PR review comments not showing up in "Files changed" (and are not "Outdated")

[alexjpwalker]

Generally, when writing a PR review comment on a specific line of code, it becomes visible to anyone viewing the "Conversation" or "Files changed" tabs of the PR. In "Files changed", the comment shows up immediately below the line(s) of code it was written against.

However, posted comments intermittently do not show up at all. We know this occurs if the comment goes "Outdated" due to code changes, but this is not happening in this case - the comment simply disappears without trace, with no further code changes.

One common scenario I've experienced (for example) is, when adding a review comprising multiple comments, the first one I write sometimes fails to show up in "Files changed". It exists, and can be accessed from the small "Conversations" button on the "Files changed" tab, but that is a significantly worse UX than seeing the thread at the relevant code line - highly frustrating during the review process.

[alxndr-w]

It also looks like comments are only visible to me as a reviewer, as long as i don't finish my review.

This results in bad UX because while I thought, my comments and code change suggestions are visible to the PR creator, in fact, they were only visible to me, as long as I don't finish my review.

The drafted comments should be somehow be different visually (grayed out, not 100% opaque, or a flag / badge) as long as the review was not submitted.

[kroggen]

That is a good idea, to show the comments with a different color somewhere while the review was not submitted. Or display that they are not yet visible to other users.

In our company we had problems related to this twice, because people do not know that single comments are not appearing to others while they don't click on finish review, but their thought is "I will finish the review once the PR creator answers this question" but the question is not visible to others!!!

[kroggen]

The current UX is not intuitive! Just on our company 2 different developers already did the exact same thing described above.

They write a question linked to code and instead of clicking on "Add a single comment" they click on "Start a review", but it is not directly clear that the written messages are only visible to themselves!

This causes misunderstandings and make us lose a lot of time

[Lasse-BeCause]

Agreed, very confusing, as there is nothing that indicates that the comment is only visible to the reviewer.

[PureGero]

Agreed. I just did the same thing and spent ages trying to debug why my coworker couldn't see my comment.

[michaelloo]

I experienced exactly the same problem as the original post:

One common scenario I've experienced (for example) is, when adding a review comprising multiple comments, the first one I write sometimes fails to show up in "Files changed". It exists, and can be accessed from the small "Conversations" button on the "Files changed" tab, but that is a significantly worse UX than seeing the thread at the relevant code line - highly frustrating during the review process.

And I had submitted my review, so other comments from the same review would be shown in "Files changed" but not the first one and it was not marked as "Outdated" and no code was updated.

I'm not sure if it is related but the Sys Engineers in my company had updated GHE from 3.8.8 to 3.8.10 a week or two before this started happening.

[Zebiano]

I'm also experiencing the same issue: comments appearing in the "Conversation" tab, but not all of them appearing on the "Files changed" tab.

Highly frustrating, as I'd go through all the comments on the "Files changed" tab (since they're easier to grasp there because of having the rest of the code as context), only to go back to the "Conversation" tab when I'm done and realise there are still comments left unresolved.

The weird part is when in the "Files changed" tab, there's a small "Conversations" button at the top that lists all the comments left unresolved. If you click on one, you'll automatically get scrolled to where the comment is. But when you try to click on a comment that is not showing up, clicking it does absolutely nothing. But it does exist, as it's part of the list of comments left. So there's definitely something wrong here.

[NoureddinBakir]

Issue is persistent with one of my PR's as well. I saw my coworker's screen as having placed the comments but even after clearing cookies, cache, logging out and in numerous times the comments are no where to be found.

[JetFault]

Same is happening for us on GHE Server 3.8

[xepelinalf]

I have the exact same problem; I cannot see the responses to my own comments. How is this still happening?!

[MatheusIam]

As far as I can tell, it seems like you need to finalize the review based on the information provided above

It also looks like comments are only visible to me as a reviewer, as long as i don't finish my review.

[nazreen]

Just realised the past few times when my dev said " I can't see the comments" it was due to this too.

please fix.

[Daerst]

I kinda see that for larger PRs it might be worth adding all the comments before showing them to the developer in question, so that they don't start reworking code based on incomplete information. Still, I think this needs to be made clearer to the commenter.

[filipbatur]

Had this happen to us on several occasions, the UX is completely misleading.

[samakshkhatri-crest]

This process of "starting" and "finishing" a review seems does not intuitive at all. As a reviewer, my review should be considered "finished" when I provide an approval.

Also it's unusual that PR authors cannot approve their own authored PR even if they are added as a reviewer and the assignee is someone else. This should take precedence over author restrictions. Neither Gitlab or Bitbucket restrict this inherently, of course approval rules can be configured separately but those are optional.

[RazorBach]

This process of "starting" and "finishing" a review seems does not intuitive at all. As a reviewer, my review should be considered "finished" when I provide an approval.

Also it's unusual that PR authors cannot approve their own authored PR even if they are added as a reviewer and the assignee is someone else. This should take precedence over author restrictions. Neither Gitlab or Bitbucket restrict this inherently, of course approval rules can be configured separately but those are optional.

+1. github please consider improving this.

[RazorBach]

For anyone expericing similar issue. Please follow the steps in the offical guide to send out review comments.
especially step 10

When you're done, click Start a review. If you have already started a review, you can click Add review comment.
Before you submit your review, your line comments are pending and only visible to you. You can edit pending comments anytime before you submit your review. To cancel a pending review, including all of its pending comments, click Review changes above the changed code, then click Abandon review.