C formating question between Unix and Windows

[sitko]

Hi,

I've continued with my VB->C conversion, and have it working on the Linux box, now I'm going back and getting it to work on the Windows box.

Something strange occurred and I thought I'd try to get you guys input on it.

The code:

fprintf(datFile, "%10d%10.3e%10.2f%10.3f\n",(intIndex*100+intPart),gdblX,gdblY,gdblZ);
 

Creates on Unix:

"       101 8.800e-05 180000.00     0.400"
 

But, in windows the same code creates:

"       1018.800e-005 180000.00     0.400"
 

Which is the same thing except for an extra "0" in the "e" number.

Are there different rules for window formating than unix formating?

Thanks,
Sitko.

[sitko]

My co-workers(engineers) are telling me that 8.800e-05 and 8.800e-005 are the same number, and that it shouldn't be an issue.

But, if anyone has any ideas how to fix this, I'd still be interested in your thoughts.

Note: this file is made up of 10 character wide 'numbers' (padded with spaces) so the fact that the 8.800 runs into the 101 isn't an issue. If you were wondering about that.

Thanks,
Sitko.

[sicarie]

I'd be interested in what compilers you were using on both. Did you use GCC? Or is this possibly a VC/GCC implementation issue.

[sitko]

I'd be interested in what compilers you were using on both. Did you use GCC? Or is this possibly a VC/GCC implementation issue.

On unix, I used GCC, and on windows, I'm working in Visual C++ (2005).

Its really bizarre how many odd errors are showing up on the windows side which work perfectly fine on the Unix side. Glad there is the ability for platform compiling directives...otherwise, i'd never be able to get this to work on both OSs.

Thanks,
sitko.

[sicarie]

On unix, I used GCC, and on windows, I'm working in Visual C++ (2005).

Its really bizarre how many odd errors are showing up on the windows side which work perfectly fine on the Unix side. Glad there is the ability for platform compiling directives...otherwise, i'd never be able to get this to work on both OSs.

Thanks,
sitko.

Yeah, I'm betting it's just the way libraries are implemented on each. Did you compile GCC with the -wall option? Depending on what they are, you might see some warnings about some of the things that VC++ is complaining about...

[weaknessforcats]

Creates on Unix:

Code: ( text )
" 101 8.800e-05 180000.00 0.400"

But, in windows the same code creates:

Code: ( text )
" 1018.800e-005 180000.00 0.400"

These don't look the same to me. I mean 101 8 isn't the same as 1018.

That space causes the e-05.

Have you triedn the -pedantic on g++ to disable non standard C++ extensions?/

[sitko]

These don't look the same to me. I mean 101 8 isn't the same as 1018.

That space causes the e-05.

Have you triedn the -pedantic on g++ to disable non standard C++ extensions?/

As I mentioned above, the program that uses these data files looks for 10 character entries, so even though the 101 is right next to the 8.800..., it still separates the two numbers in its code.

In a further update on my effort, I first got the Unix version to work, then started working on the Windows version, which was WAY harder to get it to work...but eventually I got it working, then for giggles I took the window version back to Unix, to see what additional changes I'd need to make, and it still worked in Unix without having to make (or add ) any changes including additional conditional compiling, which I figured I was going to have to do more so than I had already.

This really impressed me that Unix is so forgiving while windows is so non-forgiving...

Thanks for everyone's help,
Sitko.