From: Duy Nguyen <pclouds@gmail.com>
Date: Tue, 06 May 2014 00:30:17 GMT
Subject: Re: Watchman support for git
Message-ID: <CACsJy8B_NiL8LnYXk_O+tbsZxfNbg5EvkbbPtO-crHMmKdZhEA@mail.gmail.com>
In-Reply-To: <CACsJy8DWK9ysCJobh4oQZe2hxdTaE=n7hH=Nm6bfWzx+yJ=KZg@mail.gmail.com>

On Tue, May 6, 2014 at 7:26 AM, Duy Nguyen <pclouds@gmail.com> wrote:
> This discard_index() code was added
> recently in e28f764 (unpack-trees: plug a memory leak - 2013-08-13).
> As a workaround, perhaps we only do so when the sequencer is running.

Hmm.. as o->result does not carry cache-tree anyway, the next
assignment after discard_index() will "destroy" cache-tree anyway. I
wrote and deleted the following sentence, but it looks like I should
rewrite again. So if cache-tree has always been destroyed before, be
careful when you try to keep cache-tree as cache-tree invalidation
code in unpack-trees.c may not be well tested.
-- 
Duy