I can't deploy to clojars ... 😱
Checksum validation failed, expected 306592ef671ef8c6f6f84e7341b9c8e417adcb01 but is f9a5fb105c8dd4b00a1fc335c8eaad8e8439eb44The full output:
$ bb publish --skip-bump
Deploying org.babashka/cli-0.2.10 to repository clojars as borkdude
Sending org/babashka/cli/0.2.10/cli-0.2.10.pom (1k)
to
Sending org/babashka/cli/0.2.10/cli-0.2.10.jar (5k)
to
Retrieving org/babashka/cli/maven-metadata.xml (1k)from
Checksum validation failed, expected 306592ef671ef8c6f6f84e7341b9c8e417adcb01 but is f9a5fb105c8dd4b00a1fc335c8eaad8e8439eb44
Retrieving org/babashka/cli/maven-metadata.xml (1k)from
Could not transfer metadata org.babashka:cli/maven-metadata.xml from/to clojars (): Checksum validation failed, expected 306592ef671ef8c6f6f84e7341b9c8e417adcb01 but is f9a5fb105c8dd4b00a1fc335c8eaad8e8439eb44
Execution error (ChecksumFailureException) at org.eclipse.aether.connector.basic.ChecksumValidator/validateExternalChecksums (ChecksumValidator.java:174).
Checksum validation failed, expected 306592ef671ef8c6f6f84e7341b9c8e417adcb01 but is f9a5fb105c8dd4b00a1fc335c8eaad8e8439eb44
Full report at:
/var/folders/j9/xmjlcym958b1fr0npsp9msvh0000gn/T/clojure-12973975526547264868.edn
Error while executing task: publish Publishing a new version also didn't work. Halp!
Also tried using lein deploy clojars but didn't work :(
I tried it here to see if I had the same issue but doesn't look like it, I managed to do a mvn deploy for com.lambdaisland/facai
yeah, something is screwed with this repo I'm afraid
Something is obviously wrong here:
$ curl -s | sha1sum
f9a5fb105c8dd4b00a1fc335c8eaad8e8439eb44 -
$ curl -s
306592ef671ef8c6f6f84e7341b9c8e417adcb01% Hmm. Looking now.
Can you share the contents of
that you see? I think this may be a CDN caching issue, because I get:
curl -s
f9a5fb105c8dd4b00a1fc335c8eaad8e8439eb44 I may need to purge the cache.
Wait, you just did share the contents, duh.
Looking at the caching settings now.
I purged the cache, can you try again?
We have logic to purge files on deploy, but that runs in a future w/o error handling, so if the request fails, you'll have stale data in any nodes that had the old versions cached. The node(s) you hit would have them cached since you deployed the prior version.
It worked!
Great! Sorry for the trouble.
Is there anything on my end that I should do to prevent this next time?
Unfortunately, no. I need to make the purging more robust on the server side. I'll create an issue to look at that.
Thanks a lot!
My pleasure!