Error 500 creating PR with assignees #2042

New Issue

2025-11-02T04:22:04-06:00

GiteaMirror commented

2025-11-02 04:22:04 -06:00

Originally created by @lukas010 on GitHub (Jul 11, 2018).

Gitea version (or commit ref): bbf9abde49
Git version: 2.7.4
Operating system: Ubuntu 16.04.4 LTS
Database (use [x]):
- PostgreSQL
- MySQL
- MSSQL
- SQLite
Can you reproduce the bug at https://try.gitea.io:
- Yes
- No
- Not relevant
Log gist:

Description

When I try to create pull request with one or more assignees I get 500 error. Error doesn't show up if nobody is assigned. I'm able to add assignees only for already existing pull requests.

Log error:

2018/07/10 15:41:42 [...routers/repo/pull.go:836 CompareAndPullRequestPost()] [E] NewPullRequest: newIssue: loadPullRequest: pull request does not exist [id: 0, issue_id: 24, head_repo_id: 0, base_repo_id: 0, head_branch: , base_branch: ]

Originally created by @lukas010 on GitHub (Jul 11, 2018).  - Gitea version (or commit ref): bbf9abde4937285c77154129ae9934b22b4ea7a6 - Git version: 2.7.4 - Operating system: Ubuntu 16.04.4 LTS - Database (use `[x]`): - [ ] PostgreSQL - [x] MySQL - [ ] MSSQL - [ ] SQLite - Can you reproduce the bug at https://try.gitea.io: - [x] Yes - [ ] No - [ ] Not relevant - Log gist: ## Description When I try to create pull request with one or more assignees I get 500 error. Error doesn't show up if nobody is assigned. I'm able to add assignees only for already existing pull requests. Log error: ``` 2018/07/10 15:41:42 [...routers/repo/pull.go:836 CompareAndPullRequestPost()] [E] NewPullRequest: newIssue: loadPullRequest: pull request does not exist [id: 0, issue_id: 24, head_repo_id: 0, base_repo_id: 0, head_branch: , base_branch: ] ```

GiteaMirror added the type/bug label 2025-11-02 04:22:04 -06:00

GiteaMirror closed this issue

2025-11-02 04:22:04 -06:00

GiteaMirror commented

2025-11-02 04:22:05 -06:00

@SagePtr commented on GitHub (Jul 29, 2018):

There is form of spaghetti in code:

First, newIssue is called. It creates new issue and puts it into database.
newIssue function calls ChangeAssignee for each assigner in created issue.
ChangeAssignee calls loadPullRequest to load pull request information (for supsequent webhook calls), but pull request is not ready yet.

I see two possible solutions:

NewPullRequest should create new issue without assignees, then add pull request to database, and then add every assignee with changeAssignee function.
changeAssignee should not enqueue webhooks calls for pull requests in the middle of creation (if corresponding record is not in database yet), but after successful PR record insertion NewPullRequest can loop through assignees and enqueue webhook call for each of them.

Both look like dirty workarounds for me, but if some of them is acceptible, implement it (or I could implement it, although i'm not golang coder, but syntax looks clear for me)

@SagePtr commented on GitHub (Jul 29, 2018): There is form of spaghetti in code: First, newIssue is called. It creates new issue and puts it into database. newIssue function calls ChangeAssignee for each assigner in created issue. ChangeAssignee calls loadPullRequest to load pull request information (for supsequent webhook calls), but pull request is not ready yet. I see two possible solutions: 1. NewPullRequest should create new issue without assignees, then add pull request to database, and then add every assignee with changeAssignee function. 2. changeAssignee should not enqueue webhooks calls for pull requests in the middle of creation (if corresponding record is not in database yet), but after successful PR record insertion NewPullRequest can loop through assignees and enqueue webhook call for each of them. Both look like dirty workarounds for me, but if some of them is acceptible, implement it (or I could implement it, although i'm not golang coder, but syntax looks clear for me)

GiteaMirror commented

2025-11-02 04:22:05 -06:00

@ptman commented on GitHub (Aug 17, 2018):

This seems to affect 1.5.0, why not backport?

@ptman commented on GitHub (Aug 17, 2018): This seems to affect 1.5.0, why not backport?

GiteaMirror commented

2025-11-02 04:22:05 -06:00

@ptman commented on GitHub (Aug 17, 2018):

Ah, it was backported. Maybe this could be made clearer some way?

@ptman commented on GitHub (Aug 17, 2018): Ah, it was backported. Maybe this could be made clearer some way?

GiteaMirror commented

2025-11-02 04:22:05 -06:00

@lafriks commented on GitHub (Aug 17, 2018):

Yes for backported issues milestone should be changed

@lafriks commented on GitHub (Aug 17, 2018): Yes for backported issues milestone should be changed

Sign in to join this conversation.

Branches Tags

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: github-starred/gitea#2042