World: r3wp

Browser is FF 1.5.

Fired up a fresh instance of FF, and it worked there.

Please add a ticket about that error in CureCode, I'll try to find 
a FF v1.5 to try to reproduce it.

I registered okay, FF 2004

.

Curecode new release : 0.9.1 - http://softinnov.org:8000/curecode/

Changelog:

	o Localization framework (english and french catalogs provided).
	o ChangeLog page added.
	o Ticket history added.
	o Tickets list sorting added.
	o Numerous bugfixes (see CureCode project)
	
Needs Cheyenne 0.9.16 (to be released soon).

Package download URL : http://softinnov.org/dl/curecode-r091.zip

Looks like there's some bugs in the online version with localization 
(french browsers get a mix of french and english).

Seems to be ok now after a reboot.

Looks nice.  I noticed that on column sorting, depending on the sort, 
columns resize differently.

I'm far better at REBOL coding than at CSS, at least, with REBOL 
(or any decent language), I can predict the behaviour...;-) If anyone 
knows how to fix such things in a cross-browser manner, I'll be happy 
to know.

If you want to see the french translation of CureCode, just add french 
language in your browser language options (and ensure that it appears 
first in the list), or...just use a french browser version ;-). By 
default, the lang is controlled by the browser, but it can be overloaded 
by the application if required (next release will add a language 
option in user profile).

Curecode needs new Cheyenne ?

ie. unreleased version?

Nenad, why build this, as opposed to use Trac, or RAMBO?


(sometimes people think I'm being sarcastic when I ask such questions, 
but I'm genuinely curious what your logic is).

It's a demonstration of Cheyenne

Hmmm. teach me...how does this demo Cheyenne, I'm missing the connection...

Needs Cheyenne 0.9.16 (to be released soon).

Graham: yes, the localization framework has been added to 0.9.16, 
so CureCode won't work on 0.9.15.

CureCode is built using the RSP engine provided by Cheyenne.

Reichart, it's a good and logical question. There're several reasons 
:


1) To have a real application to test the concepts and implementation 
of Cheyenne's RSP engine. It has proved to be really useful in helping 
me see the gaps and find bugs in RSP engine.


2) To demo the features of RSP framework and Cheyenne's performances. 
(CureCode is really fast !)


3) RAMBO is very limited (able to handle only one project), the bugtracking 
part of Trac is quite limited too (compared to more popular products 
like Mantis or Bugzilla). Mantis was my favourite one but it's badly 
coded and failed once loosing all my tickets! So CureCode will gave 
me a lightweight Mantis-like solution for all our bugtracking needs.


4) I want to be able to build such kind of web application in just 
a few days. So I need to study each steps of the full creation process 
to see how some parts could be automated by a smart framework.

Good reasons, and I agree, the best way to do this.
We nee da universal XML export import with ID compare.  



In other words, CureCode needs to be able to export itself as XML 
(Easy).
It needs to import XML (Easy)


It needs to import XML and update the old XML keying on some ID value. 
 Not so easy.


Once done, building things like RSS feeds with just the top 10 recent 
changes becomes easy.


Also, then we can sync Qtask and CureCode.  There are a lot of cool 
side effects though…


-	We have an integrated file share, so you can link to screen shots.
-	Qtask has the ability to have conversation about each issue.
-	You can throw your issues onto a calendar view.


I know this is not your concern right now, but this is the direction 
we are going in, and need to learn how to best allow this type of 
integration.  Much as you have made one thing to test another, we 
have the same problem constantly.

R: if both QTask and CureCode are Rebol, why the need for the XML 
middle man?  Especially when Rebol exchange data is more versatile 
than XML.

Because this is not about Qtask, or Cheyenne (CureCode), it is about 
building things that lets others know that what they have can be 
made better by working with us, and if we work well with others (and 
each other) then we all win.


XML sucks, and sure we could exchange RBF (Rebol Block Format).  
But millions of programs work with XML, so just use it.  

Let's pick out battles…

I really like having a full example of a Cheyenne RSP solution and 
appreciate the effort that Doc made to build and disclose his coding 
approach.

I've thought about interfacing with external world. I need first 
to define a consistent kernel for CureCode (partially done in the 
DB abstraction layer) then I could easily interface with outside 
world through reb|web-services, REST, XML, RSS,...

Syncing with Qtask, why not, CureCode just needs to be a little bit 
more mature. One cool thing I'd like to have is an embeddable version 
of CureCode : I'd like to distribute a single binary with Cheyenne 
and CureCode bundled together (using sqlite as backend) for fast 
and very easy enterprise deployment. All the product will remain 
FOSS, but my company could sell support, customization and specific 
interfacing with third-party softwares.

All cool stuff...

typo: Reproductibility = should be Reproducibility

Where can I download the latest version of curecode ? The links posted 
are dead. thanks

TRETBASE might work as a backend, at the minimum the newest builds 
since it is server based.

The files were not restored after last disk crash. It's done now, 
so this URL will work : http://softinnov.org/dl/curecode-r091.zip

Btw, here's also the data model used by Curecode : http://softinnov.org/dl/DB-schema.png

Paul : if TRETBASE can support the same DB structure, support SQL 
JOINs (or similar feature), user and security management, not being 
more verbose than SQL, and being as fast as MySQL without using more 
memory, I'll drop MySQL at once and use TRETBASE as a backend in 
all my projects.

Doc give me a bit more time and I will have that.  It isn't ACID 
compliant yet so wouldn't recommend in its current state except in 
very controlled situations.

Doc, nice DB schema pic. What have you used to create it?

How does virtual domains work?

Schema : I'm using Microsoft Visio.

Virtual domains: Is your question related to CureCode or Cheyenne 
?

Sorry, Cheyenne...

Replying in Cheyenne group.

that picture reminds me Ladybird a Database designer wirted in rebol 
(project by reprotizer and Coccinelle)

hum curecode is a ticket tracker ?  like rambo but done with cheyenne 
that's right ?

sorry  i'm discovering it ....

CureCode is a lightweight bugtracker inspired by Mantis (a popular 
tracker in php). It's implemented using RSP and use MySQL as a backend.

it's nice anyway  good job dock  ^^

Doc, how hard would it be to switch to SQLite in the backend?

It's a planned feature. It just needs a SQLite REBOL driver with 
an API compatible with my MySQL driver or RT's DB drivers.

I guess that it would also needs a few changes in the SQL requests 
also, MySQL specific features like LAST_INSERT_ID( ) need to be ported. 
Anyway, it shouldn't take more than a couple of hour to make it work 
with SQLite (once you have the right driver).

On http://softinnov.org:8000/curecode I have tried to create an account 
to test it a bit.

But, when using the link in the mail to confirm my account I have 
an error in curecode :

> RSP Error Trapped<
Error Code : 	310
Description : 	script error !
validate-user has no refinement called update
Near : 	[if not validate-user/update request/content [err?: yes]]
Where : 	protected-exec