#2927: comment id model update to long instead of int #2928

arxange1 · 2024-06-04T23:22:24Z

Resolves #2927

Before the change?

GetIssueComments() leads to System.OverflowException: Value was either too large or too small for an Int32.

After the change?

GetIssueComments() returns comments.

Pull request checklist

Tests for the changes have been added (for bug fixes / features)
Docs have been reviewed and added / updated if needed (for bug fixes / features)

Does this introduce a breaking change?

Please see our docs on breaking changes to help!

Yes
No

Bazeloth · 2024-06-05T06:42:09Z

If i could approve this i would. This is holding us back big time.

Edit:
What the hell i can? I'll patiently wait for an official approval though.

kevinobruno · 2024-06-05T12:31:07Z

I think we also need to change IIssueCommentsClient and its implementation IssueCommentsClient?

arxange1 · 2024-06-05T12:45:26Z

I think we also need to change IIssueCommentsClient and its implementation IssueCommentsClient?

Indeed, thank you!

ioshm · 2024-06-05T13:25:49Z

I believe RepositoryCommentsClient & its interface should also be updated?

arxange1 · 2024-06-05T13:53:03Z

I believe RepositoryCommentsClient & its interface should also be updated?

Please check

ioshm · 2024-06-05T14:01:11Z

Octokit/Clients/ICommitCommentReactionsClient.cs

Looks like a few comment id/number parameters were not updated to long in the other methods.

ioshm · 2024-06-05T14:03:06Z

Octokit/Clients/IIssueCommentReactionsClient.cs

Looks like a few comment id/number parameters were not updated to long in the other methods.

ioshm · 2024-06-05T14:05:14Z

Octokit/Clients/IPullRequestReviewCommentReactionsClient.cs

Looks like a few comment id/number parameters were not updated to long in the other methods.

ioshm · 2024-06-05T14:06:07Z

Octokit/Clients/IPullRequestReviewCommentReactionsClient.cs

@@ -75,7 +75,7 @@ public interface IPullRequestReviewCommentReactionsClient
        /// <param name="commentId">The issue id</param>


Documentation is wrong here?

ioshm · 2024-06-05T14:07:24Z

Octokit/Clients/IPullRequestReviewCommentReactionsClient.cs

@@ -85,6 +85,6 @@ public interface IPullRequestReviewCommentReactionsClient
        /// <param name="commentId">The issue id</param>


Documentation is wrong here?

ioshm · 2024-06-05T14:08:04Z

Octokit/Clients/IRepositoryCommentsClient.cs

@@ -18,17 +18,17 @@ public interface IRepositoryCommentsClient
        /// <remarks>http://developer.github.com/v3/repos/comments/#get-a-single-commit-comment</remarks>
        /// <param name="owner">The owner of the repository</param>
        /// <param name="name">The name of the repository</param>
-        /// <param name="number">The comment id</param>
+        /// <param name="commentId">The comment id</param>


Not sure if we want to rename this to commentId. Perhaps the maintainers should chime in on this one.

It is a breaking change fix anyway. I would like to see the maintainer here to reduce the id/number confusion around the code.

ioshm · 2024-06-05T14:09:36Z

Octokit/Clients/IRepositoryCommentsClient.cs

Looks like a few comment id/number parameters were not updated to long in the other methods.

number should stay as int. The fix is about comment id only. I believe there is a huge mess around numbers/ids, and I changed it to long in all places where I think it is comment_id actually.

Unless I am mistaken, number do not exists on comments, and only have ids which should be int64 by now. It seems number and id are used interchangeably when it comes to comments.

Take for example this method

Definition:

octokit.net/Octokit/Clients/IRepositoryCommentsClient.cs

Lines 123 to 131 in fe329e7

/// <summary>

/// Updates a specified Commit Comment.

/// </summary>

/// <remarks>http://developer.github.com/v3/repos/comments/#update-a-commit-comment</remarks>

/// <param name="owner">The owner of the repository</param>

/// <param name="name">The name of the repository</param>

/// <param name="number">The comment number</param>

/// <param name="commentUpdate">The modified comment</param>

Task<CommitComment> Update(string owner, string name, int number, string commentUpdate);

Implementation:

octokit.net/Octokit/Clients/RepositoryCommentsClient.cs

Lines 209 to 225 in fe329e7

/// <summary>

/// Updates a specified Commit Comment.

