Re: [RFC/PATCH 3/3] receive-pack: allow a maximum input size to be specified

From: Christian Couder <christian.couder@gmail.com>
Date: Aug 16, 2016, 08:25 UTC
Message-ID: <CAP8UFD2Bpp-gcxEQGNSFxvKyptkyeAsMtaHZaBebJO2XiF+W3g@mail.gmail.com>
In-Reply-To: <20160816010309.zstrljqknc3vjziq@sigill.intra.peff.net>

On Tue, Aug 16, 2016 at 3:03 AM, Jeff King <peff@peff.net> wrote:

Show 20 quoted lines

> On Mon, Aug 15, 2016 at 03:48:55PM -0700, Junio C Hamano wrote:
>
>> Jeff King <peff@peff.net> writes:
>>
>> > The simple fix is to call register_tempfile() in open_pack_file(), and
>> > just have index-pack clean up the file on its way out.
>> >
>> > But there are harder cases. For instance, imagine somebody pushes a
>> > 500MB file, and you have a pre-receive hook that says "too big; I won't
>> > accept this". And then they push in a loop, as before. You've accepted
>> > the incoming pack into the repository by the time the pre-receive runs.
>> > You can't just delete it, because you don't know if other simultaneous
>> > processes have started to depend on the objects.
>> >
>> > To solve that, I have patches that put incoming packfiles into a
>> > "quarantine" area, then run the connectivity check and pre-receive hooks
>> > with the quarantine accessible via GIT_ALTERNATE_OBJECT_DIRECTORIES. And
>> > then we either move the quarantine packs into the real repo, or blow
>> > away the tmpdir, depending on whether the hooks said the objects were
>> > OK.

Thanks Peff for the review and all these explanations and suggestions!

>> > Those are patches I plan to share upstream but just haven't gotten
>> > around to yet.

I would be happy to help if I can on these patches too!

>> I think these other patches can come later, independent from this
>> three-patch series resurrected from the archive, so I can take a
>> reroll of these once the integer-size issues you pointed out are
>> sorted out.

Ok, I just sent a non RFC reroll of the series with those issues hopefully fixed.

> Yeah, definitely it's independent. I was mostly suggesting to Christian
> that he may want to look into the "easy" register_tempfile() case, as
> GitLab may find this is only half of the fix that they want.

Yeah, thanks again for this suggestion!

> Also a patch I may try to polish and share in the future, but not as
> high priority as some of the other stuff.

I can help polishing this patch if you want.

Thanks, Christian.