Error 503 Response object too large from Gentoo git server

piggy · n00b Joined: 28 Sep 2015 Posts: 26

Sorry if this is the wrong section, then I want to understand if I'm wrong somewhere or the Gentoo Git server has some limitations or something or if it is even a bug.

I need this:

https://gitweb.gentoo.org/repo/gentoo.git/commit/sys-kernel/gentoo-sources?id=9028ce75da8404b77d0f92df381941a36119277f

and I get this:

Hu · Administrator Joined: 06 Mar 2007 Posts: 22601

Gitweb is a convenience to let you browse the git history without needing to download it locally. While it can return some git objects, you should not use it to download arbitrary large blobs. If you need a substantial portion of the git history, you should git clone the relevant repository and extract the objects locally.

That link works fine for me. It is just a removal of one ebuild.

pjp · Administrator Joined: 16 Apr 2002 Posts: 20476

The page loads fine for me. What is it you're trying to do / get?

I have on occasion received an error when trying to search for something. I can't say it is a 503 error, but it might be. I have presumed the search was taking too long.
_________________
Quis separabit? Quo animo?

piggy · n00b Joined: 28 Sep 2015 Posts: 26

Banana · Posted: Fri Jul 19, 2024 10:21 am Post subject:

as far as I can tell, it should be this feature: https://git-scm.com/docs/gitweb.conf#Documentation/gitweb.conf.txt-snapshot

So a better way would be to clone the project from https://gitweb.gentoo.org/repo/gentoo.git/ or https://github.com/gentoo-mirror/gentoo and produce your own snapshot https://stackoverflow.com/questions/615816/git-how-to-get-a-snapshot-of-a-git-repository

But cloning it localy makes the need of the snapshop rather obsolete. What exactly do you want to achive?
_________________
Forum Guidelines

PFL - Portage file list - find which package a file or command belongs to.
My delta-labs.org snippets do expire

piggy · n00b Joined: 28 Sep 2015 Posts: 26

Hu · Administrator Joined: 06 Mar 2007 Posts: 22601

If I interpret this right, you were not trying to get a single file. You were trying to get all files in the repository as of that commit, since you requested a snapshot of that commit. If I am right, that would explain why it was too big to serve to you. If you only want one file, you should download that file, raw. If you want a snapshot of that directory, you should focus your snapshot request on the relevant tree.

piggy · n00b Joined: 28 Sep 2015 Posts: 26

Hu · Administrator Joined: 06 Mar 2007 Posts: 22601

The link you posted requests that the server generate an archive of all the files in that commit, compress the archive, and serve it to you. Hence, I said you were trying to get all files in the repository as of that commit (>100k files). A test locally says that when I archive that commit and compress it with gzip, I get a 37M result. That does not seem huge, but since it is dynamically generated, I can see that the server administrators might have set a fairly low limit on maximum object size.

Banana · Posted: Fri Jul 19, 2024 12:26 pm Post subject:

pjp · Administrator Joined: 16 Apr 2002 Posts: 20476

piggy · n00b Joined: 28 Sep 2015 Posts: 26

pjp · Administrator Joined: 16 Apr 2002 Posts: 20476

Hu · Administrator Joined: 06 Mar 2007 Posts: 22601

OP appears to have reported this as https://bugzilla.kernel.org/show_bug.cgi?id=219061. OP was then directed to e-mail it to LKML. I cannot find that mail, but I see what appears to be the first response to that mail at https://lkml.org/lkml/2024/7/19/397.