[GH-ISSUE #760] Create process for testing GCS and diagnosing incompatibilities #128

New issue

Open

opened 2026-03-03 12:08:34 +03:00 by kerem · 0 comments

kerem commented

2026-03-03 12:08:34 +03:00

Owner

Originally created by @fsouza on GitHub (Apr 25, 2022).
Original GitHub issue: https://github.com/fsouza/fake-gcs-server/issues/760

Every now and then someone opens an issue about something missing or not behaving the same as in GCS, and sometimes it's tricky to understand what GCS is doing, specially when it does nonstandard or undocumented things (for example, the Java SDK uses Content-Encoding to gzip the request, which is not HTTP-compliant, but we have to support it), and I always need to ask for more information or setup an ad-hoc bucket with some data and a script from scratch.

This is a meta issue to setup a process that will simplify this (maybe it's as simple as having a bucket and making sure that GHA has access to it, idk).

People will still need to provide snippets or something like that, but having that bucket will make things easier. Perhaps we could even have some sort of automation where we run pieces of code against the real server and fake-gcs-server and compare the results.

Obviously this will cost money, but as long as the cost isn't too high, I don't mind.

Originally created by @fsouza on GitHub (Apr 25, 2022). Original GitHub issue: https://github.com/fsouza/fake-gcs-server/issues/760 Every now and then someone opens an issue about something missing or not behaving the same as in GCS, and sometimes it's tricky to understand what GCS is doing, specially when it does nonstandard or undocumented things (for example, the Java SDK uses Content-Encoding to gzip the _request_, which is not HTTP-compliant, but we have to support it), and I always need to ask for more information or setup an ad-hoc bucket with some data and a script from scratch. This is a meta issue to setup a process that will simplify this (maybe it's as simple as having a bucket and making sure that GHA has access to it, idk). People will still need to provide snippets or something like that, but having that bucket will make things easier. Perhaps we could even have some sort of automation where we run pieces of code against the real server and fake-gcs-server and compare the results. Obviously this will cost money, but as long as the cost isn't too high, I don't mind.