using clang optimizing compiler yields run-time errors not encountered when using CVI standard compiler

[Wolfgang]

Hi,

I have compiled an executable using Clang with the template provided by NI.

Interestingly I receive errors (generated by checking the UI return code) that I do not get when running the executable created using CVI's standard compiler.

There are no compilation warnings / errors by neither compiler.

Any hints what one could do to track down the problem?

Thanks!

[nickb]

Hey Wolfgang -

I've not seen anything like this while using the Clang compiler.  Is there anyway you can post some code that reproduces what you're seeing?

NickB

National Instruments

[Suse]

Hey Wolfgang,

I subscribe to NickB's view. Without being able to reproduce the issue, we cannot help you.

Best regards

Suse

[stef_fr]

Hi Wolfgang,

I know that your issue is quite old but may be this post will help.

I have also observed differences between Clang and CVI2012 compiler.

In the following expression :

a[i] = b[i++];

CVI compiler will evaluate '++' after '=' operator while Clang does it before.

The consequence is that CVI compiler performs a[N] = b[N] while Clang performs a[N+1]=b[N]. 

The project in attachment shows this behaviour, the message box will display different strings when compiled with CVI or Clang compiler. 

Regards,

Stephane

[Peter_Ilberg]

Hello!

The behavior of your code snippet is undefined. These links explain the issue better than I can.

http://c-faq.com/expr/evalorder1.html

http://c-faq.com/expr/evalorder4.html
http://c-faq.com/expr/seqpoints.html
http://en.wikipedia.org/wiki/Sequence_point

Sequence points in the C99 standard: section 5.1.2.3 and appending C. See also the index.

Regards,

Peter