Author Topic: Analyze Problem 2.21  (Read 11263 times)

Offline labtech

  • Newbie
  • *
  • Posts: 2
    • View Profile
Analyze Problem 2.21
« on: August 16, 2005, 02:47:58 pm »
When I try to use the analyze function, and begin to analyze befor install image, the analyze program hangs.  I finally kiiled the process in task manager  after 1000 seconds!  :'( I guess this doesn't work on my PC, I haven't tried it on a different one yet.

Here is the PC specs.

HP/Compaq D530
2.2Ghz P4 - 512mb ram
Integrated Intel Extreme Graphics

Any Ideas?

Offline labtech

  • Newbie
  • *
  • Posts: 2
    • View Profile
Re: Analyze Problem 2.21
« Reply #1 on: August 16, 2005, 02:57:13 pm »
When I try to use the analyze function, and begin to analyze befor install image, the analyze program hangs.  I finally kiiled the process in task manager  after 1000 seconds!  :'( I guess this doesn't work on my PC, I haven't tried it on a different one yet.

Here is the PC specs.

HP/Compaq D530
2.2Ghz P4 - 512mb ram
Integrated Intel Extreme Graphics

Any Ideas?


Oh, sorry, I forgot to mention the OS;  Windows XP Pro. SP2


Offline Admin

  • Administrator
  • Hero Member
  • *****
  • Posts: 522
    • View Profile
    • ZSoft Software
Re: Analyze Problem 2.21
« Reply #2 on: August 16, 2005, 05:25:45 pm »
Have you only tried (and failed) once, or have it happened several times?

You say that the "Analyse program hangs" - does it stop responding or is the form repainted if you minimize it and bring it back up?

When you killed it, did you see how many threads (in the job-list under "Processes") the program had?

Also - you shouldn't have to wait 1000 seconds... The form should update with the newest data - if it haven't found any files or registry entries within a couple of seconds (wait 10 to be sure) I guess it doesn't work...

Taking that the 1000 seconds was read from the form (so it wasn't frozen) - didn't it start "counting" the files and/or registry entries found? What exactly happened?
« Last Edit: August 16, 2005, 05:31:31 pm by Admin »

Wolfi

  • Guest
Re: Analyze Problem 2.21
« Reply #3 on: February 07, 2006, 11:18:07 am »
I have the same fault in version 2.3.  The analyzer (bevor install) will not finish.  The timer is still running after finishing counting files and registrations.  Usage of the buttons is not possible. Exit is only possible with the task manager.

OS: Windows XP home SP2.

I hope you can help.

Offline Admin

  • Administrator
  • Hero Member
  • *****
  • Posts: 522
    • View Profile
    • ZSoft Software
Re: Analyze Problem 2.21
« Reply #4 on: February 07, 2006, 03:17:50 pm »
Does it happen every time, or have it just happened once?

When you killed it, did you see how many threads (in the job-list under "Processes") the program had?

Offline idlerwheel

  • Newbie
  • *
  • Posts: 7
    • View Profile
Re: Analyze Problem 2.21
« Reply #5 on: February 23, 2006, 08:36:15 am »
This has happened to me twice, once in the current release version; once tonight in the latest debug version. The error occurs during the second of the two analyze stages. The system hung tonight for two hours with the "compare" statement flashing. I had installed System Mechanic 5. The System Mechanic installation does not incorporate a reboot. All menu selections in the analyze window were dimmed. I broke out of the routine by hitting the "X" at the top right cornier of the window.

The earlier time when this happened, I broke out with CTRL+ALT+DEL. I don't recall the processes that were running.? I rebooted the computer and tried to pick up where I left off, with the "continue an 'old' anlysis. The routine hung in this second attempt, too.

I rebooted the computer and returned to the "Continue an 'old' analysis" routine. The process hung again, after counting the files and then allowing me to enter a name for the installation. The name that I gave the installation contained illegal DOS symbols (/ and spaces). The hang is only during the "comparing" phase, with the blinking red statement on the screen. Each time tonight, the timer froze at 29 seconds. It's been hanging for a half hour now.

It's hung right now. 24 processes are showing in Task Manager. I don't know which ones are associated with your program. CPU usage is low -- 2-6%.? Strong resident utilities that are currently loaded are Sygate Firewall, Grisoft AVG, and Nero InCD. I know that Sygate runs a few processes very frequently -- apparently in rotation. I haven't looked for this activity with AVG.

To me, this problem is serious.

I suggest changing the English name of this second analysis function to "Finish an analysis that has already been started,"  or "Continue an analysis that hasn't been finished."
« Last Edit: February 23, 2006, 08:41:18 am by idlerwheel »

Offline Admin

  • Administrator
  • Hero Member
  • *****
  • Posts: 522
    • View Profile
    • ZSoft Software
Re: Analyze Problem 2.21
« Reply #6 on: February 23, 2006, 04:12:35 pm »
So to sum up what idlerwheel said:

It never crashes in the first phase.
It sometimes crashes in the second phase.
It sometimes crashes when comparing.

Is this correct?