/// </summary>

/// <param name="owner">The owner of the repository</param>

/// <param name="name">The name of the repository</param>

/// <param name="number">The comment number</param>

/// <param name="commentUpdate">The modified comment</param>

/// <remarks>http://developer.github.com/v3/repos/comments/#update-a-commit-comment</remarks>

[ManualRoute("PATCH", "/repos/{owner}/{repo}/comments/{comment_id}")]

public Task<CommitComment> Update(string owner, string name, int number, string commentUpdate)

{

Ensure.ArgumentNotNullOrEmptyString(owner, nameof(owner));

Ensure.ArgumentNotNullOrEmptyString(name, nameof(name));

Ensure.ArgumentNotNull(commentUpdate, nameof(commentUpdate));

return ApiConnection.Patch<CommitComment>(ApiUrls.CommitComment(owner, name, number), new BodyWrapper(commentUpdate));

}

The number parameter still has the type int when it should be long. The docs states that the comment_id (in this case, this parameter is called number) is of type integer which means any double without a fractional part - long should be good enough.

nickfloyd · 2024-06-05T20:01:28Z

Hey y'all,

First off, thank you for tackling this and thank you for moving forward on these updates to make things work/better - apologies for the trouble. This was fallout from a recent change. You are correct this issue and ones like it are why we are moving to a fully generated SDK.

It looks like the tests need to be updated as well before we can ✅ and get this merged in and shipped.

To answer the question regarding scope of change right now the only ID which cascades across a number of SDK APIs is comment_id, which in some cases is simply named number as a parameter. Changing the name of this parameter and it's type is the right move here. Especially since this will be releases as a breaking change.

Side note: The newer, generated .NET SDK has already accounted for this. That SDK gets generated nightly against the latest OpenAPI description from the GitHub REST API.

nickfloyd · 2024-06-05T20:05:41Z

If i could approve this i would. This is holding us back big time.

Edit: What the hell i can? I'll patiently wait for an official approval though.

Hey @Bazeloth While you can "approve" - because we want/need community involvement and input the merge criteria would not be met given that requires an approval from a core maintainer/owner. We do take those ☑️ checks seriously as we are going through approval though ❤️ .

BrianCArnold · 2024-06-06T01:08:35Z

It looks like there's a lot more places where int number should actually be long commentId, I'm going to see if I can patch all of them, looks like a bit of work.

BrianCArnold · 2024-06-06T02:33:17Z

@arxange1 @nickfloyd

Take a look at https://github.com/BrianCArnold/octokit.net/tree/bugfix/%232927
I fixed what I think are the remainder of all the Int32 commentId and reactionId issues, and all the tests should be passing now.

Just cherry-pick that commit onto this branch if it looks good, I'm trying to get my bot working again.

Bazeloth · 2024-06-06T08:21:02Z

If i could approve this i would. This is holding us back big time.
Edit: What the hell i can? I'll patiently wait for an official approval though.

Hey @Bazeloth While you can "approve" - because we want/need community involvement and input the merge criteria would not be met given that requires an approval from a core maintainer/owner. We do take those ☑️ checks seriously as we are going through approval though ❤️ .

Thanks that explains a lot. I was just wondering, since in our repo people that can approve usually can merge as well. Good to see you guys took precautions 👍

…sts fix

…okit.net into bugfix/octokit#2927

BrianCArnold · 2024-06-06T21:59:01Z

... did you copy my changes into your own commit, or just do (virtually) the same thing?

arxange1 · 2024-06-06T22:03:01Z

... did you copy my changes into your own commit, or just do (virtually) the same thing?

I made the changes independently and didn't see your message before doing so.

BrianCArnold · 2024-06-06T22:23:22Z

Ahh, gotcha, yeah, I looked through and I see that as well. There's actually several fixes that I made that look like they got missed, let me see if I can make a patch to fix them as well.

BrianCArnold · 2024-06-06T22:27:20Z

(Also, you fixed several things I missed as well)

BrianCArnold · 2024-06-07T14:28:33Z

If we're doing a breaking change on this release to avoid confusions on parameter names, such as replacing "number" with "commentId", I'd like to include fixing as many of them as possible.

Should we include them on this PR?

(https://github.com/BrianCArnold/octokit.net/tree/FixParameterNames)

nickfloyd · 2024-06-07T21:46:58Z

Should we include them on this PR?

Normally I'd be a little more risk adverse, but given the scope of change here (an id field, the no longer matches the backing type + breaking change - I can definitely get behind a little bigger changeset to resolve these issues in one go.

nickfloyd · 2024-06-07T21:58:58Z

I'd like to land this as soon as possible. It looks like @arxange1 had pulled in a few changes from other sources. I'll give it another day here but after that we'll accept more PRs to address this and at least get these fixes shipped out on Monday unless anyone has any issues where this should be held up.

arxange1 · 2024-06-09T08:20:05Z

If we're doing a breaking change on this release to avoid confusions on parameter names, such as replacing "number" with "commentId", I'd like to include fixing as many of them as possible.

Should we include them on this PR?

(https://github.com/BrianCArnold/octokit.net/tree/FixParameterNames)

Done

jeny-amatya-qed · 2024-06-10T00:10:14Z

Any ETA on the new version with this fix?

nickfloyd · 2024-06-10T13:08:15Z

Any ETA on the new version with this fix?

