pka
commented
3 years ago That's really great - one of the oldest feature requests! I didn't have time for a review yet, but generally I'm fine with a synchronous implementation. I'm planning to asyncify the seeder anyway and can have a look at the options then.
Regarding the CI failures:
- Do a
cargo fmt - Make the test with the real S3 access optional. Maybe you use an environment var for the S3 URL and skip the test when it is not set.
arbakker
See issue: https://github.com/t-rex-tileserver/t-rex/issues/169
Hi,
Here's an working implementation for S3 backed tilecache in t-rex. I am not sure if the solution in this PR is ready for master, mainly because of how the S3 write and read futures are resolved. This is done by using the sync method (available in the rusuto lib up to version 0.42). I suppose ideally the futures are resolved asynchronous, but I did not find a way how to do this (my experience with Rust is fairly limited).
Do you have an idea how to solve this, or what could be an acceptable solution?
Also the tests need to be improved to use a mock library, currently a S3 service is required to run for the test to run succesfully...
Will off course appreciate any other feedback on the PR as well.
Thanks for your consideration!