[RFC] Update issues via pull requests

[krrutkow]

Various improvements on how issues are referenced, opened, or closed by pull requests. Steps to accomplish:

• factor out reference, open, and close detection code
• enable commits in a PR creation or update on the tracked branch to reference issues
• enable PR and issue title or message to reference issues
• enable PR and issue comments to reference issues
• enable a PR's commits to open/close issues when PR is merged (Auto-closing of issues via PR doesn't work #2735)
• enable a PR's title or message to open/close issues when PR is merged (Auto-closing of issues via PR doesn't work #2735)
• add testing of updates to issues via PR
• enable a manually merged PR to open/close issues
• enable code comments to reference issues
• enable review comment to reference issues
• add migration to update content of existing references and retroactively create issue references

[codecov-io]

Codecov Report

Merging #3695 into master will decrease coverage by 0.41%.
The diff coverage is 21.32%.

@@            Coverage Diff             @@
##           master    #3695      +/-   ##
==========================================
- Coverage   37.58%   37.17%   -0.42%     
==========================================
  Files         318      306      -12     
  Lines       46928    45653    -1275     
==========================================
- Hits        17638    16971     -667     
+ Misses      26778    26210     -568     
+ Partials     2512     2472      -40

Impacted Files	Coverage Δ
models/migrations/migrations.go	2.61% <ø> (ø)	⬆️
routers/api/v1/repo/pull.go	13.38% <ø> (-3.75%)	⬇️
routers/repo/pull.go	34.11% <ø> (+0.38%)	⬆️
models/issue.go	49.25% <0%> (+0.85%)	⬆️
models/migrations/v73.go	0% <0%> (ø)	⬆️
models/issue_comment.go	48.13% <32.65%> (-0.24%)	⬇️
models/pull.go	50.15% <32.65%> (-1.27%)	⬇️
models/action.go	56.81% <45.73%> (-1.95%)	⬇️
routers/repo/issue.go	32.34% <62.5%> (-4.94%)	⬇️
models/update.go	26.2% <7.84%> (-5.45%)	⬇️
... and 120 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 9681c83...ac30d94. Read the comment docs.

[yasuokav]

Maybe we should update the issues when a PR got manually merged?

[lunny]

please resolve conflicts

[krrutkow]

@yasuokav I'm not sure the PR can be automatically closed if the commits are present since the PR's tracked branch could still be committed to. Commits that are pushed directly to the repo are processed with the pre-existing function CommitRepoAction, so a manual merge commit message can mention the PR to explicitly close it.

[yasuokav]

@krrutkow Thanks for your answer, you may misunderstand what I mean.
If I manually merge the PR branch into master and push to remote, this PR will be automatically marked as merged.

gitea/models/pull.go

Lines 553 to 585 in b908ac9

	func (pr *PullRequest) manuallyMerged() bool {
	commit, err := pr.getMergeCommit()
	if err != nil {
	log.Error(4, "PullRequest[%d].getMergeCommit: %v", pr.ID, err)
	return false
	}
	if commit != nil {
	pr.MergedCommitID = commit.ID.String()
	pr.MergedUnix = util.TimeStamp(commit.Author.When.Unix())
	pr.Status = PullRequestStatusManuallyMerged
	merger, _ := GetUserByEmail(commit.Author.Email)
	// When the commit author is unknown set the BaseRepo owner as merger
	if merger == nil {
	if pr.BaseRepo.Owner == nil {
	if err = pr.BaseRepo.getOwner(x); err != nil {
	log.Error(4, "BaseRepo.getOwner[%d]: %v", pr.ID, err)
	return false
	}
	}
	merger = pr.BaseRepo.Owner
	}
	pr.Merger = merger
	pr.MergerID = merger.ID
	if err = pr.setMerged(); err != nil {
	log.Error(4, "PullRequest[%d].setMerged : %v", pr.ID, err)
	return false
	}
	log.Info("manuallyMerged[%d]: Marked as manually merged into %s/%s by commit id: %s", pr.ID, pr.BaseRepo.Name, pr.BaseBranch, commit.ID.String())
	return true
	}
	return false

But MergePullRequestAction does not trigger

[InExtremaRes]

What is the status of this? Is ready for use or merge into master?

[lunny]

@krrutkow It's ready for review? @InExtremaRes need mantianers' review.

[krrutkow]

I just rebased it to master, its ready for review.

[jonasfranz]

Nice work! Thank you for your dedication!

[jonasfranz]

Fixes #4302 and #3134

[lafriks]

@MichiiL if you want you can also rebase/fix it and open new PR if original author has no time to fix this

[stale]

This pull request has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs during the next 2 months. Thank you for your contributions.

[buhtz]

What is the current status?

[techknowlogick]

I know this file previously used hardcoded ints to do an equality check against type, but could you change it to do an equality check where 24 could easily be determined? (same applies for other checks too)

[techknowlogick]

@noMICROSOFTbuhtz the status can be read above. There is feedback that needs to be addressed and conflicts need to be resolved, however due to the change freeze for 1.9.0 being in place this has been moved to the 1.10.0 milestone.

[joelft]

I've tried to rebase this, but I don't know enough about the code base to do it successfully without breaking things. I'll continue to try, but if anyone else can jump on this it would be awesome. Seems like maybe @krrutkow doesn't have time anymore. 😢

[lunny]

conflicted

[buhtz]

Btw: How can it be that a PR is open for more then one year?

Conflictes? This is not an argument for such a hugh time span.

It is more a clue that the PR is to big and complex - 32 commits solidify this argument. Split it into easer handable commits and PRs.

I would recommand to the core developers to find a standardized way to handle, accept and reject PRs depending on defined parameter (e.g. number of lines, commit count, PR age, ...).

[lafriks]

@Codeberg-AsGithubAlternative-buhtz Someone has to take over to resolve conflicts (either by maintainers or someone else by opening new PR), we can't really review while there are conflicts that could result in broken code if not resolved correctly

[guillep2k]

Mmm... I'm working on something that should greatly simplify this code; but you were first. Want me to stop over there?
Link #8137

[6543]

@krrutkow @guillep2k whats the state of this PR ?

lot of code has changed in this area so :(

[guillep2k]

@krrutkow @guillep2k whats the state of this PR ?

lot of code has changed in this area so :(

#8137 is already merged, and I'm working on closing/reopening by keyword.

[6543]

so we can considder this as closed?

[krrutkow]

This PR is free to be closed. It applies to (and was working for) ~v1.6, but it would be nearly a complete rewrite to get working with the current codebase. The feature checklist is valuable though and should be expressed in other issues/PRs to ensure the features of this PR are not lost to the project.