If you are talking about inline comments, then they are there in the comments section and you can see them using clicking on show. About the commits, it is there if the user didn't remove it. see the commits section

I think this is fixed, isn't it?

Basically, inline comments are no longer shown if the file was changed by a following revision, but the comment itself can still be found in the overview, listing all the comments.
I know it lacks context but that's not something pygit2 easily allows to extract/get so it's the best we have for the moment.

ping?

I'm going to close this, feel free to re-open if you think it should be.

Thanks for your report!

The commits are apparently lost because of the rebase. And the context is lost probably for the same reason. I'm afraid that dropping history of a pull request just because a rebase is not entirely good.

Feel free to close it if nobody's going to fix it. But the report is definitely not invalid.

The commits are apparently lost because of the rebase.

commits are not "lost", honestly. AFAIK, When you rebase (after correcting whatever was needed on existing commits in the pr), you create different commits. (git does, not exactly you)

About the history of PR, i guess better is possible like keeping the lines where the comments were actually made (the corresponding changed lines).

commits are not "lost", honestly. AFAIK, When you rebase (after correcting whatever was needed on existing commits in the pr), you create different commits. (git does, not exactly you)

I wasn't speaking about git, although it depends on your definition of “lost” when determining what happens with the git commits, as references to those commits are indeed lost and the commits themselves can be garbage-collected later. I was speaking about pagure and especially its user interface for a pull requst.

I don't know the internals of pagure and therefore I don't know what is possible and what's not... But I guess that (now obsolete) commits that have once been shown by pagure as commits belonging to that pull request could possibly be saved for later reference and history could be part of the pull request information.

Rationale: A contributer posted something and people reacted on it and even though the contributor is going to push-force a modified patchset, i.e. rebase the pull request, the history is part of the context of the previous discussion. Lack of context and disappearing information IMO break the communication attached to the pull request that should ultimately lead to a final patchset that gets accepted, or a conclution that nothing's going to happen.

Then this is likely a duplicate of #751 no?

Then this is likely a duplicate of #751 no?

Yep.

Let's close this one then :)

Thanks!

It looks like pagure doesn't support proper duplicates with displaying the information in both old and new bug so that you can look up information from all duplicate bugs if needed.