application
> > >> does not work with WAListener if there's Korean character in page/response.
> > >
> > > Yes, this is a problem with the current (bad) design of the server
> > > adaptors: automatic encoding is only supported in the WAKom hierarchy.
> > > It would be probably better if we moved this encoding transformation
> > > as a strategy into the request/response so that it can be reused with
> > > any server adapter.
> > >
> > Besides, this encoding problem, is WAListener a sort of classic WAKom
> > when not streaming and also a streaming adaptor when needed ?
>
> WAListener implements a minimal web server on its own. It uses the
> request parsing of Kom though.
>
> > In other
> > word, could we use only WAListener (if no encoding stuff...) ? Is it
> > what you do ?
>
> WAKom is equivalent to WAListener. There are no equivalents for
> WAKomEncoded, WAKomEncoded39 and WAKomEncodedSuperPlus ;-)
>
> > > So far I was probably the only one using WAListener. As you maybe know
> > > from earlier posts of mine, I don't trust the encoding of Squeak and
> > > just keep all data in UTF-8 in my images.
> > >
> > What's the drawback of having UTF-8 data in the image ? you cannot
> > "change" them in squeak nor inspect ?
>
> In my experience this is usually not a problem. For example, #size
> might not answer what you expect: it answers the bytes the string
> consists of, not the number of actual characters. All the stream
> methods, and #copyFrom:to:, #indexOf:, #findSubString:, #, etc. work
> fine (for example all that is needed in the Pier parser). Data might
> look scrambled in the inspectors,

This is also the case for WideStrings because they miss language tags.

Cheers
Philippe

> but you can easily convert it using
> #utf8ToIso and #isoToUtf8.
>
> Cheers,
> Lukas
>
> --
> Lukas Renggli
> http://www.lukas-renggli.ch
> _______________________________________________
> Seaside mailing list
> [hidden email]
> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside
>

_______________________________________________
Seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

Philippe Marschall

Re: [Q] WAKomEncoded equivalent thing for WAListener?

In reply to this post by Lukas Renggli

2007/9/9, Lukas Renggli <[hidden email]>:

There was once one. ;-)

>There are no equivalents for
> WAKomEncoded, WAKomEncoded39 and WAKomEncodedSuperPlus ;-)
>
> > > So far I was probably the only one using WAListener. As you maybe know
> > > from earlier posts of mine, I don't trust the encoding of Squeak and
> > > just keep all data in UTF-8 in my images.
> > >
> > What's the drawback of having UTF-8 data in the image ? you cannot
> > "change" them in squeak nor inspect ?
>
> In my experience this is usually not a problem. For example, #size
> might not answer what you expect: it answers the bytes the string
> consists of, not the number of actual characters.

And anything that includes upper/lower case and whitespaces (outside
ASCII) does not work.

> All the stream
> methods, and #copyFrom:to:, #indexOf:, #findSubString:, #, etc. work
> fine

Only for ASCII strings/characters of course.

> (for example all that is needed in the Pier parser). Data might
> look scrambled in the inspectors, but you can easily convert it using
> #utf8ToIso and #isoToUtf8.

Only for Latin1 strings of course.

Cheers
Philippe

> Cheers,
> Lukas
>
> --
> Lukas Renggli
> http://www.lukas-renggli.ch
> _______________________________________________
> Seaside mailing list
> [hidden email]
> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside
>

_______________________________________________
Seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside