IDE Interference (not just a another rant)

Brian Yennie briany at qldlearning.com
Thu May 27 11:51:54 EDT 2004


Xavier,

> See my previous message response with the suspendstack
> script. It's not a problem per say, true but it is a problem
> when implemented where the RR events cause a serious slow down
> each time you edit an object feature via a Revpalette and a
> suspendstack message is sent to your stack...


If you're gonna flame Rev for being slow, you should really at least 
consider that your own added script is at the crux of the issue...

Is there any reason you can't change your auto-save implementation? 
Saving stacks in Rev is quite fast, but you're probably causing havoc 
by doing it on suspendstack- unless you can provide some evidence to 
the contrary. I save 50MB stacks all the time and it's no more than a 
blip on a 500Mhz iBook.

How about using a simple timed auto-save, say every 60 seconds? What's 
the use in saving on every suspendStack when you have a screen full of 
windows and palettes?

If you really want the tightest, instant auto-saving you'll probably 
have to insert a slew of event and setprop handlers in a frontScript, 
but frankly I don't think it's worth having...

> Should the suspendstack be sent to the stack each time you edit an
> object causing a save? One missing feature (no blame anywhere) is
> the function stackwaschangedsincelastsave()=true|false... The other
> is that the suspendstack should be ignored while in the GUI design
> mode of RR...

Yes, of course it should be sent. That's how the message is defined. 
Nobody said to put a save handler in there!

- Brian



More information about the use-livecode mailing list