I have changes for s3-0.7 at https://github.com/learningobjectsinc/s3-plugin/commit/a81561d8e663cd58b588bac24ff44c8dff3d18f9 to store the user's bucket entry instead of just the bucket name to the fingerprint record.

I first tried to update the latest master, but the upload had NPEs that were beyond this particular issue blocking me and I needed something that works.

Toby,

I can confirm this fix the issue here. As the bucketName is split by / and stored in bucket and objectname, then to access the file, the bucket from the FingerprintRecord is not enough to get access the file.

Could you create a PR with your code? Please rebase to head to make sure the PR can be merged.

Thanks!

The commit has been cherry-picked and now there is a PR with it: https://github.com/jenkinsci/s3-plugin/pull/58

Hi Adrien, thanks for doing the cherry pick and PR! Do you need anything else from me?

Code changed in jenkins
User: Nicolas De loof
Path:
src/main/java/hudson/plugins/s3/S3Profile.java
src/main/java/hudson/plugins/s3/callable/S3UploadCallable.java
http://jenkins-ci.org/commit/s3-plugin/ab3afa1be121fd90693b23f353e915ec552da98c
Log:
Merge pull request #58 from ndeloof/JENKINS-26297

use the user bucketname entry that includes prefixes in fingerprint

Compare: https://github.com/jenkinsci/s3-plugin/compare/c6308fef7279...ab3afa1be121

If I understand the code commit (just by reading it), it appears that this will save the bucket as "my-bucket/key-prefix" in the build.xml file. What about builds before this fix is published? All of our builds just save "my-bucket" to the build.xml rather than "my-bucket/key-prefix," but the files are all saved to "my-bucket/key-prefix."