• bin/update add support for (re)installing packages after update
• bin/install: support user flags

rebase

• bin/install: support user flags

In T2336#42846, @DanSeraf wrote:

The problem here is that actually swh-model il calculating the persistent identifier of the input path, so it also has to exclude the specified paths.

@rdicosmo: the current version of the full example above LGTM (the surrounding text is inconsistent, e.g., it still mentions "snp" as a key and forbids snapshot anchors, but I suspect it's just that you didn't bother editing everything. Hence, we're good! :-))

the doc LGTM, but swh authentication feels like a mouthful, would it be possible to use swh auth instead? (which will also work for both "authorization", that I suspect is something we will also put behind this at some point)

(removed the last point, the hierarchy thing is in fact not relevant here, as we're pointing upward, not downward)

LGTM in general.

Now that it exists, do you want this to be part of the swh web client?

Thanks for the great review !

• client.py: make docstring start with uppercase letters

• requirements.txt: drop redundant dep on swh.core
• client.py: move getters table for .get() to an instance variable
• pytest fixture: remove left-over debug print
• tests: also test type-specific getters

• web client tests: add missing copyright headers

• Web API client: first implementation
• Web API client: add tests

Abandoned in favor of D2852, now in a dedicated repo

• web client: fix flake8 issues and update URL examples

• web client: handle symbolic refs in snapshots
• web client: handle snapshot pagination
• web client: make snapshot iteration return dictionaries

abandoned in favor of D2835

Hi Colin (@cjwatson), nice to meet you here !

Thanks a lot for this summary of a recurrent discussion we've had over the past few years now.

In D2791#66794, @swh-public-ci wrote:

Build is green
See https://jenkins.softwareheritage.org/job/DTSCN/job/tox/9/ for more details.

In D2784#66551, @vlorentz wrote:

This made me realize we didn't enable CI for this repo. I'll work on it.

In D2772#66241, @olasd wrote:

I think this was purely a diff readability helper (as our default is single quotes, it'd have changed almost every line to switch it). I agree that we should switch on string normalization as well.

+1 on the general idea.

perfect, thank you !