Re: cgit .tar.gz code snapshots

[fosslinux]

Hi Bruno,

On 12/18/24 13:05, Bruno Haible wrote:
> Hi,
>
> > we have been using downloads from cgit
> > on Savannah (https://git.savannah.gnu.org/cgit/gnulib.git, for instance,
> > https://git.savannah.gnu.org/cgit/gnulib.git/snapshot/gnulib-d271f86.tar.gz),
> > to get specific .tar.gz files of particular revisions of Gnulib.
> 'git' is the protocol that was designed for this purpose, and has
> the maximum efficiency (when you use it with --depth=1). So, that
> is the protocol that you should recommend.
I would typically agree. However, for our purposes, we do need a tar rather than a git repository eventually, and it is 
much more convenient/simpler for us to use a .tar.gz provided by upstream projects.However, we may need to revisit this.
> Since the checkbox in the Savannah admin UI carries the label
> "don't provide links to snapshot tarballs" it also controls
> the snapshot .tar.gz links, compare e.g. [4], [5].
Yes, this is the "download" link style we were using.
> If what you see is a recent change, it probably is an optimization:
> If nothing in the cgit UI displays a link with a certain form, there
> is no need for the cgit daemon to reply to that request.
>
> Maybe this has been disabled, because recently git.savannah.gnu.org
> has been overloaded with cgit requests, and what the sysadmins then
> do is to disable the requests that are abused by evil people and of
> less use to the public, in order to continue to be able to serve the
> requests that people rely on (namely the 'git' protocol).
Okay, thank you, I will clarify with Savannah staff.
> > I'd love to know if this disabling was intentional, and if so,
> > whether there are any problems with moving to downloading
> > from gitweb.
> It is possible that when the abusers notice that pulling snapshots
> via gitweb can equally damage the bandwidth of savannah, these snapshots
> will be disabled as well.
Absolutely. Understood. Some people have to make the Internet much more painful 
for everyone else :(
> > Also, in the (I hope,
> > unlikely) case that this was disabled because of downloads from live-bootstrap, 
> > I am very sorry.
> How many users of 'live-bootstrap' were downloading such a tarball, per day?

Difficult to quantify, but single digits on average I would guess, especially in recent months. There may have been 
spikes, but no chance it reached 3+ digits, unless there is someone using the project unreasonably (and wasting 
significant bandwidth) we are unaware of.

Thanks for your reply.

Samuel.