coder.ceval for printf from stdio.h

Question

Tomas Jurena 2012-5-27

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/39532-coder-ceval-for-printf-from-stdio-h

评论： Kevin Berndsen 2015-3-18

Hello, everybody! I'd like to generate C code from an m-file which makes use of MATLAB fprintf function as follows:

fprintf('Time level %-*.*f [s], Iterations: %d\n',field_width,precision,time,iter);

Since fprintf is not supported for the code generation, nor can I use coder.extrinsic method, I turn to a question:

How to rewrite the command using coder.ceval method and the standard C library function printf? To be more specific, what should the arguments be like in coder.ceval('printf', arguments); ?

While not using field_width and precision, C code generated from the following one executes without a problem:

s1=['%s%f%s%d' 10 0];  
s2=['Time level ' 0];
s3=['[s], Iterations: ' 0];
coder.ceval('printf',s1,s2,time,s3,int32(iter));

However, including the asterisks and left-justify options cause segmentation fault whenever the generated C code is compiled, built and run. Any help is appreciated!

3 个评论
显示 1更早的评论隐藏 1更早的评论

Tomas Jurena 2012-5-30

Thank you for your comment. It really seems to be a bug, because running a MEX function causes MATLAB crash as well. I'll contact Tech support.

Tomas Jurena 2012-5-30

OK, no need to do that :) See the answer given by Fred Smith.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Fred Smith 2012-5-30

2
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/39532-coder-ceval-for-printf-from-stdio-h#answer_49499

在 MATLAB Online 中打开

This example worked for me:

 function simple_printf
 s1=['%s%-*.*f%s%d' 10 0];  
 s2=['Time level ' 0];
 s3=['[s], Iterations: ' 0];
 iter = 32;
 time = 2.5;
 field_width=int32(20);
 precision=int32(3);
 coder.ceval('printf',s1,s2,field_width,precision,time,s3,int32(iter));

Is your code different?

-Fred

2 个评论
显示无隐藏无

Tomas Jurena 2012-5-30

Yes, it is just a little bit different to yours, but the difference explains the origin of the problem. Both field_width and precision must be cast to an integer type. Now, it works for me, too. Shame on me, that I forgot about it :) Thank you very much!

Kevin Berndsen 2015-3-18