Re: [PATCH v2] newlib: libc: Fix bugs in the commit 3d94e07c49b5.

[Corinna Vinschen <vinschen@redhat.com>]

Hi Takashi,

On Nov 16 21:27, Takashi Yano wrote:
> Hi Corinna,
> 
> On Wed, 15 Nov 2023 17:31:39 +0100
> Corinna Vinschen wrote:
> > Hi Takashi,
> > 
> > On Nov 13 14:07, Corinna Vinschen wrote:
> > > On Nov 11 01:09, Takashi Yano wrote:
> > > > diff --git a/newlib/libc/stdio/ungetc.c b/newlib/libc/stdio/ungetc.c
> > > > index 79914af08..5053fd6c4 100644
> > > > --- a/newlib/libc/stdio/ungetc.c
> > > > +++ b/newlib/libc/stdio/ungetc.c
> > > > @@ -125,12 +125,6 @@ _ungetc_r (struct _reent *rptr,
> > > >  
> > > >    _newlib_flockfile_start (fp);
> > > >  
> > > > -  if (ORIENT (fp, -1) != -1)
> > > > -    {
> > > > -      _newlib_flockfile_exit (fp);
> > > > -      return EOF;
> > > > -    }
> > > > -
> > > >    /* After ungetc, we won't be at eof anymore */
> > > >    fp->_flags &= ~__SEOF;
> > > >  
> > > > @@ -213,6 +207,8 @@ int
> > > >  ungetc (int c,
> > > >         register FILE *fp)
> > > >  {
> > > > +  if (ORIENT (fp, -1) != -1)
> > > > +    return EOF;
> > > >    return _ungetc_r (_REENT, c, fp);
> > > >  }
> > > >  #endif /* !_REENT_ONLY */
> > > 
> > > As per the reasoning in terms of old standards, we shouldn't check or set
> > > the orientation in ungetc at all.
> > 
> > Given all that has been discussed, I attached a followup proposal.
> > I hope that's ok.  We may want to release 3.4.10 in the next weeks.
> 
> LGTM, except for:
> 1) Shouldn't we add flags2 = 0 also for _fmemopen_r(), _fopencookie_r(),
>    _funopen_r() and internal_open_memstream_r()?

Not necessary.  All three functions get the FILE* from a call to __sfp,
which always sets _flags2 to 0 anyway.

> 2) __swbuf_r() has ORIENT (fp, -1) call, however, result is not checked.

Good catch.  Looks like setting the orientation at this point is
wrong.  __swbuf_r is called via the __sputc_r macro in stdio.h.
But __sputc_r is also called from __fputwc as well as, potentially,
from _fputws_r.

AFAICS, we already set the orientation from the calling functions
_putc_r (covering putchar, putc, fputc) and from _puts_r as
well as _fputs_r (covering puts, fputs).

On the wide-char side, we set the orientation from _fputwc_r
(covering putwchar, putwc, fputwc) and from _fputws_r (covering
fputws).

So I *think* we can safely remove the ORIENT call from __swbuf_r,
or did I miss some other way to call __swbuf_r?


Thanks,
Corinna