LabVIEW
Windows occasionally stop redrawing
@TurboPhil wrote:
Yes, it appears that disabling Aero is a viable workaround. It's a bit of a bummer to have to disable that feature, since there are elements of Aero that are actually somewhat useful (window previews, etc.), but we can live with it for now.
I strongly support your investigation to try to solve the fundamental problem.
Not only is it a "bit of a bummer," but what do you say to a customer? "By the way, the app only works if you disable Windows Aero?" It's almost like saying, "By the way, the app only works in Windows XP."
On the other hand, this issue is apparently a problem at the deepst levels of LV, otherwise it would've been fixed already. To that, I give NI a sincere "good luck" and hope they can resolve the issue without a major rewrite.
I personally wouldn't get hung up about Aero having to be disabled. Especially for control applications!
It's another layer of user interface complexity that can at best slow down the machine, at worst introduce all kinds of issues, not only with LabVIEW itself, and definitely increases the number of code paths by a significant amount, to make the chances for hitting bugs in video drivers significantly bigger.
For mission critical systems it's always a very good idea to not use all the gimmicks that come with a new OS, but go with as few features as possible.
@rolfk wrote:
I personally wouldn't get hung up about Aero having to be disabled. Especially for control applications!
It's another layer of user interface complexity that can at best slow down the machine, at worst introduce all kinds of issues, not only with LabVIEW itself, and definitely increases the number of code paths by a significant amount, to make the chances for hitting bugs in video drivers significantly bigger.
For mission critical systems it's always a very good idea to not use all the gimmicks that come with a new OS, but go with as few features as possible.
From a technical perspective, I agree. But if this is to be deplyed on a customer's computer, they may not be so understanding. If the customer doesn't like it, then you're wrong. Period. Laying the explanation above on a customer usually sounds like you're making lame excuses for sub-par work. A savvy customer would understand and agree, but all too often, some VP is in charge of a project and is going to complain about having an "ugly" looking computer interface. You and I know that a high end computer stripped of all unnecessary crud is a developer's dream, but sometimes it's impossible to achieve the dream. 😉
One reason that it might be not in the list of known bugs might be that it is difficult to reproduce since it depends on a certain hardware/driver configuration. It is very possible that this is really a bug in the video driver and only exhibits itself with certain video controllers and according drivers. Bugs have to be confirmed by NI people to be bugs before they can go into the bug database and an engineer needs to look at it before it makes it into the list of known bugs. If they would add every "bug" that is reported by users to that list, the list would be much longer than you ever would care to read through and would be totally useless.
