Redirection heuristic for #113 - #114
Conversation
|
@Mrmaxmeier Thanks for tackling this! Do you understand the problem that's occurring in #113, then? I've been puzzled since I simply don't get that amount of redirection when I run the program myself. It doesn't seem like the redirects should be location-dependent, but I don't know a lot about how modern CDNs work. |
|
@pkgw I guess Bintray only redirects to a CDN if it knows of a server that's in the general location of the request IP. For reference: Here's the chain of redirects I get when resolving the URL: |
This fix is needed for Rust < 1.19 (see rust issue #39817)
|
Yes, this is it. And yes, this fixes it. Good job and -- thanks. 🍻 |
|
Ah, interesting. I've been blocking the 0.1.6 release on writing a skeleton manual but this seems like an important enough fix to be worth cutting a release even if I can't complete that task, which is looking like it's going to take a long time due to non-Tectonic obligations. I will aim to do so this weekend. Thanks for the very rapid fix! |
I've tried to keep the heuristic general and avoid false-negatives as it only interferes if the last segment of the URL doesn't contain a
.(This will separates hash-like URLs from ones that contain a file extension)
Closes #113