thanks !

In T2288#42054, @olasd wrote:

Your swh-identify backtrace looks like a $PATH caching issue (see also the path of the swh-identify script in the first line). I can't reproduce it in a fresh venv and swh-identify works fine.

a draft implementation of this idea is in D2461

Given this is done, where can one see the timeline of visits for a given origin coming from GNU?

In D2623#62446, @vlorentz wrote:

"Missing" isn't perfect, maybe, but it's consistent with SQL storage tables at least.

No, SQL tables use "skipped". But I prefer "missing" because it's more generic (it also includes content we couldn't find)

s/non missing/present/
would be an improvement.

tnx for debugging and fixing this!

I'm fine with switching to IRIs in the doc, just please expand what it means on first use (with a mention like "they are like URIs but"), as I don't think the acronym is that well-known yet, especially in the US.

In T1791#41498, @anlambert wrote:

I propose to remove those id leaks from that endpoint and contact OpenAIRE to tell them to use the Link header from now on to paginate
the results.

@anlambert is this done and, if so, can you close it?

this has been implemented by @DanSeraf in D2582

thanks Morane!

For the time being we do not want to leak origin PIDs to the world, as they're not intrinsically computable on the origin itself (but only on the URLs, which is not the artifact itself, which is not digital, but just a locator for it). Let's consider ori PIDs only an internal shorthand useful in places where having fixed-length shorthands for full URLs is needed (e.g., swh-graph).

In D2461#59540, @douardda wrote:

IMHO this should be a 'standalone' repo/project. I mean it should be possible to pip install it at least.

In D2492#59143, @vlorentz wrote:

I'm not a big fan of using absolute URLs. It makes the code more verbose, tests more complex, and prevents clients from using proxies. Why do we want/need that?

(marking as done as it was moved to the done column on the sprint board, please reopen if not ok)

(marking as done as it was moved to the done column on the sprint board, please reopen if not ok)

Turns out this was due to a left-over username/password in my ~/.netrc, which apparently requests uses by default and curl doesn't.
(But why the heck the caching effect? I've no idea...)

• webclient: add missing error checking

• simplify code and streamline returned snapshot types

In T2147#39796, @anlambert wrote:

@zack, this is now deployed to production.

Darn, sorry, I've looked for a dupe before submitting, but I obviously failed at that. Thanks for closing the duplicate.