Re: [boost] [system][filesystem v3] Question about error_codearguments

Subject: Re: [boost] [system][filesystem v3] Question about error_codearguments
From: Detlef Vollmann (dv_at_[hidden])
Date: 2009-10-31 21:11:42

On 10/31/09 04:49, Gottlob Frege wrote:
> bool diff(file srcA, file srcB, file output)
> {
>    // get our files ready...
>    error_code srcA_err = file_open(srcA);
>    error_code srcB_err = file_open(srcB);
>    error_code out_err = file_open(output);
> 
>    if (srcA_err || srcB_err || out_err) // check for errors
>       return false;
> 
>    // do the diff...
> 
>    return true;
> }
> 

> So that took a bit of thought for me to come up with, but the code
> doesn't look completely unreasonable, does it?
Hmmm, that depends on your background.
If you look at a respective POSIX program:

   int srcA_err = open(srcA);
   int srcB_err = open(srcB);
This will overwrite errno.
So system level programmers are used to write:
   int err = open(srcA);
   check_error(err, "first source file");
   err = open(srcB);
   check_error(err, "second source file");

So this specific issue should be teachable.

> Basically, you never want more than one actual error_code to exist at
> a time (per thread).
Hmmm, what about:

void do_the_diff(file &srcA, file &srcB, file &out)
{
   error_code readA_err, readB_err, out_err;
   // chunksize, chunkA, chunkB

   bytes_read = read(&readA_err, srcA, &chunkA, chunksize);
   // check for and handle EINTR (only)
   bytes_read = read(&readB_err, srcB, &chunkB, chunksize);
   // check for and handle EINTR (only)

   // ...
}
Here we would throw in case any of the error_code object would
have another error than EINTR.

This becomes more interesting both contain such an error.
I'm not so sure that I like this approach anymore...

   Detlef