Stalled
> is it possible to change it to pyjamas which is python base ? to
> reduce technology involved ?
When I started sao, the state of pyjamas was not enough for such big
application. But now, I don't know if it will be or not.
--
Cédric Krier
B2CK SPRL
Rue de Rotterdam, 4
4000 Liège
Belgium
Tel: +32 472 54 46 59
Email/Jabber: cedric...@b2ck.com
Website: http://www.b2ck.com/
A Dilluns, 18 d'abril de 2011 10:25:10, C�dric Krier va escriure:
> On 17/04/11 19:42 -0700, gops wrote:
> > What is the state of tryton web client ?
>
> Stalled
>
> > is it possible to change it to pyjamas which is python base ? to
> > reduce technology involved ?
>
> When I started sao, the state of pyjamas was not enough for such big
> application. But now, I don't know if it will be or not.
Another option I evaluated recently was using jwt [1] with Jython and seemd to work well. It may also not be extremely difficult (even if it's a quite a lot of work) to port Koo [2] to using jwt... (did a very very raw version of porting the login and server selection dialogs a few days ago).
Wt is really cool because it allows creating great interfaces with latest web technologies without having a clue of javascript and browser independent. Anyone who knows Qt can use jwt easily and with jython it also avoids the need for learning yet another language. Python only, and not even JavaScript!
[1] http://www.webtoolkit.eu/jwt
[2] http://www.nan-tic.com/en/koo-platform
--
Albert Cervera i Areny
Tel: +34 93 553 18 03
skype: nan-oficina
It is based on javascript and the backend. can use every-think
On Mon, Apr 18, 2011 at 9:09 AM, Albert Cervera i Areny
<alb...@nan-tic.com> wrote:
> A Dilluns, 18 d'abril de 2011 10:25:10, Cédric Krier va escriure:
> --
> tryto...@googlegroups.com mailing list
>
--
Andres Vargas
www.zodman.com.mx
A Dilluns, 18 d'abril de 2011 16:09:05, Albert Cervera i Areny va escriure:
> A Dilluns, 18 d'abril de 2011 10:25:10, C�dric Krier va escriure:
> > On 17/04/11 19:42 -0700, gops wrote:
> > > What is the state of tryton web client ?
> >
> > Stalled
> >
> > > is it possible to change it to pyjamas which is python base ? to
> > > reduce technology involved ?
> >
> > When I started sao, the state of pyjamas was not enough for such big
> > application. But now, I don't know if it will be or not.
>
> Another option I evaluated recently was using jwt [1] with Jython and seemd
> to work well. It may also not be extremely difficult (even if it's a quite
> a lot of work) to port Koo [2] to using jwt... (did a very very raw
> version of porting the login and server selection dialogs a few days ago).
>
> Wt is really cool because it allows creating great interfaces with latest
> web technologies without having a clue of javascript and browser
> independent. Anyone who knows Qt can use jwt easily and with jython it
> also avoids the need for learning yet another language. Python only, and
> not even JavaScript!
Just in case someone is curious about JyWt:
First, I would like to clarify some thoughts about a web client.
The primary client of Tryton should stay the GTK one because it is the best
way (aka thin desktop client) to have great usability, integration, security
and fast responsiveness.
I think the approach of having a clone of the GTK client is not the best.
Because we could have the need to embed some parts/functionnalities inside
other website (like displaying the invoices of customer on company website).
So we (Bertrand and me) think that having a modular/library approach is
better. For that we need to build some javascript libs:
- proteus.js: a clone of proteus but in JS. The goal of this lib is to
provide an API to access to the server Models through the JSON-RPC.
- ???.js: a library that will convert XML view into HTML and could be
linked to a record (provided by proteus.js).
- a library that runs wizards.
- a library that runs reports.
- a library that manages login.
- ...
Like that we will have blocks that we can assemble to build a complete web
client but also we can use just some part to build a specific application
(like a web mobile addressbook or a shipment assignation application).
We think about using JQuery as an helper to build the XML2HTML and the
JSON-RPC parts.
Bertrand is working on a prototype of proteus.js. He will submit ASAP a
codereview to discuss the implementation.
jquery is after a lot of tries my preferred solution. The only bib which
is an other favourite of mine is extjs - but isn't gpl I thing.
