forgejo/models/issues/pull_list.go
Earl Warren ceea9c4334
Revert "avoid superfluous synchronized pull_request run when opening a PR"
The fix against the race incorrectly assumes the sha of the commit being
pushed belongs to the base repository. It finds the highest possible
pull request ID from the head repository instead of looking it up in
the base repository.

Figuring out if a PR was created in the future based on the highest
index of the base repository would require collecting all of them
because there is no way to know in advance which repository may be
involved in the race.

Fixing this race can be done either by:

* Introducing a new field in the pull_request table https://codeberg.org/forgejo/forgejo/pulls/2842
  which feels more like a hack than a real solution
* Refactoring the logic
  which would be a significant undertaking

The race has been in the codebase for a very long time and manifests
itself in the CI, when events happen in quick succession. The only
concrete manifestation was however fixed by https://codeberg.org/forgejo/forgejo/issues/2009

Since this race now only exists in theory and not in practice, let's
revert this bugous commit until a proper solution is implemented.

Fixes: https://codeberg.org/forgejo/forgejo/issues/2817

This reverts commit 036f1eddc5.

Conflicts:
	services/pull/pull.go
2024-03-28 07:08:17 +01:00

220 lines
6.8 KiB
Go

// Copyright 2019 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
package issues
import (
"context"
"fmt"
"code.gitea.io/gitea/models/db"
access_model "code.gitea.io/gitea/models/perm/access"
"code.gitea.io/gitea/models/unit"
user_model "code.gitea.io/gitea/models/user"
"code.gitea.io/gitea/modules/log"
"code.gitea.io/gitea/modules/util"
"xorm.io/xorm"
)
// PullRequestsOptions holds the options for PRs
type PullRequestsOptions struct {
db.ListOptions
State string
SortType string
Labels []int64
MilestoneID int64
}
func listPullRequestStatement(ctx context.Context, baseRepoID int64, opts *PullRequestsOptions) (*xorm.Session, error) {
sess := db.GetEngine(ctx).Where("pull_request.base_repo_id=?", baseRepoID)
sess.Join("INNER", "issue", "pull_request.issue_id = issue.id")
switch opts.State {
case "closed", "open":
sess.And("issue.is_closed=?", opts.State == "closed")
}
if len(opts.Labels) > 0 {
sess.Join("INNER", "issue_label", "issue.id = issue_label.issue_id").
In("issue_label.label_id", opts.Labels)
}
if opts.MilestoneID > 0 {
sess.And("issue.milestone_id=?", opts.MilestoneID)
}
return sess, nil
}
// GetUnmergedPullRequestsByHeadInfo returns all pull requests that are open and has not been merged
func GetUnmergedPullRequestsByHeadInfo(ctx context.Context, repoID int64, branch string) ([]*PullRequest, error) {
prs := make([]*PullRequest, 0, 2)
sess := db.GetEngine(ctx).
Join("INNER", "issue", "issue.id = pull_request.issue_id").
Where("head_repo_id = ? AND head_branch = ? AND has_merged = ? AND issue.is_closed = ? AND flow = ?", repoID, branch, false, false, PullRequestFlowGithub)
return prs, sess.Find(&prs)
}
// CanMaintainerWriteToBranch check whether user is a maintainer and could write to the branch
func CanMaintainerWriteToBranch(ctx context.Context, p access_model.Permission, branch string, user *user_model.User) bool {
if p.CanWrite(unit.TypeCode) {
return true
}
if len(p.Units) < 1 {
return false
}
prs, err := GetUnmergedPullRequestsByHeadInfo(ctx, p.Units[0].RepoID, branch)
if err != nil {
return false
}
for _, pr := range prs {
if pr.AllowMaintainerEdit {
err = pr.LoadBaseRepo(ctx)
if err != nil {
continue
}
prPerm, err := access_model.GetUserRepoPermission(ctx, pr.BaseRepo, user)
if err != nil {
continue
}
if prPerm.CanWrite(unit.TypeCode) {
return true
}
}
}
return false
}
// HasUnmergedPullRequestsByHeadInfo checks if there are open and not merged pull request
// by given head information (repo and branch)
func HasUnmergedPullRequestsByHeadInfo(ctx context.Context, repoID int64, branch string) (bool, error) {
return db.GetEngine(ctx).