And what I would like to know is not how many processes there was running, but how many threads Uninstaller.exe ran. I didn't explain this properly before. To see this you have to open the "job manager", go to "view" => "Choose columns" and check "number of threads". Then go to the "processes" tab, find Uninstaller.exe end look in the "Thread" column.

I don't know if names is 100% correct, my version of windows isn't English, but is probably something like that.

-----

I will go over the code again.
« Last Edit: February 23, 2006, 05:11:58 pm by Admin »

Offline idlerwheel

  • Newbie
  • *
  • Posts: 7
    • View Profile
Re: Analyze Problem 2.21
« Reply #7 on: February 23, 2006, 10:52:36 pm »
Jens,

Your summary of what I said is almost correct. As I recall, the only time the program has crashed for me is during the compare mode.

One thing that I really appreciate about your uninstall program is the distinction between the first and second phases. With other installation trackers that I've used, there's been too much automation. The totally-automated large process will work fine with some installations, but will go bonkers when faced with an installation routine that incorporates a reboot. At that point, the user is confused -- typically faced with simultaneous messagess from both the installer and the tracker -- and unable to make a decision as to what to do; and this is before the actual reboot. One severe problem with most error messages is that they don't inform the user what the consequences of decisions will be, so the user is forced to act blindly. At this point, the entire purpose of the uninstaller goes out the window because there's a good chance that the procedure won't have been carried out properly; and this will be revealed when the time comes to uninstall the program.

Your manual activaton of the first and second stages is a good solution, and brings some sanity to this. A discussion in the documentation will be helpful for this. As I see it, your program separates the tracking operation into three phases: first anallysis, second analysis, and comparison.

There are obviously differences in the methods by which different installation routines operate, and different challenges presented to an uninstaller designer. A truly brilliant uninstaller would anticipate the different ways in which instllations differ, and cover the bases in advance -- the user needs to know what to expect ahead of time, in order to head off those unpleasant surprises. (Still with me?)

What would be especially useful, and probably especailly difficult, is the ability to track and uninstall, _perfectly_, humungous deeply-embedded system-wreckers such as Real Player and Quick Time, and to a lesser degree, Roxio.

To me, a perfect uninstaller would also incorporate a _safe_ registry editing function or reporting funciton  showing the exact changes. In fact, I'd also include this in the install tracking stage.

Now to your question about processes: I'll try to do this the next time it crashes. But, of course, I hope it doesn't crash because I'm reallly trying to use the program! Note that I'm not a programmer and don't think like one: I'm a usability and interface person, and an explainer.

There is one other place in which I feel that  the program has failed me, and that's when the uninstallation isn't complete. The documentation says that if the uninstallation is incomplete, the program will attempt to complete the uninstall process each time that the program is run. However, what happened was that a piece of the original remained no matter how many times I reran Uninstaller. I believe that this is a registry issue. Here, the user needs to know what to do. The success of a program like this one is largely a matter of its documentaton -- and user documentation does not come naturally to most coders.

Here, the operation is stuck. I don't know the reason. The user needs to know what to do when this happens, and be presented with a list of the registry entries that won't come out -- and some guidance as to what to do about it. And a _safe_ way to edit these entries manually. If you don't want to provide this edit function, perhaps you could direct the user to an existing free (or paid) program that is manageable. As some of us know, Microsoft's Regedit may be the most sadistic utility ever devised this side of computer viruses; people need to be steered away from it, and especially, toward well-crafted alternatives.

Ricchard Steinfeld

Offline Admin

  • Administrator
  • Hero Member
  • *****
  • Posts: 522
    • View Profile
    • ZSoft Software
Re: Analyze Problem 2.21
« Reply #8 on: February 24, 2006, 05:02:06 pm »
As I recall, the only time the program has crashed for me is during the compare mode.

Well I've found a bug in the compare function when dealing with "\" "/" "*" "?" and other illegal windows-characters. I hope that was it :)

A truly brilliant uninstaller would anticipate the different ways in which instllations differ, and cover the bases in advance -- the user needs to know what to expect ahead of time, in order to head off those unpleasant surprises. (Still with me?)

Nope ::)

To me, a perfect uninstaller would also incorporate a _safe_ registry editing function or reporting funciton  showing the exact changes. In fact, I'd also include this in the install tracking stage.

Please elaborate :)

There is one other place in which I feel that  the program has failed me, and that's when the uninstallation isn't complete. The documentation says that if the uninstallation is incomplete, the program will attempt to complete the uninstall process each time that the program is run. However, what happened was that a piece of the original remained no matter how many times I reran Uninstaller.

I'll add a function showing a list of files and registry entries that couldn't be deleted so the user can do it him/herself. That was what you meant, right?

The success of a program like this one is largely a matter of its documentaton -- and user documentation does not come naturally to most coders.

Got that right ::)

Here, the operation is stuck. I don't know the reason. The user needs to know what to do when this happens, and be presented with a list of the registry entries that won't come out -- and some guidance as to what to do about it.

? :-[

And a _safe_ way to edit these entries manually. If you don't want to provide this edit function, perhaps you could direct the user to an existing free (or paid) program that is manageable. As some of us know, Microsoft's Regedit may be the most sadistic utility ever devised this side of computer viruses; people need to be steered away from it, and especially, toward well-crafted alternatives.

