EFS causing eSignal error report/system rebooting

[rcameron]

I have been using the attached script for some days and seemingly it invokes the eSignal crash report module (and sometimes just a pure system crash with rebooting) and I cannot find what in the script could be causing the problem. I cannot explain why my system or eSignal would behave in this way except for this script I believe but am not really sure.

I would appreciate any assistance to find the bug if one exists or at least know that it is not the script so that I can investigate some other contributing issue.

I am using version 7.5 build 628 on an XP Pro (fully updated) AMD 2100 processor system.

Robert

[DuaneG]

Hi Robert,

I have been running the EFS you provided and have not had any issues. I'm running eSignal 7.6 build 635. Our hardware setups are comparable so I'm not sure where the error is generating from. If you have any specific debug messages from the crash log, they may be helpful in determining what exactly is occurring. Please let me know.

[rcameron]

Hi Duane,

Thanks for testing the efs script. Whenever I had crashes that were caught by eSignal I sent those through to eSignal using the report transfer utility, however I did not take down any details of the crash (is there a way to do that?). Whenever I had system reboot crashes that were caught by Windows, I sent those through to Microsoft and when I inspected the error it was to do with an .xml file posted in a temporary directory (presumably by eSignal).

Anyway I have incrementally been adding various error trapping routines to see if I could eliminate the crashes, in fact my system has not crashed today at all (probably will now!) since I added the following lines yesterday:-

Near the start of main:-

////////////////////////////////////////////
// This will return null if no value for rawtime
if (getValue("rawtime") == null) return;


At the end of main:-

return null;


Whether this has done the trick remains to be seen, it would be nice for other users to confirm that the script is stable on their systems so I can look at other scripts if my system continues to crash.

Robert

[Guest]

Robert,

I looked at your code and nothing stood out.

If you continue to have problems, and it appears that you do not, I have a solution which I guarantee will narrow the problem down to the offending step. I have attached the trace program, which can be disassembled and easily added to your code. At the end of each line, paste these two steps in (this example would for example go at the end of step 356)

trace("356:");trace("save");

this will write to a file and close the file prior to the last step that works. I personally have had to do this several times. For your code, I estimate it would less than an hour to insert. (especially if you do a global replace of ";" with ";trace(":");trace("save");" and fill in the step numbers later). The program is also in my fileshare area entitled functions

Regards,

[rcameron]

Steve,

Thanks for looking over the code. I did not crash my system yesterday after extensive usage. So maybe my recent error trapping routines have solved the problem, I hope it has, as it has been a very frustrating couple of weeks for me.

Anyway I'd like to know from others if the code is stable on their system or if there are any problems at all.

Robert

[Gavishti]

When I have problems (i.e., crashes) it usually turns out to be something I did in an efs. These take a while to weed out. Another thing I've noticed is that charts can get corrupted. The only way I have found to get around that problem is to rebuild them from scratch. Several versions ago I was having lots of problems. I rebuilt every chart and layout from scratch and the problems magically disappeared. Yes, several things in eSignal do not work as advertised or implied. However, it seems to be stable. I have yet to trace a crash directly back to eSignal. Perhaps someone that has a better understanding of what is going on under the hood can elaborate on these issues.