Reproduceable Bug that crashes IDE

Questions and discussion about the ZBasic IDE.
pjc30943
Posts: 220
Joined: 01 December 2005, 18:45 PM

Post by pjc30943 »

EDIT: The last few days the IDE crashes about every third download, hanging the system. Other modes of crashes today: right at the end of downloading, the phrase "Download Complete" appears; verification, however, never begins; and the IDE hangs indefinitely.

======================================
Included is the extensive exported system info.

Briefly:
Win XP professional sp2
xeon 1.7GHz cpu, 1 gb ram
Dell model M-869 mouse
Dell model RT7D10 keyboard

Display interaction (one of several)

-doing nothing; just downloading. Verifying causes a crash generally *without* the VB debug info. Oddly, after restarting the IDE, a large percentage of the time a crash occurs on the very next download, this time *with* the VB debug info. This paired sequence occurs fairly often.

-other times the IDE is attempting download to a device which is powered off; upon powering on the device during its attempts, the IDE crashes.
Attachments
sys info.txt
System Information
(300.68 KiB) Downloaded 3889 times
pjc30943
Posts: 220
Joined: 01 December 2005, 18:45 PM

Post by pjc30943 »

Second Edit:

Any progress on solving the crash issues?

---------------------------------------------

[did not Edit the previous post, instead added one on here: not sure if editing bumps a post]

Is there any other information required that might be of use?

Perhaps someone else could also posted their info...
pjc30943
Posts: 220
Joined: 01 December 2005, 18:45 PM

Post by pjc30943 »

Any progress on solving the crash issues? Please let me know if there anything else I can help with regarding the issue:)
dkinzer
Site Admin
Posts: 3120
Joined: 03 September 2005, 13:53 PM
Location: Portland, OR

Post by dkinzer »

pjc30943 wrote:Any progress on solving the crash issues?
In a word, no. We've attempted to reproduce the problem on several different systems running either XP (SP2) or Win2K. Obviously, there is something we're missing - either some part of the procedure, UI action/timing or a certain version of system DLLs.

What we need is a step-by-step, very detailed set of instructions that anyone can use to reliably reproduce the problem. If we can't get a recipe for reproducing it, we'll probably have to go to a site where the problem can be reliably reproduced, install Visual Studio on the box and work on it there. Another alternative is if someone has a laptop that exhibits the problem and they can bear to be without it for a week or so it could be shipped here. Of course, we'd be obliged to pay for shipping both ways.

We would be ecstatic if we could resolve this issue.
- Don Kinzer
spamiam
Posts: 739
Joined: 13 November 2005, 6:39 AM

Post by spamiam »

I have visual studio and VC++6.0, (as you know, Don). I still get crashes on an irregular but reasonably frequent basis after I compile. When I hit "Go", then it will cease functioning about 20% of the time, but rather randomly.

Right now I am not in the cycle where I am developing on the ZX, but I will be doing so in the reasonably near future.

Unfortunately, ZBasic only crashes on my primary computer, the one I can not bear to be without, so sending it anywhere is not a possibility. On my laptop, it is fine.

Is there anything I need to do to get Studio and VC to give the best debugging info?

In the past when we were trying to track this down, even with all the debugging data, I still had trouble getting useful info to you. I just did not know enough about VC/Studio to trace anything, and since the bug only strikes once in a while, it is a bugger to even see it happen! Recently I have been doing a little work on an old version of VB (2003) and I am slightly more conversant with studio.

-Tony
dkinzer
Site Admin
Posts: 3120
Joined: 03 September 2005, 13:53 PM
Location: Portland, OR

Post by dkinzer »

spamiam wrote:ZBasic only crashes on my primary computer [...]. On my laptop, it is fine.
That is an interesting observation. It would be interesting to see the differences in DLLs that exist on each. You may recall an earlier post about using ListDLLs to get information about what DLLs are installed on a system.

A second possible clue is that someone reported that they don't have the problem if they turn off the download progress indicator - the dots displayed to indicate progress during download and verification. You can try that out by opening the Serial Port Options dialog from the Options menu. In the group box in the upper left corner, uncheck the checkbox entitled "Display progress indication".
Attachments
download_quiet.jpg
download_quiet.jpg (20.5 KiB) Viewed 6375 times
- Don Kinzer
spamiam
Posts: 739
Joined: 13 November 2005, 6:39 AM

Post by spamiam »

dkinzer wrote:It would be interesting to see the differences in DLLs that exist on each. You may recall an earlier post about using ListDLLs to get information about what DLLs are installed on a system.

