Two PCs, same EXE & Data - different answer!

[GriffMG]

Anyone else come across this?

I have an application which is running on about 100 machines altogether, and go a call today from a user who said his one was reporting a different answer to his neighbours (his was wrong and was MISSING info).

I got a copy of his data, loaded it on mine and I get the right answer. To be sure, I sent him my exe file and his is still wrong.

I got him to load the system & data on another (fourth) machine and it gives the right answer.

The user says he is getting a new laptop next week, because his is 'playing up' - but this seems a rather weak explanation!

Data is held locally in all cases, as is the .exe.

TIA

Regards

Griff
Keep [Smile]ing

 

[Elkari]

How old is the computer that is giving the wrong answer? Try having them run a few diagnostics on it. It is possible a memory chip went bad. It happen to one of my clients when they upgrade the RAM. They did not bother checking he parity of the chips and all calculations came out screwy.

Another possiblity, but then the computers have to be really old, is that Pentiums did have a math problem when they first came out. If the computer is that old, it time to toss it.

 

[GriffMG]

Elkari,

I don't know for sure the age, I know the user is a little bit 'unlucky' shall we say with his laptop (dropped, repaired, overheated etc.) and he says it is problematic.

I was just a tad worried that this might be more common than I thought!

I would suspect a hardware problem - dodgy motherboard or somesuch then - and hope for the best!

I've been running clipper apps since about '87 and this isn't the very first time I've come across this kind of thing - but it has been years since I had to think about it!

Thank you for your thoughts!

Griff Regards

Griff
Keep [Smile]ing

 

[Bozz]

Griff,
Was the wrong answer tied to a date/time or any other "movable" part of a calculation? What sort of data was missing.
Ian Boys
DTE Systems Ltd

 

[GriffMG]

Hi Ian,

The result was the omission in a report program of all records of a certain type... denoted by a 'twice removed' kind of a look-up... i.e. not adding all the records whose type (look-up A) is found in look-up B.

On all other machines, since about 1995 (no kidding) the answer has worked out fine, but, as of last Friday apparently, this one laptop - held by a chum of mine (who is a BIT clumsy) -, has ignored those records (in this one report) which have a 'twice removed' look-up yielding a 'Y' in the field that should make them included!

Sorry, this was mildly frustrating earlier - now it begins to feel like a pain!

Thanks for your interest (no dates or times or other 'local' variables/fields in the calculation!)

Thanx again... Regards

Griff
Keep [Smile]ing

 

[Bozz]

I just hope it's not catching! I know silly thing happen, I changed fron 5.01a to 5.2e compiler recently and the only problem was a corrupt index, but the same corrupt index on systems hundreds of miles apart, but only on 20% of the PC's using it had the problem. No logic, just write it down to experience. As someone on another forum says "I want to live in theory, everthing works in theory" Ian Boys
DTE Systems Ltd

 

[GriffMG]

Can you not emigrate to Theory... got to be on a map somewhere!
Regards

Griff
Keep [Smile]ing

 

[circes9]

may want to search that PC's local drive and paths for existance of files that dont belong there. Because they are found in the PATH setting, it would be as though they were local to that application and could be affecting the report logic.

 

[GriffMG]

circes9

I'm going to see the users machine next week - I'll have a look!

Thank you for the thought

Regards

Griff
Keep [Smile]ing