Version 6 misc

[Jeremy Collake]

ALPHA == DEVELOPMENT BRANCH

1. INCOMPLETE CODE BASE
2. INCOMPLETE FEATURES
3. NOT TESTED
4. ETC ....

In version 6, we have lots of cool new things. Let me go through A FEW them:

1. Much better multi-core support. Cool new graph(s) supporting any number of cores, as seen in ThreadRacer.
2. Spliced sleep support, so you can prevent the display or PC from sleeping.
3. Greatly expanded watchdog rules. Now you can do LOTS MORE. I mean, LOTS MORE. You can have CPU and I/O priorities raised or lowered, affinities changed, process terminations and restarts, etc.. all based on more criteria than ever before. The new rules dialog is a big part of version 6.
4. Greatly expanded services support.
5. More precise system user context resolution. While few really need this, the mechanism works better and no longer has TS/RDS as a pre-requisite. Basically, it resolves the process's security token to a user context, instead of its session ID to a user context.
6. Easier to navigate menus, and a generally more fluid product.
7. No longer requires TS/RDS service, since it uses the process security token to resolve user contexts instead of the session ID
8. Lots more .... its been in the works for so long, its hard to document them all. Some things aren't done, and I've not mentioned those.




UPDATE:

Download Links
Early ALPHA
NOT FINISHED
PARTS DYSFUNCTIONAL
PARTS INCOMPLETE
FOR TESTING ONLY
MUCH MORE NEEDS TO BE COMPLETED, ADDED
BUGS TO BE EXPECTED, NOT TO REPORT

REGISTERED USERS SEE SUPPORT AREA.

[Jeremy Collake]

I don't think I posted many screenshots on the Forum, so will post a few here and there as I work. Here is what v6 looks like, main window. Yes, it actually runs for me, which is why I got so thrown off on the initial release. I believe it was a configuration related issue/crash. I have fixed more and more. Should have a new build out again today.

Note the ThreadRacer style per-core usage display I developed will also be used for other metrics, as I mentioned, and overlaid on the graph too. For instance, per-core CPU frequency.

Service names *and* groups are included with process names. Right click on a process, and any services it hosts you can manage.

Rules dialog is awesomely improved, lots of possible actions, on different behaviors.

Username can be specified in the process names now too, with a colon ':' character, e.g. 'myprocess*.exe:Jeremy' ...

[Jeremy Collake]

Rules dialog .. impossible to show all options since they hide/show as necessary, but...

[Jeremy Collake]

Release came. Do not that I got lazy and did NOT upload new Server Edition builds, but did errantly update the database. That means users who downloaded the Server Edition were getting an OLD alpha. I am fixing this now. Now that I'm awake again, there will be more updates tonight that will bring things MUCH closer to beta. I anticipate beta by the end of the week.

[Jeremy Collake]

v6.0.0.8 ALPHA issued

[edkiefer]

Any major changes /additions to test in 4.0.0.8 as I don't see any changelog to go by. Maybe if there major changes post them .
Only noticeable issue which probably is really not one . after updating column settings are lost in GUI , I am assuming you have a reset UI flag on updater ?
Any performance on governor that should be tested . I have never tried getting a metric on it .
So far so good from what I see .

[Jeremy Collake]

At this time to keep track of changes would be another slow down, as there are a gazillion each build. They will start when I hit beta. I will try to note large functional improvements or changes though.

[edkiefer]

At this time to keep track of changes would be another slow down, as there are a gazillion each build. They will start when I hit beta. I will try to note large functional improvements or changes though.

sounds good, that is what I meant, I didn't want a list of bugfixes as that not really important to end user , just major changes or things that were added, enabled , improved in big way .
I surely don't want you to list every small error,bug etc , just things the end user could test , evaluate .

[Victek]

v6.0.0.8 ALPHA issued

I must be blind as I can't find this...? 

[gman68w]

Me neither. You say 'issued' like you've distributed it to everyone who's supposed to get it.

[Jeremy Collake]

I am sorry for the confusion. I created a beta test group a while back. So they got notified, but others haven't YET because we'll be at BETA in probably a week, so why have people run a half-functional build? Some languages (e.g. Russian) haven't been updated and will totally malfunction. The rules dialog is still torn apart, etc.. etc...

