Re: Fluid Application Environment 3.0.alpha1 Released!

new topic     » goto parent     » topic index » view thread      » older message » newer message
eukat said...
jimcbrown said...
eukat said...
jimcbrown said...
eukat said...

certified to not crash as if it would if hacked in by the random user who does not understand FAE so well.

I don't think this is possible either. function_to_do() could simply call crash("some msg") and crash that way. I think what you want is integrity checking inside of FAE itself (akin to what a kernel does at its syscall interface when being called by user-space code), to reject garbage if garbage is passed in. I agree that this would be a good idea, but I must reluctantly acknowledge that it's difficult to catch all cases - and even more difficult to cerify in any way that has some sort of reliable meaning.

Ok, then the certified code that won't crash in FAE's widget

if sequence(dir elephant_3legs.e) then include elephant_3legs.e end if 
 
procedure do_user(sequence what) 
  elephant_3legs:gui_interface(what)-- must exist in my elephant_3legs.e 
end procedure 

Well, I just tried that out (commenting out the one line in do_user since I don't have a file called elephant_3legs.e) and it crashed.

Ok, make it so it doesn't crash? If elephant_3legs.e does exist, and i don't sequence(dir elephant_3legs.e) then why would it crash?

In that case, it wouldn't .. but then why present that line in the first place?

You seem to be asking for FluidAE to also be a preprocessor. There's nothing wrong with this, though this is typically outside the remit of a GUI toolkit. Still, I know that ryanj has plans to make FluidAE special, and maybe this is just one more method to make it stand out.

new topic     » goto parent     » topic index » view thread      » older message » newer message

Search



Quick Links

User menu

Not signed in.

Misc Menu