[FIXED] ProBalance inoperable with 3.80

[watermelon]

Process Lasso doesn't lower priorities anymore. I am not sure if this came with 3.80 or was also present in the beta releases. Manually setting the priorities as well as custom default priorities are the only things still set.

[MickeyG]

I appear to have a similar problem but can't tell what is happening since the display shows nothing.  I am running Vista on a tri-core system.

[Jeremy Collake]

Blank window problem
If the display window is blank, use the 'View / Reset Window layout' menu option for a quick fix, or update to v3.80.2. This was a minor bug in v3.80 caused by the new pane positioning code, user manipulation of the pane sizes, hiding/showing of the graph, and main window resizing. Basically, the pane positions could simply get out of whack under certain conditions. It was rare, but happened.

Reported ProBalance problem
In v3.80 and the first build of v3.80.1, if (and only if) the user manually tweaked the 'time before restraint' ProBalance parameter to less than 1/2 the refresh speed, then ProBalance would fail to act. At the default refresh speed, that means a value lower than 499ms could cause a malfunction. I have since updated v3.80.1 to fix this issue. It did not apply to the older v3.70 release., and would not apply to anyone simply using the default ProBalance parameters. Therefore, to fix it without updating, simply reset your ProBalance parameters to their defaults.

[watermelon]

The problem definitely isn't any exclusion settings I would be aware of. What you have suggest, trying the CPU Eater, was indeed the second thing after realizing the ProBalance issue.

Other than upgrading Process Lasso I did not have any changes at all on its configuration or otherwise on my system. Well ..except for the recent IE8 hotfix.

I will spare you posting the default configuration, ProBalance doesn't work with a clean install either. That's also about all I can think for possible causes. If you could point me to an older version of Process Lasso I could check if the problem persists.


Hope to see the portable version soon.

[Jeremy Collake]

This report bothered me all day (when I first read about it) .

I have discovered a scenario where this would occur if the user lowered the 'time before restraint' ProBalance parameter to less than 1/2 of the refresh speed. In the new code, this would cause a malfunction under certain circumstances, causing ProBalance to not act.

Does this apply to you? What is your 'time before restraint' ProBalance setting? If you slowed the refresh speed to 2 seconds, then the 'time before restraint' could cause problems when it goes below 1000ms (1 second).

Thanks for your feedback. I would bet this fixes what you see.. as its the only applicable code that was modified in v3.80. I know you said you tried the default config, but I'm hoping that test was an anomaly.. as this is the only possibility I've seen. I've ran ProBalance through more tests than I can count, and have been walking through the code checking for any other theoretical possibilities. I've seen no other issues.

...And the portable edition *IS* coming quite soon.. it is on my todo list, and has been planned upon the release of of v3.80. I've been busy doing language updates and such, but will get to it very soon.

UPDATE: I've now updated v3.80.1 with a fix the few users affected by this bug (maybe 0.1%). So, re-download v3.80.1 and the issue will be solved. This bug was only present in v3.80 and v3.80.1.

[watermelon]

It indeed was the bug you described. Thanks for providing both a solution and fix. 



One minor suggestions for the localized versions of your website. I'd probably be better if the language detection would respect the headers of the clients browser, instead of using geolocating to redirect people.

[Jeremy Collake]

Thanks for reporting back to me that it worked . I am glad you reported the problem. It was a very minor mistake I made in one of the last betas. A minor change to a line of code that resulted in an expression being treated as unsigned of signed .. sigh.. At least not many people were affected by it. It is something I should have caught though, in my harsh opinion of myself, lol.

I now wish I would have released the minor update to v3.80.1 as v3.80.2, then pushed it out. I promised a bunch of little odds and ends in v3.80.2 to various people, and all this new code is taking me time to test and review. I will get it on out tonight though..

The new features include an expanded system tray tooltip with more info, but only for English users at present. I had to write the code so that it was optimal, since this tooltip is refreshed often. It was this ensuring that the code is optimal that took the most time. If only I wasn't such a freak about performance, but I don't want to bloat my code. Any wasted CPU cycle is doing more harm than good, from my perspective. So, I try not to waste any...

Then there is the new indication of hard-coded exclusions in the GUI. That change just took a minute. I've also made a few other tweaks. Truthfully, this should be v3.82 instead of v3.80.2.. but, it is what it is.

Thanks for the tip on the geolocation. I will look into tweaking the geolocator code soon. It is needing some maintenance anyway. I need to store a cookie so that people who opt-out of it aren't presented with it again until the cookie expires or is deleted.

Thanks a bunch