Re: git 2.16.0 segfaults on clone of specific repo

From: Duy Nguyen <pclouds@gmail.com>
Date: Jan 19, 2018, 05:31 UTC
Message-ID: <CACsJy8BTFm_0sv=roL1OKKW=1DyU3vqD50NKyHg3KQ7G+mAepQ@mail.gmail.com>
In-Reply-To: <20180119034025.GB222163@genre.crustytoothpaste.net>

On Fri, Jan 19, 2018 at 10:40 AM, brian m. carlson <sandals@crustytoothpaste.net> wrote:

Show 20 quoted lines

> On Thu, Jan 18, 2018 at 10:06:10PM -0500, Eric Sunshine wrote:
>> > I have a guess about what the problem might be.  Can you try this patch
>> > and see if it fixes things?
>>
>> That does fix the crash. Thanks for the quick diagnosis.
>>
>> Can the commit message go into more detail as to why this was crashing
>> (or your speculation about why)? Perhaps give more detail about what
>> 'clone' is doing that led to the crash.
>
> Sure.  I ran into this as I was expanding the hash structure abstraction
> into my next series.  I'll send a follow-up patch with a more
> descriptive answer.
>
> I'm still extremely puzzled as to why this doesn't fail on Linux.  If
> it's failing in this case, it should very, very clearly fail all the
> time we access an empty blob or an empty tree.  I've tried with gcc and
> two versions of clang, using -fno-lto, with address sanitizer, with -O0,
> and so on.  I'd really like to catch this kind of issue sooner in the
> future if I can figure out how to reproduce it.

I think it's file system related, case-insensitive one in particular.

The call trace posted at the beginning of this thread should never trigger for an initial clone. You only check if an _existing_ entry matches what you are about to checkout when you switch trees. For this to happen at clone time, I suppose you have to checkout entry "A", then re-checkout "A" again. Which can only happen on case-insensitive file systems and a case-sensitive repo where the second "A" might actually be "a".

vim-colorschemes.git has these two entries, which meets one of the conditions, I suppose macos meets the other

colors/darkblue.vim colors/darkBlue.vim

On Linux, after I hacked all over the place to force ce_match_stat() to eventually call index_fd() which in turns must use one of the hashing function, I got a crash.

-- 
Duy