Paradox Community

Items in pnews.paradox-dos

Subject:Re: Qbe date parse results - final?
Date:Thu, 19 Apr 2018 16:02:10 +1000
From:Bernie van't Hof <>
The client is very simple. Just enough to get menus to work and some other fields like status
and messages. It does 
respond to keyboard, not mouse. Mouse works on menus, in fact at the moment only the mouse works
on menus.

When the websocket server detects a valid new client it kicks off a new php cli instance for
the app and the sockets for 
the app to communicate with the wss server. The app sends stuff to the client eg a menu, and
waits for a response.

Its all very primitive.

Re validation: Obviously this is reqd at the server end, and it would be nice to have it a the
client as well, but that 
matters nought until some presentation logic is sorted ie forms, windows, images etc.

I had a good laugh at your response re the CSS! Well, it is supposed to look like a DOS screen,
and the thinking at the 
time was that only fixed pitch fonts translated to character based forms! I think the CSS has
loads of crap left over 
from a previous project. Haven't been there in ages.

I'm still reworking some of the sql generation after discovering some shortcomings, and it was
getting too complicated. 
Hopefully should be done in the coming days. I want this finished! Been going on for too long.

Having said that, you be willing, am keen to try and work out why Windows is not playing nice.
Please feel free to 
emailing me results, etc and I'll try to point in the right direction.

As soon as the sql is done I'd like to take this further and put as much debug code in as necessary.

If you want to start a new branch on github pls feel free. But please don't splat master!
I'm hesitant to be running more than one dev branch for fear of stuffing it up, but I should
be able to copy your branch 
here and cooperate on it whilst still keeping my master and dev branch. I think.

If you stay away from the converter stuff we should have no issues merging later.

Thanks Larry,

- Bernie
On 19/4/18 1:50 pm, Larry DiGiovanni wrote:
> Bernie van't Hof wrote:
>> In my project I originally envisaged the bulk of the app running on the server, communicating
>> with some sort of simple glass teletype on the web client. Not very efficient perhaps,
but given
>>  the nature of the project I really didn't care.
> I took a deeper dive into your code this evening (I still can't get it to run in any useful
fashion, which is a barrier 
> to understanding), and I can't figure out the client event model, but it appears to be
driven to the keystroke.  Not 
> sure how mouse interactions are trapped and translated.  It appears that the actual form
rendering is not yet 
> implemented (at least as of my most recent pull in March).
> But I'm guessing that the HTML in #PXdata is intended to be well formed, CSS-driven DOM. 
If your event capturing 
> doesn't interfere/override the html events, then you could dynamically inject validation
as part of #PXdata. Was that 
> part of the strategy?
> You weren't kidding about the teletype approach as evidenced by this little gem in your
> * {
>     margin: 0;
>     padding: 0;
>     /*font-family:Verdana, Geneva, sans-serif;*/
>     /*font-family: Arial, Helvetica, sans-serif;*/
>     font-family: "Courier New", Courier, mono;
> }
> There's something you don't see every day.  :-)
> -- 
> Larry DiGiovanni

Copyright © 2004