[openbeos] Re: Init scripts
- From: "Balint, Jess" <JBalint@xxxxxxxxxxx>
- To: "'openbeos@xxxxxxxxxxxxx'" <openbeos@xxxxxxxxxxxxx>
- Date: Tue, 14 May 2002 15:58:54 -0400
That's a mighty fine idea, but for existing programs, and to keep OBOS
compatible with R5, the Flash installer would not work correctly.
-----Original Message-----
From: François Revol [mailto:revol@xxxxxxx]
Sent: Tuesday, May 14, 2002 3:44 PM
To: openbeos@xxxxxxxxxxxxx
Subject: [openbeos] Re: Init scripts
I'm more in the SysV way, with some folders, and eah scripts in this folder
gets executed, the order depending on the begining of the name.
This way if a proggy wants to hook in, just put another script in this
folder, and it just has to check if the script exists to know it was
already installed.
So when in SAFEMODE, the sysinit script would ignore them.
And this one would never change, taking down all risk of damaging
it by a faulty prog install.
En réponse à Fred K Ollinger <follinge@xxxxxxxxxxxxx>:
> > This sounds reasonable. One that I can think of off-hand if the
> Macromedia
> > Flash plug-in for BeOS. There is a required plug-in server that is put
> in
> > the startup scripts if requested. If the scripts are re-designed, it
> needs
> > to be that if Joe Web-Surfer tries to install this in the start-up
> scripts
> > that it won't break anything. I'm not sure how to go about doing
> anything
> > about this, but it is a thought.
>
> Thanks for being specific. Makes things easier.
>
> One way to do it is to take script, then install macromedia. Then do a
> diff on old init and new init. Now we know what changed. Now find what
> macromedia looks for when it installs itself and keep that there, but
> change other parts to make things more compatible.
>
> In fact, things can change radically, and the macromedia player might
> still work if this is thought of ahead of time. Anyone know anything
> else
> that would break if scripts were "improved"? I'm guessing that since
> newos
> is going to be the kernel, then things are going to have to change
> anyway
> even if the krusty, hard-to-configure script model is used.
>
> I could be wrong about this.
>
> Fred
>
>
>
Other related posts:
