Re: [bug-gawk] Possible regression in gawk 4.x regarding I/O errors

[Assaf Gordon]

Hello,

On 07/24/2014 03:14 AM, Aharon Robbins wrote:
> Interesting test case. I was able to reproduce it. An enormous THANK YOU
> for the excellent bug report.
>
> Here is a patch that doesn't break 'make test'.  Let me know if it works
> for you.
>
> Thanks,
>
> Arnold
> ------------------
> diff --git a/io.c b/io.c
> index 57d4af2..c291dc9 100644
> --- a/io.c
> +++ b/io.c
> @@ -596,6 +596,8 @@ inrec(IOBUF *iop, int *errcode)
>                 INCREMENT_REC(NR);
>                 INCREMENT_REC(FNR);
>                 set_record(begin, cnt);
> +               if (*errcode > 0)
> +                       retval = 1;
>         }
>
>         return retval;

Yes, the above patch fixes the issue.

I wonder, though, looking at the above "if" block in the same "inrec" function, should 
"update_ERRNO_int" also be called?

Thanks for the fix!
 - Assaf