From: Duy Nguyen <pclouds@gmail.com>
Date: Tue, 14 Jun 2016 09:21:22 GMT
Subject: Re: [PATCH 00/27] nd/shallow-deepen updates
Message-ID: <CACsJy8BAawP+599R7iKYcEa+DJ=Gc_Zv11ma5SSqA0eCki1p=g@mail.gmail.com>
In-Reply-To: <xmqqr3c1c8we.fsf@gitster.mtv.corp.google.com>

On Tue, Jun 14, 2016 at 12:10 AM, Junio C Hamano <gitster@pobox.com> wrote:
> Eric Sunshine <sunshine@sunshineco.com> writes:
>
>> I agree with Junio that moving the sigchain_pop() into the error
>> handling code-path, if possible, would be a nice improvement.
>
> Yeah, "if possible" is really what I was not sure about---is it safe
> to do the _push() thing before start_command(), which presumably
> would affect both the main and the forked processes?

Good point. But we do exec() very shortly in the forked process, which
should reset signal table back to default. So probably not a bad
thing.
-- 
Duy