From: Stefan Beller <sbeller@google.com>
Date: Wed, 19 Nov 2014 01:41:17 GMT
Subject: Re: [PATCH 1/4] error: save and restore errno
Message-ID: <CAGZ79kbNatX51P3kocndyR+UVeACQR_u3OhkZiJmZrq_UBoJxg@mail.gmail.com>
In-Reply-To: <20141119013710.GA2135@peff.net>

This one makes my day.
A really good fix as it minimizes maintenance burden for checking
incoming patches for that pattern.

Reviewed-by: Stefan Beller <sbeller@google.com>


On Tue, Nov 18, 2014 at 5:37 PM, Jeff King <peff@peff.net> wrote:
> It's common to use error() to return from a function, like:
>
>         if (open(...) < 0)
>                 return error("open failed");
>
> Unfortunately this may clobber the errno from the open()
> call. So we often end up with code like this:
>
>         if (open(...) < 0) {
>                 int saved_errno = errno;
>                 error("open failed");
>                 errno = saved_errno;
>                 return -1;
>         }
>
> which is less nice. Let's teach error() to save and restore
> errno in each call, so that the original errno is preserved.
> This is slightly less efficient for callers which do not
> care, but error code paths are generally not performance
> critical anyway.
>
> Signed-off-by: Jeff King <peff@peff.net>
> ---
> It's pretty minor to just handle errno in the callers, but I feel like
> I've wanted this at least a dozen times, and it seems like it cannot
> possibly hurt (i.e., I imagine there are callers where we _should_ be
> doing the errno dance but have not realized it).
>
>  usage.c | 3 +++
>  1 file changed, 3 insertions(+)
>
> diff --git a/usage.c b/usage.c
> index ed14645..ee44d57 100644
> --- a/usage.c
> +++ b/usage.c
> @@ -142,10 +142,13 @@ void NORETURN die_errno(const char *fmt, ...)
>  int error(const char *err, ...)
>  {
>         va_list params;
> +       int saved_errno = errno;
>
>         va_start(params, err);
>         error_routine(err, params);
>         va_end(params);
> +
> +       errno = saved_errno;
>         return -1;
>  }
>
> --
> 2.1.2.596.g7379948
>