Skip to content

Merge Request Version

Description including problem, use cases, benefits, and/or goals

To be able to review different changesets, it should be possible to change the comparison in a merge request.

TODO

Proposal

See latest proposal in comment https://gitlab.com/gitlab-org/gitlab-ce/issues/13570#note_12909009 and https://gitlab.com/gitlab-org/gitlab-ce/issues/13570#note_13002460 with example https://gitlab.com/gitlab-org/gitlab-ce/issues/13570#note_12909564. Feel free to contact @dzaporozhets during implementation for questions/details.

OLD Proposal

DZ explains it best in this comment.

We create a path that allows you compare the different commits in the merge request and a set of dropdowns to change this.

For example, merging the branch feature-1 into master:

  • master's commit A
  • feature-1 has commits B and C, where C is the latest

Right now we only compare A...C.

We would introduce a path and dropdowns that allow you to change this to:

  • A...B (see what has changed between master and the first commit on the source branch)
  • B...C (see what has changed between the two commits on the source branch)
  • A...C (original)

coms

The dropdown would should the commit hash and a bit of the commit message, so you know what you're looking at.

Reviewboard does a nice thing, where they call A the orig, so that it's clear where the target is.

Links / references

Original issue

ZD: https://gitlab.zendesk.com/agent/tickets/15786

Description

Currently to get a Merge Request we use GET /projects/:id/merge_requests/:merge_request_id. Is it possible to add the following syntax?

GET /projects/:id/merge_requests/:merge_request_id/:merge_version_id

Adding a merge version ID at the end while the head reference still points to the latest version without any changes.

The benefit would be that different versions of a MR can be compared very easily even if the MR is squashed / rebased etc which loses normally any signs of changes in between (compared to new commits all the time).

Reference

What do you think @V@vsizov?