Hey @jeny-amatya-qed I am working on a release right now... I'll let you all know when it drops.

nickfloyd · 2024-06-10T13:24:23Z

These changes have been released - https://www.nuget.org/packages/Octokit

dotnet add package Octokit --version 12.0.0

jessehouwing · 2024-06-10T20:34:30Z

This applies to other models as well I found today. My guess is that any int id should be a long id.

See #2931

Old Octokit versions used int rather than long for Comment.Id; which causes overflow error when parsing responses from the GitHub Comments API. See octokit/octokit.net#2928

octokit#2927: comment id model update to long instead of int

4f3c8e8

bucataruelena approved these changes Jun 5, 2024

View reviewed changes

octokit#2927: code review fixes (1)

4b2e677

octokit#2927: code review fixes (2)

fe329e7

ioshm reviewed Jun 5, 2024

View reviewed changes

nickfloyd added Type: Breaking change Used to note any change that requires a major version bump Type: Bug Something isn't working as documented labels Jun 5, 2024

Victor Vorobyev and others added 4 commits June 6, 2024 23:54

octokit#2927: comment id model update to long instead of int: unit te…

3022734

…sts fix

Merge branch 'main' into bugfix/octokit#2927

1977413

octokit#2927: code review fixes

f5d035b

Merge branch 'bugfix/octokit#2927' of https://github.com/arxange1/oct…

10dc141

…okit.net into bugfix/octokit#2927

dcairney approved these changes Jun 6, 2024

View reviewed changes

premun mentioned this pull request Jun 7, 2024

Maestro keeps commenting on PRs with failed policies dotnet/arcade-services#3601

Closed

missymessa mentioned this pull request Jun 7, 2024

Production - [Alerting] DotNetEng Status Failed Requests/Hour alert dotnet/dnceng#3003

Closed

Fixed most names of parameters

b52b57c

nickfloyd mentioned this pull request Jun 10, 2024

[BUG]: Int32 overflow for Github object ids #2893

Closed

1 task

nickfloyd approved these changes Jun 10, 2024

View reviewed changes

nickfloyd merged commit 6c43183 into octokit:main Jun 10, 2024
5 checks passed

Jericho mentioned this pull request Jun 20, 2024

Value was either too large or too small for an Int32 GitTools/GitReleaseManager#598

Closed

Mankarse mentioned this pull request Jul 1, 2024

Update Octokit 10.0.0 -> 13.0.0 ZeroK-RTS/Zero-K-Infrastructure#2987

Merged

		@@ -75,7 +75,7 @@ public interface IPullRequestReviewCommentReactionsClient
		/// <param name="commentId">The issue id</param>

		@@ -85,6 +85,6 @@ public interface IPullRequestReviewCommentReactionsClient
		/// <param name="commentId">The issue id</param>

	/// <summary>
	/// Updates a specified Commit Comment.
	/// </summary>
	/// <remarks>http://developer.github.com/v3/repos/comments/#update-a-commit-comment</remarks>
	/// <param name="owner">The owner of the repository</param>
	/// <param name="name">The name of the repository</param>
	/// <param name="number">The comment number</param>
	/// <param name="commentUpdate">The modified comment</param>
	Task<CommitComment> Update(string owner, string name, int number, string commentUpdate);

#2927: comment id model update to long instead of int #2928

#2927: comment id model update to long instead of int #2928

Conversation

arxange1 commented Jun 4, 2024 • edited Loading

Before the change?

After the change?

Pull request checklist

Does this introduce a breaking change?

Bazeloth commented Jun 5, 2024 • edited Loading

kevinobruno commented Jun 5, 2024

arxange1 commented Jun 5, 2024

ioshm commented Jun 5, 2024

arxange1 commented Jun 5, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ioshm Jun 5, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nickfloyd commented Jun 5, 2024

nickfloyd commented Jun 5, 2024

BrianCArnold commented Jun 6, 2024

BrianCArnold commented Jun 6, 2024

Bazeloth commented Jun 6, 2024

BrianCArnold commented Jun 6, 2024

arxange1 commented Jun 6, 2024

BrianCArnold commented Jun 6, 2024

BrianCArnold commented Jun 6, 2024

BrianCArnold commented Jun 7, 2024

nickfloyd commented Jun 7, 2024

nickfloyd commented Jun 7, 2024

arxange1 commented Jun 9, 2024

jeny-amatya-qed commented Jun 10, 2024

nickfloyd commented Jun 10, 2024

nickfloyd commented Jun 10, 2024

jessehouwing commented Jun 10, 2024

arxange1 commented Jun 4, 2024 •

edited

Loading

Bazeloth commented Jun 5, 2024 •

edited

Loading

ioshm Jun 5, 2024 •

edited

Loading