LabVIEW
error 1097 in call library function
mhs100 wrote:
Hi Guenter Mueller
I can not uderstand why all the time when i run the program , the error as i described above occurs again and again.
The error is mentioned in the error.doc file which is attached in the above post.
Please any other volunteer look into this.
There are several possible reasons:
1) Your CallLibrary Node is not configured properly. Calling convention,number of parameters, parameter types and size of parameters.
2) Size of parameters is especially important. If you pass in a buffer for the DLL to write data in, this buffer needs to be preallocated in the correct size before passing it o the CLN.
3) another possibility is simply that you have a bug in your DLL.
And ooh before I forget, I do not open doc files from strangers.
mhs100 wrote:Hi Rfolk
You have given a general answer. I have checked all of the things which you have mentioned like calling convention ,
number of parameters , data types etc.
Please check the coding of dll in mhsdlldef.zip in post number 17 and reply to post number 20 and please also check the mhsVI.zip also.
Well some code simply doesn't deserve much time. Your VI has no data dependency at all. How do you hope that your LoadDLL() is executed BEFORE any of the other functions????? There is a 1 to 5 chance that it won't, making LabVIEW call into functions which try to call invalid function pointers. So my point 3 already applies.
After such an obvious error like this I have absolutely no interest in checking the entire DLL for parameter errors or anything else. Also it is your DLL not mine so please be a bit more polite about requesting me what to do.
Also I don't like code sprinkled with global variables. My problem I know but it just turns me off.
You should really look for some references about dataflow programming. Dataflow is what defines the execution order in LabVIEW. If you do not have any dataflow such as with your intilialization functions you can cause a sequential operation by using the ... exactly, Sequence Structure.
Now since you are a newbie and can be educated a bit: Sequence Structures are in general bad, especially the stacked one. It hides code and makes thing hard to overview. The parallel Sequence Structure has that problem at least not, but it is still a cludge for most practical uses.
The proper way is to put all your little Call Library Nodes into subVIs, add proper error handling to all of them (input error cluster, not executing anything if that cluster indicates an incoming error except for the close function which always should attempt to close but needs of course protection so that it does not crash if the DLL never has been initialized (check of the close function pointers to be valid), then if the function returns an error (and does not exit(1) as it does now) evaluate that error and put it in the error out cluster, and last but not least wire those VIs together, with the error cluster being the natural data flow.
Your use of exit(1) is VERY bad for a DLL. This will tear down the entire application, looking like a crash. Instead, make the functions return an error code, with 0 if everything was ok and another value if something went wrong.
