Re: [tig PATCH] continue updates when pipe read has errno "Success"

From: Jonas Fonseca <jonas.fonseca@gmail.com>
Date: Aug 22, 2008, 10:10 UTC
Message-ID: <2c6b72b30808220310v1083e860tef9fd6efa57767e3@mail.gmail.com>
In-Reply-To: <20080821014043.GA14452@coredump.intra.peff.net>

On Thu, Aug 21, 2008 at 03:40, Jeff King <peff@peff.net> wrote:

Show 5 quoted lines

> When we are reading from a pipe and receive a signal, our
> read call fails and ferror() returns true. The current
> behavior is to call end_update and report failure. However,
> we can detect this situation by checking that errno is set
> to success and continue the reading process.

Thanks for the fix/workaround!

Show 5 quoted lines

> I am not convinced this is the right solution. Specifically:
>
>  - there are a few other calls to ferror. Maybe they should be
>    converted, too, which implies that perhaps there is a better idiom
>    for checking this.

Well, perhaps something like this can work around the issue for tig-0.12:

bool check_ferror(FILE *file) { return ferror(file) && errno != 0; }

For a possible "better" fix, I have been working on moving tig to use the run-command.[ch] code from git, which means that ferror() will no longer be needed. It is still not ready but looks promising.

Show 6 quoted lines

>  - I have no idea how portable this is. Do all stdio implementations
>    fail to restart on signal? Do they all set ferror and have errno ==
>    0 (I would have expected EINTR, or at the very least a 0-read
>    without ferror set)?
>
> But it works for me (Linux, glibc 2.7).

I have tested it on FreeBSD which doesn't seem to have this problem with ferror(). Anyway, the workaround doesn't break anything so applied.

-- 
Jonas Fonseca