Where("head_repo_id = ? AND head_branch = ? AND has_merged = ? AND issue.is_closed = ? AND flow = ?",
repoID, branch, false, false, PullRequestFlowGithub).
Join("INNER", "issue", "issue.id = pull_request.issue_id").
Exist(&PullRequest{})
}
// GetUnmergedPullRequestsByBaseInfo returns all pull requests that are open and has not been merged
// by given base information (repo and branch).
func GetUnmergedPullRequestsByBaseInfo(ctx context.Context, repoID int64, branch string) ([]*PullRequest, error) {
prs := make([]*PullRequest, 0, 2)
return prs, db.GetEngine(ctx).
Where("base_repo_id=? AND base_branch=? AND has_merged=? AND issue.is_closed=?",
repoID, branch, false, false).
OrderBy("issue.updated_unix DESC").
Join("INNER", "issue", "issue.id=pull_request.issue_id").
Find(&prs)
}
// GetPullRequestIDsByCheckStatus returns all pull requests according the special checking status.
func GetPullRequestIDsByCheckStatus(ctx context.Context, status PullRequestStatus) ([]int64, error) {
prs := make([]int64, 0, 10)
return prs, db.GetEngine(ctx).Table("pull_request").
Where("status=?", status).
Cols("pull_request.id").
Find(&prs)
}
// PullRequests returns all pull requests for a base Repo by the given conditions
func PullRequests(ctx context.Context, baseRepoID int64, opts *PullRequestsOptions) ([]*PullRequest, int64, error) {
if opts.Page <= 0 {
opts.Page = 1
}
countSession, err := listPullRequestStatement(ctx, baseRepoID, opts)
if err != nil {
log.Error("listPullRequestStatement: %v", err)
return nil, 0, err
}
maxResults, err := countSession.Count(new(PullRequest))
if err != nil {
log.Error("Count PRs: %v", err)
return nil, maxResults, err
}
findSession, err := listPullRequestStatement(ctx, baseRepoID, opts)
applySorts(findSession, opts.SortType, 0)
if err != nil {
log.Error("listPullRequestStatement: %v", err)
return nil, maxResults, err
}
findSession = db.SetSessionPagination(findSession, opts)
prs := make([]*PullRequest, 0, opts.PageSize)
return prs, maxResults, findSession.Find(&prs)
}
// PullRequestList defines a list of pull requests
type PullRequestList []*PullRequest
func (prs PullRequestList) LoadAttributes(ctx context.Context) error {
if len(prs) == 0 {
return nil
}
// Load issues.
issueIDs := prs.GetIssueIDs()
issues := make([]*Issue, 0, len(issueIDs))
if err := db.GetEngine(ctx).
Where("id > 0").
In("id", issueIDs).
Find(&issues); err != nil {
return fmt.Errorf("find issues: %w", err)
}
set := make(map[int64]*Issue)
for i := range issues {
set[issues[i].ID] = issues[i]
}
for _, pr := range prs {
pr.Issue = set[pr.IssueID]
/*
Old code:
pr.Issue.PullRequest = pr // panic here means issueIDs and prs are not in sync
It's worth panic because it's almost impossible to happen under normal use.
But in integration testing, an asynchronous task could read a database that has been reset.
So returning an error would make more sense, let the caller has a choice to ignore it.
*/
if pr.Issue == nil {
return fmt.Errorf("issues and prs may be not in sync: cannot find issue %v for pr %v: %w", pr.IssueID, pr.ID, util.ErrNotExist)
}
pr.Issue.PullRequest = pr
}
return nil
}
// GetIssueIDs returns all issue ids
func (prs PullRequestList) GetIssueIDs() []int64 {
issueIDs := make([]int64, 0, len(prs))
for i := range prs {
issueIDs = append(issueIDs, prs[i].IssueID)
}
return issueIDs
}
// HasMergedPullRequestInRepo returns whether the user(poster) has merged pull-request in the repo
func HasMergedPullRequestInRepo(ctx context.Context, repoID, posterID int64) (bool, error) {
return db.GetEngine(ctx).
Join("INNER", "pull_request", "pull_request.issue_id = issue.id").
Where("repo_id=?", repoID).
And("poster_id=?", posterID).
And("is_pull=?", true).
And("pull_request.has_merged=?", true).
Select("issue.id").
Limit(1).
Get(new(Issue))
}
// GetPullRequestByIssueIDs returns all pull requests by issue ids
func GetPullRequestByIssueIDs(ctx context.Context, issueIDs []int64) (PullRequestList, error) {
prs := make([]*PullRequest, 0, len(issueIDs))
return prs, db.GetEngine(ctx).
Where("issue_id > 0").
In("issue_id", issueIDs).
Find(&prs)
}