Anyway, you might want to log in to the Support area ;p. Do note that I don't want to see this alpha issued out to the general public yet, because so much remains unfinished.

I shouldn't have even mentioned it yet if I knew what was good for me, BUT I get as eager and excited as everyone else.

Since we'll be at beta in no time, I'm just trying to protect everyone from:

1. Premature judgement on the code base
2. Problems or issues with the code base
3. General displeasure
4. SO MANY things that are not done or not functional yet

But, if you want it, you now know where to find it.

[Jeremy Collake]

v6.0.0.9 ALPHA being issued in a few minutes. I *will* mention one change. It eliminates that annoying graph flicker of the new display

[Jeremy Collake]

v6.0.0.10 ALPHA is coming shortly, I fixed A LOT today, and finished a lot more ... but there is work I'm still doing. Many things were dysfunctional, as you surely noticed (or maybe not). We're getting there though.

[DeadHead]

Where do you want us to post issues with the alpha? I'm not sure which issues you're aware of or not.

Anyway, per-core activity don't work on my system. Also, can that window be toggled on/off?

[Jeremy Collake]

I anticipated per-core activity to not currently work on English systems. You are non-English? This was something I wanted to find out for sure before I made modifications designed to allow proper operation on all systems.

Other than specific questions I ask (like the above I asked somewhere at some point), don't report issues with the alpha - as I generally know about them, just haven't gotten to them. Once we get to beta, then start flooding with issues . The alpha probably should have been kept internal, BUT ... blah, this is how I do it I guess

[Jeremy Collake]

.. and the window will be toggle-able, when I'm done. As with the other windows.

[DeadHead]

/quote]

Roger that!

[Jeremy Collake]

Slightly smaller per-process utilization graphs within the listview. This allows for a border that shows selection state, and just looks better.

Subtle, but very definite, improvement. This one I'm backporting to v5.1.0.92.

[Jeremy Collake]

And whoever (if anyone) was waiting on Energy Saver Exclusions .. you now have them as of v6.1.0.12 alpha (now at v6.1.0.13 alpha).

[hanemach_gt]

Energy Saver exclusions are wanted by me, big plus to you for extending this wonderful feature.

[Jeremy Collake]

Energy Saver exclusions are wanted by me, big plus to you for extending this wonderful feature.

Good, just got done adding it also to the process context menu; under a new 'Excluded from ...' sub-menu to try to keep things organized. That will be in the next build. Hopefully it works fine, I have not reviewed or tested that new code yet.

To be clear, the current build already supports it, of course, but you have to use the configuration dialog found in the Energy Saver Dialog.

[Jeremy Collake]

v6.0.0.18 ALPHA will be issued in about an hour.

Progress this, and latest few, versions:

+ Rules dialog increasingly functional (needs more testing and review)
+ Core engine support for expanded rules worked out and better (needs more testing and review)
+ Energy Saver dialog spruced up with Apply button for current power profile and Energy Saver Exclusions
+ Energy Saver Exclusions got their own rule
+ Fixed a possible race condition also found in v5, though it is theoretical. I back-ported the change for the next v5 minor update, whenver that may be <-- might be the cause of startup crashes some saw in v5, before I implemented a work-around. When I re-enabled this code and was running through a debugger, I just happened to catch the race condition at its worst.
+ Let user set update check frequency
+ Fot per-core metrics displayed like I wanted, but there is a little flicker from time to time I'm working on
+ A million more things ....

[Jeremy Collake]

There is still a known crash in 2K/XP in the Active Processes Tab. I haven't got to even look at it yet, but will make it my next priority.

[Jeremy Collake]

Something many wanted ...

[Jeremy Collake]

v6.0.0.19 alpha, amongst other things, fixes a crash seen most often in XP while viewing the Active Processes tab, but it could have easily manifested on other machines - e.g. Windows 7, etc... Will be uploaded in about an hour - now available.

[edkiefer]

Something many wanted ...

What was the old "periodically check for updates " frequency rate ?

