merged in abbcf03b7bb2f1425db154dbe6e43e10c647354c

This is now done: swh-users@inria.fr

This has now been discussed on the sourcehut mailing list and I took part in the conversation.

For future reference, it looks like we are still "small" players as "big" packages go on PyPI: https://pypi.org/stats/ (e.g., tf-nightly is currently the largest package on PyPI and it weights 427 GiB).
While it is still not nice to ship a big fat JAR in a PyPI package, our extension requests will likely be granted.

In T3560#85153, @KShivendu wrote:

Hey @vlorentz @zack, I've been using sourcegraph.com for almost a year now and I feel that they have worked a lot on polishing their search query language. I think we can learn from them and adapt our language. Here are a few suggestions:

Monumental documentation work, thanks!
I think this is generally great, and I've pointed out only some minor issues/suggestions here and there.

In D7814#203592, @seirl wrote:

In D7814#203336, @zack wrote:

I'm fine with this code cleanup, with one caveat: that we document/ship the systemd startup service (and its meaning, including some intuitions about the trade-offs you mention) somewhere, in replacement of the cachemount command.

It's already in puppet (swh-site/site-modules/profile/templates/swh/deploy/graph/swhgraphshm.service.erb).

I'm fine with this code cleanup, with one caveat: that we document/ship the systemd startup service (and its meaning, including some intuitions about the trade-offs you mention) somewhere, in replacement of the cachemount command.

Just a nitpick: either add thousand separators to the node/edge counts, or summarized them with M/B suffixes.
Rationale: those numbers are so huge that are hard to read without that.

In D7192#187159, @vlorentz wrote:

Sorry, that is a bit rambly and not very helpful. @anlambert @zack What do you think?

Then let's just go for it (insert here ref. to upcoming separate task :-)).

In T3887#77949, @olasd wrote:

Now that I've written it out loud, of course, Releases don't have extra_headers so the package loaders can't make use of this workaround/hack for now.

@marla.dasilva @anlambert: let's go for "Until Jan 30th" then. (I'll also ping you about this in the chat, just in case.)

Thanks Marla, I also planned to raise this.

Just to be sure: test_pipeline() from test_cli.py is now run with all new passes as well, and as such it also testes the LLP step(s), correct?
It seems that way to me because test_pipeline() seems to be running all passes, but I'd like this to be double-checked before landing.