On Sunday 11 Sep 2011 Duncan wrote: > Anne Wilson posted on Sun, 11 Sep 2011 18:35:32 +0100 as excerpted: > > I take your point, but in truth, any well-written piece of code should > > be able to exit gracefully if something fails. If it can't, I'd rather > > not run it at all. > > You're absolutely correct. But the point is, one of the big selling- > point features of plasma has been its extensibility... by coders who may > in fact be rather bad at it, beginners or whatever. In that sort of > environment, you have to COUNT on some of the code being bad, because it > is GOING to happen. An app designed for that can't simply crash when one > of the extensions goes bad, because it WILL heppen, and it WILL look bad > for the main app as a result. It MUST be designed to be robust and have > the rest keep running in spite of whatever badly coded barf-code someone > throws at it, or it wasn't so properly designed for that extensibility as > it might seem to be after all. > > Which is the problem we have. > > So are you going to quit running plasma, then, because you'd "rather not > run it at all"? [Haha only serious.] > > [Caveats about not being a (C/C++) coder apply. My skills tend more > toward technical sysadmin, scripting, etc. And /as/ a sysadmin that does > at least speak some programming lingo, plasma is quite good in a lot of > ways including its extensibility, but it could certainly be better in > regard to robustness in view of that extensibility, is what I'm saying.] By and large, when something goes wrong with plasma, it drops out, then restarts, so you could say this is a real effort to deal with the problem. Having no coding skills whatsoever, I can only guess at what could make a plasmoid freeze everything. I imagine that plasmoids that get official blessing have had some code eye-over, but as you say, plasmoids are relatively easy to write and distribute, so anything goes, I guess. You can't safeguard against everything. What, to me, is more important, is that the author of that plasmoid should be told what is happening. Most applications have a contact email address for the author - I don't know whether it is true of any plasmoids, but it's certainly not true of all of them. It should be. Anne
Attachment:
signature.asc
Description: This is a digitally signed message part.
___________________________________________________ This message is from the kde mailing list. Account management: https://mail.kde.org/mailman/listinfo/kde. Archives: http://lists.kde.org/. More info: http://www.kde.org/faq.html.