A second possible clue is that someone reported that they don't have the problem if they turn off the download progress indicator - the dots displayed to indicate progress during download and verification. You can try that out by opening the Serial Port Options dialog from the Options menu. In the group box in the upper left corner, uncheck the checkbox entitled "Display progress indication".
I can do that. Both my laptop and my desktop (floortop) are running XP home, but not identical versions. The laptop is newer than the desktop, and I think it might even be SP2 out f the box.

I will try unchecking the progress indicator and see what happens.

-Tony
mikep
Posts: 796
Joined: 24 September 2005, 15:54 PM

Post by mikep »

dkinzer wrote:A second possible clue is that someone reported that they don't have the problem if they turn off the download progress indicator - the dots displayed to indicate progress during download and verification. You can try that out by opening the Serial Port Options dialog from the Options menu. In the group box in the upper left corner, uncheck the checkbox entitled "Display progress indication".
Note that both "Display progress indication" and "Display firmware version" must be unchecked and the workaround should hopefully work. Once several people have verified that this actually works, it will be easier for Don to implement a solution.

When you do this will you see a slight pause while the download and verify occur and then the standard Signon message will be displayed, assuming it is not turned off using "Options Signon"

Code: Select all

ZBasic v2.3
Mike Perks
spamiam
Posts: 739
Joined: 13 November 2005, 6:39 AM

Post by spamiam »

Well, I did a few compiles and got it to crash as I hit "Go". I de-selected the progress indicator, and did some more compile/go iterations. I had another crash, but this time it was as I hit compile. The error box was the same as for when it crashed during the previous "go". It is the second error box that was posted in a previous posting by someone else. It was a failed assertion....

Any other suggestions? As I said, I do have VC++ 6.0 and the accompanying Studio.

-Tony
dkinzer
Site Admin
Posts: 3120
Joined: 03 September 2005, 13:53 PM
Location: Portland, OR

Post by dkinzer »

I believe that we have resolved both the access violation and the lockup issue. We have posted an update containing a few changed files that should be extracted to the IDE subdirectory of the ZBasic installation directory.

The new version also supports selectable help file format (HTML or PDF) via the Help menu. In order for this to function properly, you must delete entries that you may have added to your User Properties files to control the help file. The lines that might have been added have prefixes like those in the list below. If you didn't manually add such lines to your User Properties file, there is nothing that you need to do.

Code: Select all

command.syslib.help=
command.help.$(file.patterns.bas)=
- Don Kinzer
dkinzer
Site Admin
Posts: 3120
Joined: 03 September 2005, 13:53 PM
Location: Portland, OR

Post by dkinzer »

dkinzer wrote:I believe that we have resolved both the access violation and the lockup issue.
I neglected to mention that much of the credit for the ultimate solution belongs to Mike Perks. From his observations of the conditions under which the access violations occurred, he deduced what might be causing the problem and ultimately found some relevant posts on the SciTE newsgroup that described how to avoid it.

Thanks again, Mike.
- Don Kinzer
Don_Kirby
Posts: 341
Joined: 15 October 2006, 3:48 AM
Location: Long Island, New York

Post by Don_Kirby »

Yes, thank you Mike. You're possibly the most prolific contributor here (the extraordinary Mr. Kinzer excluded).

-Don
pjc30943
Posts: 220
Joined: 01 December 2005, 18:45 PM

Post by pjc30943 »

Thank you Mike and Don! The fix appears to work.
dkinzer
Site Admin
Posts: 3120
Joined: 03 September 2005, 13:53 PM
Location: Portland, OR

Post by dkinzer »

pjc30943 wrote:The fix appears to work.
Thanks for the confirmation. I only wish that it could have been corrected earlier. I never was able to replicate the access violation problem and was only able to replicate the lockup problem after Mike gave a precise description of a method for doing so. Again, many thanks to Mike as well as the others whose descriptions gave him clues as to the nature of the problem.
- Don Kinzer
Post Reply