Unexpected output in Code::Blocks for specifier %Lf

Here's the code that works in Xcode C on OSX but not Windows XP using Code::Blocks version 10.05 May 27, 2010 build:
Extra info, C code using MinGW/GCC

Code:
long double a = 3.4;
printf("a is %Lf \n",a);

Output on OSX is "a is 3.4"
On Win XP it is "a is -42307582002573505000000000000000000000000000000000 00000000000000000000000000000000000000000000000000 000000000000000000000000.000000"

Why please?

Originally Posted by FloridaJo

Here's the code that works in Xcode C on OSX but not Windows XP using Code::Blocks version 10.05 May 27, 2010 build:
Extra info, C code using MinGW/GCC

Code:
long double a = 3.4;
printf("a is %Lf \n",a);

Output on OSX is "a is 3.4"
On Win XP it is "a is -42307582002573505000000000000000000000000000000000 00000000000000000000000000000000000000000000000000 000000000000000000000000.000000"

Why please?

C data types - Wikipedia, the free encyclopedia --- see the table

I don't see what you are pointing out.
Shouldn't a long double be able to handle 3.4 or 1.0 for that matter?

It could be a compiler or standard library bug. You could try:

Code:

long double a = 3.4L;

and if it works, then perhaps the bug is with conversion from double to long double, otherwise the bug might be with printf. Of course, I am assuming a bare bones test program, otherwise the bug might be with the code that you did not show.

Yes, bare bones test code.
That is the only code.
Using 3.4L did not change anything.
It's a 10 year old computer, the Code::Blocks is newly loaded on it,
but I'm not sure if the Compiler came with it or it's pulling it from
within the computer and there's a bug in that.

From reading several sites Google turned up; MinGW GCC supports a long double that is NOT a perfect match to what Microsoft C run-time does. Since MinGW uses Microsoft C run-time to do printf it fails to work right.

NOTE: If you compiled with warning on you would see warnings.

Tim S.

Originally Posted by FloridaJo

Yes, bare bones test code.
That is the only code.
Using 3.4L did not change anything.
It's a 10 year old computer, the Code::Blocks is newly loaded on it,
but I'm not sure if the Compiler came with it or it's pulling it from
within the computer and there's a bug in that.

Sorry about my non-specific initial reply. I thought it would just be a matter of the using the right type specifiers. However, I came across this (after having the same problems you did on W7 and Code::Blocks, MinGW).

Printing "long double" type via printf() on MinGW g++ 3.2.3 - C / C++

It hints that MinGW might be broken in this respect, which is news to me. Bears some more reading anyway.

Don't think it's your fault. On linux, using %Lf worked fine, with or without appending "L" to the literal assignment. Had the exact same problems with CB and MinGW

EDIT: stahta01 beat me to the punch!

Yes, it appears it is a MinGW problem or the library MinGW uses. Not sure which.
But either way, thanks everyone for saving some more hairs on my head.
I just use that computer in my shop to play around with, the iMac is a better station.