I for one actually like regedit :)

How would you want a "manageable" tool to work?

Offline idlerwheel

  • Newbie
  • *
  • Posts: 7
    • View Profile
Re: Analyze Problem 2.21
« Reply #9 on: February 25, 2006, 02:16:36 am »
R: As I recall, the only time the program has crashed for me is during the compare mode.

J: Well I've found a bug in the compare function when dealing with "\" "/" "*" "?" and other illegal windows-characters. I hope that was it

R today: Great! Now, I'm looking forward to the next debug!!!

R: To me, a perfect uninstaller would also incorporate a _safe_ registry editing function or reporting funciton  showing the exact changes. In fact, I'd also include this in the install tracking stage.

J: Please elaborate

R Today: I realize that I'm asking for something that may require a lot more coding than I may realize. What I'm envisioning is a "before" and "after" comparitor, with the changes highlighted, perhaps in a different color. Let's say, the original keys are in green, and the new ones are in red. This would be a display only, but an "advanced" mode would be offered for those inclined to edit the entries. This might be especially useful in those nasty cases in which something refuses to come out.

R: There is one other place in which I feel that  the program has failed me, and that's when the uninstallation isn't complete. The documentation says that if the uninstallation is incomplete, the program will attempt to complete the uninstall process each time that the program is run. However, what happened was that a piece of the original remained no matter how many times I reran Uninstaller.

J: I'll add a function showing a list of files and registry entries that couldn't be deleted so the user can do it him/herself. That was what you meant, right?

R today: Yes; that's it! There's nothing worse than a routine that says, "The operation got messed up," without providing any clues as to what went wrong. I'm talking from experience with a few different uninstallers. Now, if the program could provide ability to jump to an edit mode at that point, the app would begin to become a killer. You're already thinking beyond other designers with your basic concept, I think.

Oh - before I forget: Please add to the list of tracked installations, the date and time of the analysis. This way, anyone who wants to do a reverse-sequence uninstallation would be able to determine the order. We had a discussion about this in the freeware newsgroup recently, and the consensus was that reverse-order uninstallation was the preferred way to do this. I read your reply in this forum, and I get it -- but I have this gnawing feeling that there may be some applications that have bad manners and overwrite things. I admit that I don't have the experience to back up what I'm saying; it's a feeling that comes from using maintenance utilities and somehow messing up the system despite being cautious.

R: Here, the operation is stuck. I don't know the reason. The user needs to know what to do when this happens, and be presented with a list of the registry entries that won't come out -- and some guidance as to what to do about it.

J: ?

R Today: I think we've covered this above. What I'm getting at is simply that when the routine produces unexpected results, the user needs to be presented with a clear path to solve the new problem.

R: And a _safe_ way to edit these entries manually. If you don't want to provide this edit function, perhaps you could direct the user to an existing free (or paid) program that is manageable. As some of us know, Microsoft's Regedit may be the most sadistic utility ever devised this side of computer viruses; people need to be steered away from it, and especially, toward well-crafted alternatives.

J: I for one actually like regedit

How would you want a "manageable" tool to work?

R Today:
OK. Here's my issue with Regedit. You sound like you're experienced enough to use Regedit wtih confidence. When we know what our goals are with this program, we can work with it confidently. It handles pretty well. Of course, we should back up the registry first!!!!!

I said that this utility is sadistic, and I mean this. Regedit, as I remember it, it's incredilby devoid of fail-safes. It's almost as if it was designed deliberately to enable people to do great permanent damage to their Windows.
- There's no backup/restore provision
- All changes are permanent because there's no "save" or "abandon" options. Whatever you change will be saved when you exit and that's that.

And Microsoft keeps this hostile design from one version of Windows to the next, despite a history of writers criticizing this method of working and warning their readers about the utility. That's what I meant. I can think of almost no situation in which safety steps are more important than right here. And, I'd add, that any registry editing functions that are incorporated, let's say, in an enlightened uninstall program (if they are), should have safety stages built-in.

I hope that all this is clear.

Richard

Offline Admin

  • Administrator
  • Hero Member
  • *****
  • Posts: 522
    • View Profile
    • ZSoft Software
Re: Analyze Problem 2.21
« Reply #10 on: February 25, 2006, 12:06:37 pm »
So a "Rededit" that makes backups of whatever the user changes..?

Offline idlerwheel

  • Newbie
  • *
  • Posts: 7
    • View Profile
Re: Analyze Problem 2.21
« Reply #11 on: February 26, 2006, 05:06:14 pm »
That would be an interesting feature of an ideal registry editor. I've been trying diffferent registry editor/maintenance programs, and each one I've tried lacks at least one essential ability, or works in a way that's user-hostile.

Richard

Offline Admin

  • Administrator
  • Hero Member
  • *****
  • Posts: 522
    • View Profile
    • ZSoft Software
Re: Analyze Problem 2.21
« Reply #12 on: February 26, 2006, 06:46:36 pm »
Okay, I've added it to my todo-list, but as it is quite a "biggie", don't expect anything soon :-[