[Jeremy Collake]

What was the old "periodically check for updates " frequency rate ?

Two days. The alpha is accelerated in its frequency though.

[Tarnak]

There is still a known crash in 2K/XP in the Active Processes Tab. I haven't got to even look at it yet, but will make it my next priority.

Just tried v6.0.0.19, and it is fixed for me!

[edkiefer]

Two days. The alpha is accelerated in its frequency though.

Ok , The per core overlay graph been fixed in last 2 builds and seems I wasn't affected by XP active process crash , maybe because i didn't have enough processes running .

[Jeremy Collake]

Ok , The per core overlay graph been fixed in last 2 builds and seems I wasn't affected by XP active process crash , maybe because i didn't have enough processes running .

Glad the graph is getting better, but it is not perfect yet. It is close to what I want it to behave like.

On the crash .. it depended upon the value of uninitialized virtual memory. People affected were just unlucky and/or pseudo-randomly affected.

[Jeremy Collake]

Just tried v6.0.0.19, and it is fixed for me!

Good, yes it was also fixed here. However, that may not be the only crash in this early alpha, so let us not jump for joy yet

[Tarnak]

[.46]Addition.GUI: Clicking the graph will rotate through available application color schemes (currently only white and black) See -

The above change introduced in V5.1.0.46 is not functioning in v6.0.0.19 with XP.

[edkiefer]

[.46]Addition.GUI: Clicking the graph will rotate through available application color schemes (currently only white and black) See -

The above change introduced in V5.1.0.46 is not functioning in v6.0.0.19 with XP.

confirmed not working in 0.19 but was working in last few .

[Jeremy Collake]

Yea, I know, thanks though. I disabled it. It will be fixed up. The rotation didn't work well with the new components yet.

[Tarnak]

Yea, I know, thanks though. I disabled it. It will be fixed up. The rotation didn't work well with the new components yet.

At the moment I get the new 'Core graph', blinking in and out.

[Jeremy Collake]

At the moment I get the new 'Core graph', blinking in and out.

Thanks. Yea, I know. At this stage of development, there's no need to even report things like that, as I see them too.

[Jeremy Collake]

v6.0.0.20 alpha will be issued within the next hour to fix up the per-core utilization graph when the RAM Load is not visible. The white/black theme toggle I'll have back real soon, don't worry.

[Jeremy Collake]

Process Lasso 6.0.0.21 alpha

* Restores white/black theme change on click
* Restores click on RAM Load to enable/disable
* Finishes affinity settings for watchdog rules, they are now functional - including the undocumented "random" affinity which will select 1/2 of available cores
* Disables some unused code
* Etc.... Etc....

Note you do have to click the actual graph now to change themes. In v5, any extraneous click could change themes, something that was both good and bad. I left it as an easy way to accidentally introduce the user to these types of theme changes, but v6 locks it down to the graph region only.

Some users who have seen crashes with the alpha may still see them. I will be nixing these soon, but wanted to complete the functionality first. Keeping the GUI minimized to the system tray would mitigate any known crash I've heard of with the alpha. Most users should have no problems, except for missing or incomplete features here and there. In my haste, I surely made a mistake or two though, and those will be quickly discovered in the debugging stage.

32-bit: http://bitsum.com/files/alpha/prolasso.zip
64-bit: http://bitsum.com/files/alpha/prolasso64.zip

[GTX2GvO]

Note you do have to click the actual graph now to change themes. In v5, any extraneous click could change themes, something that was both good and bad. I left it as an easy way to accidentally introduce the user to these types of theme changes, but v6 locks it down to the graph region only.

I really like this change. Seeing the graph theme flip on any UI click was(/is) rather annoying.
Is there also going to come a user setting that "locks" it to either the dark or light theme?
I prefer the dark theme over the light one. 

Still on v5.x here. Don't know If I should start using a alpha/beta build. 

[dougaf4t]

Very minor point: Using 6.0.0.21 x64 alpha with tray icon set to show total processor utilization and self-restart set to once per day.  When it does restart, it causes the GUI to open ...

