llvm dev - May 2010 - [LLVMdev] "\01__isoc99_fscanf"

I have a bitcode file Y that has a symbol __isoc99_fscanf in it.  When 
I disassemble the bitcode, the symbol is written as 
"\01__isoc99_fscanf".  I am wondering where this symbol is coming
from?
  I grep'ed throughout /usr on my machine for both names, and I only 
found __isoc99_fscanf, without the special character in front.

Regardless of where the symbol is coming from, I am trying to write a 
wrapper function in C, for each external function called in the bitcode 
file Y.  By my own convention, the wrapper for a function f() is given 
the name f___xxxx___().  However, I can not seam to imitate the special 
character at the front of __isoc99_fscanf, when I'm writing the C code. 
  Is there a way to do it, or do I need to revise my naming convention.

Regards,
Ryan

On 29 May 2010 02:43, Ryan M. Lefever <lefever at illinois.edu>
wrote:
> I have a bitcode file Y that has a symbol  __isoc99_fscanf in it.  When
> I disassemble the bitcode, the symbol is written as
> "\01__isoc99_fscanf".  I am wondering where this symbol is coming
from?
>  I grep'ed throughout /usr on my machine for both names, and I only
> found __isoc99_fscanf, without the special character in front.
I'd say this is a bug in your disassembler. Even C++ mangling routines
don't use non-printing chars (like \01) in function names.

If you compile the bitcode directly to assembly or object file, do you
still see the weird symbol? Can you send the bitcode file to the list?

cheers,
--renato

Hello, Ryan
> "\01__isoc99_fscanf".  I am wondering where this symbol is coming
from?
It's comming from the gcc's asm("foo") symbol renaming
extension.
The "\01" prefix just tells the mangler that it should not touch the
symbol name at all.

-- 
With best regards, Anton Korobeynikov
Faculty of Mathematics and Mechanics, Saint Petersburg State University

Hello, Renato
> I'd say this is a bug in your disassembler. Even C++ mangling routines
> don't use non-printing chars (like \01) in function names.
No, it's ok. See my prev. e-mail to list.

-- 
With best regards, Anton Korobeynikov
Faculty of Mathematics and Mechanics, Saint Petersburg State University