Dustin Ingram
Dustin Ingram
There's enough of these that we should probably just detect and handle the encoding, I don't think a spec update is specifically needed. ``` warehouse=> select count(*) from releases where...
Additionally, the locale negotiation on the backend might be unnecessary in most cases: if we just use the `PyPI-Locale` header set by Fastly (assuming this is set for all requests,...
Short term, instead of exposing this to end users to configure, we could do this by default for GitHub Actions to prevent exfiltration.
Agreed, I think we should probably just hide it when there's only one wheel, rather than whether that wheel is pure or not.
I think there's a benefit to both testing the specs and testing the application as a black box (via CLI). Can we agree that it's worthwhile to do both?
Agree wholeheartedly with @tetsuo-cpp, one point though: > What's the appeal of running them centrally? I think one advantage would be getting good insight into whether the 'online' services have...
(Sorry, cc @kairoaraujo as well!)
#8586 has been merged, next step on our end here is setting up the CDN to serve the hashed pages from storage.
`tuf==1.0.0` has been released: https://pypi.org/project/tuf/1.0.0/
> XML-RPC (N/A?) I think it would be wise to not bother porting these docs. > Integration guide (N/A?) I think this can probably get included in the top-level page...