This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] gold: don't assume read/pread never return short counts


On Fri, Dec 7, 2012 at 1:47 PM, Roland McGrath <mcgrathr@google.com> wrote:
>
> gold/
> 2012-12-07  Roland McGrath  <mcgrathr@google.com>
>
>         * testsuite/binary_unittest.cc (read_all): New function.
>         (Sized_binary_test): Use it instead of ::read.
>         * fileread.cc (do_read): Don't assume pread always reads the whole
>         amount in a single call.

> +          bytes = ::pread(this->descriptor_, read_ptr, to_read, read_pos);
> +          if (bytes < 0)
> +            {
> +              gold_fatal(_("%s: pread failed: %s"),
> +                         this->filename().c_str(), strerror(errno));
> +              return;
> +            }

No  need to return after gold_fatal (I know the current code has that
too).  Just take out the return statement.

This is OK with that change.

Thanks.

I don't think we need to worry about 2.23 since this is not the normal
case anyhow.  The normal case is that whole_file_view_ is not NULL.

Ian


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]