Jan
I can understand that. Too bad it's a nightmare from a UX point of view.
But please don't take that the wrong way. I'm very happy with the 2.0
changes - though I fully agree with the analysis by erp@northman as I
understand it on the tryton list: keep up the good work, and cleanup
some more :-)
> I think the approach of having a clone of the GTK client is not the best.
Yippee! Let's not repeat the old mistakes...
> Because we could have the need to embed some parts/functionnalities inside
> other website (like displaying the invoices of customer on company website).
> So we (Bertrand and me) think that having a modular/library approach is
> better. For that we need to build some javascript libs:
>
> - proteus.js: a clone of proteus but in JS. The goal of this lib is to
> provide an API to access to the server Models through the JSON-RPC.
>
> - ???.js: a library that will convert XML view into HTML and could be
> linked to a record (provided by proteus.js).
Just those two for now will do really nice for me. A xml2html python
sibling for proteus.py would also do of course, but I'm not picky.
At least templating should be supported by the javascript version. I
didn't get too exited about jquery's capabilities in that regard. On the
other had, Mochikit has excellent support for generating html! And I'm
sure other toolkits can be used to cover that area as well.
>
> - a library that runs wizards.
>
> - a library that runs reports.
>
> - a library that manages login.
Those can easily be handled by server apps using the xml-rpc channel.
Stuff like reports and authentication I prefer to handle through a
separate middleware anyway - mostly for security and performance reasons.
> - ...
>
> Like that we will have blocks that we can assemble to build a complete web
> client but also we can use just some part to build a specific application
> (like a web mobile addressbook or a shipment assignation application).
>
> We think about using JQuery as an helper to build the XML2HTML and the
> JSON-RPC parts.
JQuery is at the very least a safe choice to get started. I love it's
ajax interfaces.
> Bertrand is working on a prototype of proteus.js. He will submit ASAP a
> codereview to discuss the implementation.
Looking forward to it.
--
________________________________________________________________
Paul Stevens paul at nfg.nl
NET FACILITIES GROUP GPG/PGP: 1024D/11F8CD31
The Netherlands________________________________http://www.nfg.nl
Mochikit and JQuery has almost the same capabilities. To generete html, you
just need to have access to the DOM.
> Mochikit and JQuery has almost the same capabilities. To generete html, you
> just need to have access to the DOM.
Eh, that's just not true. The part about 'same capabilities' that is.
In mochikit you can do stuff like:
var rows = [
["dataA1", "dataA2", "dataA3"],
["dataB1", "dataB2", "dataB3"]
];
row_display = function (row) {
return TR(null, map(partial(TD, null), row));
}
var newTable = TABLE({'class': 'prettytable'},
THEAD(null,
row_display(["head1", "head2", "head3"])),
TFOOT(null,
row_display(["foot1", "foot2", "foot3"])),
TBODY(null,
map(row_display, rows)));
which goes a long way toward full dynamic templating.
Too bad mochikit appears to be dead in the water.
Don't know enough both libs but I would mean they provide *almost* the same
facilities. Of course one will have this cool feature the other doesn't etc.
But it doesn't matter a lot. We need to choose one lib that is stable, robust
and be there in the next 5 years. Because at the end, we just generate HTML :-)
I finally managed to upload a clean patch on codereview:
http://codereview.tryton.org/24003/
All the previous codereview may be ignored (sorry for spamming the
mailing list).
So this contains TrytonProxy which implements base RPC to the server,
test.html show a first example usage.
It use the new deferred mechanism introduced by Jquery 1.5 and the
example show the pipe() method (introduced by Jquery1.6, released
today). So IMO it gives a nice API (goodbye long callback chains!).
It already handle de-connection, it's not visible in the example, but
it allows to dynamically open a popup to ask the user password when the
session has expired (I.E. when the server send a 'NotLogged' error).
Next step: Define a Model and a List object and bind CRUD operation on
TrytonProxy.
Bertrand
--
Bertrand Chenal
B2CK SPRL
Rue de Rotterdam, 4
4000 Liège
Belgium
Tel: +32 472 54 46 59
Email: bertran...@b2ck.com
Website: http://www.b2ck.com/