[GSoC] Discuss: Implement support for reftables in ‘dumb’ HTTP transport

Hello everyone.

I hope you are doing well. I am brainstorming on the problem statement "Implement support for reftables in ‘dumb’ HTTP transport" where I encountered some doubts which I wanted to discuss with you.

Problem statement link: https://git.github.io/SoC-2024-Ideas/ (Last one)

1. Do we just need support to know about the default branch or the
entire reftable?

2. As far as I am able to understand, I just need to add support for
these reftable files to transfer through the backend incase of dumb
http protocol, is this correct or am I missing something? In simple
words what I understand is I just need to add support for the server
side nothing is to be done for the client side.

I am sorry if these sound absolutely basic but I'm genuinely eager to learn and implement them.

Thank you.

Regards Aryan Gupta

Aryan Gupta <garyan447@gmail.com> writes:

Hmph. The "dumb" transport is kept primarily to allow folks, who can only use web hosting that serves nothing but static files, to publish their repositories, and requiring more things to be done on the "dumb" server side sounds somewhat backwards to me.

To be quite honest, I personally doubt that this topic makes much sense in this age and day---I've felt that the dumb HTTP walker outlived its usefulness for the past 10 years already. But perhaps I am biased by the first-world access to the internet?

Thanks.

On Wed, Mar 13, 2024 at 02:39:24PM -0700, Junio C Hamano wrote:

Even though I proposed this project in the first place I don't disagree with you. The dumb HTTP protocol is indeed quite esoteric nowadays, and if we want to put it officially into "maintenance" mode then I would not mind to drop that proposed project.

The reason why I added this project is that I found it to be interesting as a thought experiment. The reason why we have "info/refs" is to help out clients of the dumb HTTP transport to figure out actual refs in the repository because they typically live in many separate files. But what I realized is that we don't actually need it with the reftable format anymore because we basically already have a definitive list of all files that a client needs to download to acquire all refs: "tables.list".

So theoretically speaking we can implement support for dumb HTTP with reftables by having the client download "tables.list" and then fetch all the "*.ref" files listed in it. Whether that is sensible may be a different question.

Anyway. If we think that the dumb HTTP protocol extension is dumb then I'd rather want to pull the plug on the proposed GSoC project now rather than later to not lead down potential contributors the wrong path. But in that case I think we might want to make the deprecated state of that protocol official.

Patrick

Patrick Steinhardt <ps@pks.im> writes:

Yes, YES, YES!!!

It is not even theoretical---I think it is the RIGHT way to do the dumb HTTP walker if we had reftable from the beginning. Having a file at a known location where we can start "walking" from is a very powerful thing [*1*].

If we do not have to change the server side at all, which I think what you are saying, that removes the major part of the problem I was having with this proposal. I do not even have to worry about speaking with the first-world bias and hurting folks with needs that are only served by the dumb HTTP walker, either.

By the way, about two years ago, we have talked about making the first step to deprecate dumb HTTP walker [*2*], but given that nothing concrete materialized, we are ready to move further, yet.

[Footnote]

*1* With the filesystem backed refs, because there was no standard
    and widely supported way to ask a dumb HTTP server "what are the
    files and directories under this hierarchy?" and that was here
    the ugly "info/refs" hack came from.

*2* https://lore.kernel.org/git/f8639116d2d384a6d285c75830c52d8a8230ae6b.1647243509.git.ps@pks.im/