Changeset View
Standalone View
requirements-swh.txt
# Add here internal Software Heritage dependencies, one per line. | # Add here internal Software Heritage dependencies, one per line. | ||||
swh.core >= 0.3 | swh.core >= 0.3 | ||||
swh.model >= 2.6.4 | swh.model >= 2.6.4 | ||||
ardumont: I guess it's temporary to make the ci build ok.
Why is the discovery code part implemented in… | |||||
Done Inline Actions
Yep
I feel like these pieces of code are are implementation details of the archive consumer and not something the storage (even client-side) should know about. I don't feel super strongly about it, it's just my intuition about where the code should live that may make more sense as it grows. Alphare: > I guess it's temporary to make the ci build ok.
Yep
> Why is the discovery code part… | |||||
Not Done Inline Actionsok. It feels strange to me to have the scanner consumer specifically depend on the loader.core. Maybe that means we need to push that code in swh.core which is the sole common part But same as you, i won't be too pushy. ardumont: ok.
It feels strange to me to have the scanner consumer specifically depend on the loader.core. | |||||
swh.loader.core @ git+https://forge.softwareheritage.org/source/swh-loader-core.git@1facea3cd215155f77ae4083a33837b7c6f642b0#egg=swh.loader.core | |||||
ardumontUnsubmitted Not Done Inline ActionsIf that commit is landed, we can release it so you can have the standard version number requirement here. ardumont: If that commit is landed, we can release it so you can have the standard version number… | |||||
No newline at end of file |
I guess it's temporary to make the ci build ok.
Why is the discovery code part implemented in the loader.core and not in the storage code (client side)?
(since it's used both by scanner and loader so far, that would have made sense).
Yes, i've asked it twice ¯\_(ツ)_/¯
[1] D8538#222652