MutableContentMetadata.setContentType? But Jenkins has no API for detecting content type of files.

Ideally VirtualFile.toExternalURL would allow you to pass a “view mode” parameter requesting a content type…if S3 supports that.

RobustHttpClient.uploadFile also has no way of setting a content type, which is another blocker for this I think?

(this is actually a blocker for our team to use this plugin, "I can't see my test report when I click the link ")

I doubt RobustHttpClient is involved; the content type setting would be done on the master side using the AWS SDK. I think. Will not know for sure until a fix is implemented. Perhaps dcurrie wants to take a stab at it?

Unless I'm reading wrong, by https://github.com/jenkinsci/artifact-manager-s3-plugin/blob/master/src/main/java/io/jenkins/plugins/artifact_manager_jclouds/JCloudsArtifactManager.java#L115 and https://github.com/jenkinsci/artifact-manager-s3-plugin/blob/master/src/main/java/io/jenkins/plugins/artifact_manager_jclouds/JCloudsArtifactManager.java#L134, it works by getting a presigned S3 url and `uploadFile`ing to it with RobustHttpClient.

My reading is the same as akdor1154. The S3 API indicates that the default Content-Type on a PUT is binary/octet-stream unless you specify otherwise.

Right, the question is whether the URL itself can encode the desired content type or whether this needs to be handled as an HTTP header by the uploading code. In the latter case, RobustHttpClient.uploadFile would need to be amended to set a header.

(Then the next question is what content type to pick. Files.probeContentType seems like the right choice.)

Even if you only populated content type for most common things (txt, html), this would still be useful. Is this being actively worked on? This limitation makes using HTML reports through Jenkins web UI quite painful if this plugin is active. My assumption based on what I've seen thus far is latter option is the right one but it could be that I'm just missing some details of presigned URL's.

I took a stab at this, and was able to get a custom-built version of the plugin working for me locally: it uploads files to S3 with content-type set appropriately which makes them properly viewable in the browser (for inline file types such as text/plain and text/html) instead of always forcing them to be downloaded.

This took changes to both the artifact-manager-s3 and apache-httpcomponents-client-4-api-plugin repositories, since I had to ensure when the plugin performs an HTTP upload it also specifies the "Content-Type" header and I used a new method signature in RobustHTTPClient to do so.

My changes can be viewed at:

• https://github.com/jenkinsci/apache-httpcomponents-client-4-api-plugin/pull/24
• https://github.com/jenkinsci/artifact-manager-s3-plugin/pull/100

While, as I said, a custom plugin that has these changes appears to do the right thing for me and I also added several tests to verify uploads to (and subsequent downloads from) S3 are setting the content-type now, this is my first contribution to the project so I might well have missed something. Any suggestions or feedback welcome... I hope it at least provides a starting point for a fix to this issue, if nothing else.

Thanks! I will try to review soon.

Released in https://github.com/jenkinsci/apache-httpcomponents-client-4-api-plugin/releases/tag/apache-httpcomponents-client-4-api-4.5.10-2.0

No oleg_nenashev that was just an API.