Is this just me?  I don't see anything in the config file that would seem to apply.  PL self-restarting minimized to tray would be nice, if not as a default at least as an option.  BUT if this is any more than very very very trivial, please shelve it till more important stuff is done.

[Jeremy Collake]

I have not tested the self-restart mechanism, it may be failing. There should be no need to restart itself, as no memory leaks, handle leaks, or other issues requiring a restart have been so far detected even in this alpha version.

Regardless, I will take a look at it, though I just uploaded .22, so it may be a build or two. Going through my huge list of chores one by one.

[edkiefer]

restart seems to work . Minimize GUI . then in taskmanager close processgovernor.exe , after about 20-30sec or so processgovernor restarts with a warning that core is off (red stoplight icon graphics) then a few seconds and its on .
The GUI never popped up, stayed minimized so in this very basic test seems to work ok .

PS: you can also do it with right click on PL tray icon and "stop ProcessLasso core engine" , same results as above

Tested on 6.0.0.22

[Jeremy Collake]

He is speaking about self-restart on an interval, which restarts the GUI and governor every X hours. That code is a bit separate, though does share some code paths, but is likely malfunctioning.

[Jeremy Collake]

Also note that I've got some more builds I'm publishing this weekend, have it planned out pretty well. You will notice I have perhaps prematurely listed the ALPHA as available for download (though just now fixed the links from beta to alpha).

Once you switch to the ALPHA channel, the updater WILL force ALPHA updates.

If going back to version 5:

+ No problems should be noticed except with 'Do not sleep' processes, which are now subdivided into PC vs Display
+ Saved Window Positions, including Child Windows, will be lost
+ Saved Process Listview column sizes and ordering will be lost

[edkiefer]

He is speaking about self-restart on an interval, which restarts the GUI and governor every X hours. That code is a bit separate, though does share some code paths, but is likely malfunctioning.

Ok, I see now, I missed that option (self restart PL interval ). I have to ask though why would you ever need to restart the GUI , I can see processgovernor having some kind of issue but not GUI .

[Jeremy Collake]

Ok, I see now, I missed that option (self restart PL interval ). I have to ask though why would you ever need to restart the GUI , I can see processgovernor having some kind of issue but not GUI .

Indeed, there is no good reason. I put it there as a safety, in case someone, in some distant universe, had an issue where a third-party DLL injected into Process Lasso was leaking memory or handles. Of course, with newer self-integrity checks this is even less important, as Process Lasso monitors itself to make sure nothing looks out of whack. If it does seem out of whack, it will restart with a message. Nobody has seen this occur I'm sure, except me while testing the mechanism.

[Jeremy Collake]

Oh, and I also have a BIG new document introducing all the internal features of Process Lasso v6 ... It is an important upgrade to the docs anyway. It is not done yet, not by a long shot, but I will start publicizing parts of it as I complete it. Feature by feature.

I've added a lot of new sanity checking in v6.0.0.23 (yet to be released) to help track down remaining aberrations, which I haven't spent much time on yet as I focus on features and functionality.

[edkiefer]

Oh, and I also have a BIG new document introducing all the internal features of Process Lasso v6 ... It is an important upgrade to the docs anyway. It is not done yet, not by a long shot, but I will start publicizing parts of it as I complete it. Feature by feature.

I've added a lot of new sanity checking in v6.0.0.23 (yet to be released) to help track down remaining aberrations, which I haven't spent much time on yet as I focus on features and functionality.

Soods good , knowing how new features work is important IMO . great work .

[dougaf4t]

Indeed, there is no good reason. I put it there as a safety, in case someone, in some distant universe, had an issue where a third-party DLL injected into Process Lasso was leaking memory or handles. Of course, with newer self-integrity checks this is even less important, as Process Lasso monitors itself to make sure nothing looks out of whack. If it does seem out of whack, it will restart with a message. Nobody has seen this occur I'm sure, except me while testing the mechanism.

In that case, I can turn off self-restart  .... I had turned it on sometime about two systems & four years ago (a mighty AMD K6-2 300MHz, 256MB RAM, XP SP3; what a long strange trip it's been) and somehow never changed that one option.

[Jeremy Collake]

Indeed, I may just remove the option - one less piece of code to have to maintain.