[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: FreeAES



On Fri, 19 Jun 1998, Martin-Eric Racine wrote:

>  >I think you'll find that the reason XaAES uses pipes to communicate with
>  >the applications has nothing to do with memory protection.
> 
> Still, pipes are MEANT to be accessed by several apps, without causing
> a memory violation, so it's a neat trick.

I think you are confusing pipes with shared memory. Pipes are streams used
for interprocess-communication, shared memory is blocks of memory which
can be accessed by several apps without upsetting the memory-protection.

>  >The only proper way to implement non-modal alerts [... is ]
>  >if the AES would be extended to have a non-modal alert feature.
> 
> Precisely what I meant! Rewrite form_alert() to always be non-modal,
> within the AES itself. That way, older apps don't need to change any
> thing (with the added benefit other apps can keep on running and be
> interacted with), while newer ones can specifically request their 

It's not this simple. The problem with a non-blocking form_alert() (in the
sense that it blocks only the calling application), is that the calling
application is unable to receive messages and redraw windows while the
alert is open. Freedom works around this by removing windows that can't be
redrawn, it's not exactly elegant but it does work.

> alerts be of the non-blocking type to allow multi-tasking interaction
> within the application. We just need to agree on ONE method to specify
> that extension and ensure N.AES and MagiC will follow it.

MagiC already has nonmodal dialogs, which can easily be used for alerts as
well. I don't like the API though, but it's slowly becoming a standard so
it would make sense to support it.

I also think that the new AES-calls in N.AES should be supported
(appl_control(), extended appl_search()), as they're very useful and
essential for my 'Taskbar' ;-) In an ideal world the special
MagiC-messages to AESSYS should also be supported. These could be handled
by the same code that handles appl_control().


/*
** Jo Even Skarstein   http://www.stud.ntnu.no/~josk/
**
**   beer - maria mckee - atari falcon - babylon 5
*/