Re: _REENT_SMALL IO broken

[Jeff Johnston <jjohnstn at redhat dot com>]

Paul Brook wrote:
> > > a) Every use of a file descriptor needs to check if it has the read FD,
> > > or a dummy copy. In this case there's no point having the dummy
> > > structure, we can just assign stdin/out/err magic values. As a bonus
> > > _reent gets even smaller, at the expense of a small code size increase.
>
>
> > IMO, the right solution is a) but not the 2nd part of your proposal.
> > The reason we have the dummy structure is to handle the stdio macros
> > (e.g. getc and putc).  For speed/compactness, we want them to think they
> > are operating on a real FILE * and we set the values such that we always
> > force a function call to occur based on their regular logic path.
>
>
> I guess this is the only way to do it without slowing down the !_REENT_SMALL 
> case. However consider:
>
> #include <stdio.h>
> int main()
> {
>   FILE *foo = stdout;
>   fprintf(foo, "Hello World\n");
>   if (foo != stdout)
>     abort();
>   return 0;
> }
>
> I find it surprising for this program to abort.

_REENT_SMALL was specifically added for embedded platforms that are 
extremely tight on storage (some verging on crippled IMO).  It should 
not be used by platforms that just want to save a few bytes.  It is not 
surprising that these platforms can't do everything in a Standard's test 
bucket (some of them don't even do file I/O).  Storing away the standard 
streams isn't something any reasonable program other than a concocted 
standard-test is going to do.  It certainly won't be found in real code 
written for such platforms.

I'll start working on the fix.

> There are also more subtle bugs that can occur when using the dummy FILE* with 
> the feof, ferror, and clearerr macros.

These macros should be ok.  You shouldn't be at EOF since you haven't 
issued any read and you haven't caused any error to occur prior to the 
first reference.

-- Jeff J.