Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Improve migrations to support migrating milestones/labels/issues/comments/pullrequests #6290

Merged
merged 33 commits into from May 7, 2019

Conversation

lunny
Copy link
Member

@lunny lunny commented Mar 9, 2019

This PR will improve the migrations process. It will define an interface to migrate data from external git service and at first implement from github to gitea. Below is the TODO list:

  • Define the migration interface
  • Migrate labels from github
  • Migrate milestones from github
  • Migrate issues and comments from github
  • Migrate pull requests and comments from github: remote deleted branch will made commits lost.
  • Migrate releases from github
  • Tests for download from github
  • Tests for upload to gitea
  • UI options to check wiki, milestones, labels, issues(with comments), prs(with comments), releases

Below will be ignored

  • Migrate issues or PRs' user from github: It requires we creates all the users automatically but cannot determine how to match the users between two sites since cannot retrieve users' email.
  • Migrate issues or PRs' reactions from github: reactions on gitea requires all the users IDs which cannot retrieves.

The new migration UI looks like:

image

Next steps after this merged

  • Implement GiteaDownloadV1 so that repositories could be transfer cross gitea instances.
  • Implement GitlabDownload
  • Implement BitBucketDownload
  • Implement GogsDownloadV1

@lunny lunny added the type/feature Completely new functionality. Can only be merged if feature freeze is not active. label Mar 9, 2019
@lunny lunny changed the title WIP: add migrations WIP: improve migrations to support migrating milestones/labels/issues/reactions/comments/pullrequests Mar 9, 2019
@lunny lunny mentioned this pull request Mar 16, 2019
20 tasks
@codecov-io
Copy link

codecov-io commented Mar 16, 2019

Codecov Report

❗ No coverage uploaded for pull request base (master@1c7c739). Click here to learn what that means.
The diff coverage is 39.23%.

Impacted file tree graph

@@            Coverage Diff            @@
##             master    #6290   +/-   ##
=========================================
  Coverage          ?   41.13%           
=========================================
  Files             ?      432           
  Lines             ?    59504           
  Branches          ?        0           
=========================================
  Hits              ?    24475           
  Misses            ?    31817           
  Partials          ?     3212
Impacted Files Coverage Δ
modules/auth/repo_form.go 41.73% <ø> (ø)
models/migrate.go 0% <0%> (ø)
modules/migrations/base/pullrequest.go 0% <0%> (ø)
modules/migrations/error.go 0% <0%> (ø)
modules/migrations/migrate.go 23.88% <23.88%> (ø)
routers/repo/repo.go 22.96% <38.09%> (ø)
routers/api/v1/repo/repo.go 55.59% <44.23%> (ø)
modules/migrations/git.go 52% <52%> (ø)
models/repo.go 47.82% <64.28%> (ø)
modules/migrations/gitea.go 8.3% <8.3%> (ø)
... and 1 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 1c7c739...5436028. Read the comment docs.

@GiteaBot GiteaBot added the lgtm/need 2 This PR needs two approvals by maintainers to be considered for merging. label Mar 16, 2019
@lunny lunny added this to the 1.9.0 milestone Mar 16, 2019
@lunny lunny force-pushed the lunny/migrations branch 4 times, most recently from cfa1233 to a7c2fe0 Compare March 24, 2019 08:21
@lunny lunny force-pushed the lunny/migrations branch 2 times, most recently from c5bc0ed to 4daa5c7 Compare March 30, 2019 05:05
@lunny lunny force-pushed the lunny/migrations branch 2 times, most recently from 7ceb4de to c0878f8 Compare April 12, 2019 08:35
@lunny lunny changed the title WIP: improve migrations to support migrating milestones/labels/issues/reactions/comments/pullrequests Improve migrations to support migrating milestones/labels/issues/reactions/comments/pullrequests Apr 12, 2019
@lunny
Copy link
Member Author

lunny commented Apr 12, 2019

I think this is ready for review.

Copy link
Member

@kolaente kolaente left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Also, what happens if the user creates the repo to be a mirror? Will issues/prs/wiki etc still be synced or only on the initial creation of the repo?

import "github.com/go-xorm/xorm"

// InsertIssue insert one issue to database
func InsertIssue(issue *Issue, labelIDs []int64) error {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

How is this different to createIssue?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This will not trigger notifications.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

But then I'd prefer to have createIssue call this function and trigger notifications to avoid having duplicated code.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I want to do that, but it's so different between theses two functions I would like let it duplicated currently. Maybe send a refactor PR after this PR.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fair enough, sounds good.

models/repo.go Outdated
@@ -877,6 +877,7 @@ type MigrateRepoOptions struct {
IsPrivate bool
IsMirror bool
RemoteAddr string
NoWiki bool
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please call this UsesWiki or just Wiki or something like that, then check if opts.Wiki instead of if !opts.NoWiki to avoid double negation.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done.

if err := os.RemoveAll(wikiPath); err != nil {
return repo, fmt.Errorf("Failed to remove %s: %v", wikiPath, err)
}

if err = git.Clone(wikiRemotePath, wikiPath, git.CloneRepoOptions{
Mirror: true,
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why is this always a mirror?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think the Mirror means a full clone but not a mirror repository on Gitea.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

These codes just moved from above because I added opts.Wiki.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think this is probably ok and we certainly want at least part of what it's doing regarding copying the remote tracking branches. We should check what happens when we remove the upstream at the end of the migration and check that the refspec configuration is cleaned up though.

)

// GiteaLocalUploader implements an Uploader to gitea sites
type GiteaLocalUploader struct {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

IMHO since this is already on the local Gitea instance, the word Local is double here.

The name should also make it more clear this is a migrator from github (is it?) and not just something "to move local git repos around".

Copy link
Member Author

@lunny lunny Apr 13, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@kolaente there are uploaders and downloaders here. GiteaLocalUploader will save data via gitea's local operations. Currently we have GithubV3Downloader, in future, we could implement GitLab/Gitea/Gogs/... downloaders. Furthermore we could have GithubV3Uploader so that we can migrate gitea's data to github.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So the flow is GithubV3Uploader -> GiteaLocalUploader?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Current flow is GithubV3Downloader - > GiteaLocalUploader. Always from a downloader interface implementation to an uploader interface implementation.

}

// CreateRepo creates a repository
func (g *GiteaLocalUploader) CreateRepo(repo *base.Repository, includeWiki bool) error {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

How is this (and a lot of other functions in this file) different from the implementation in models/migrate.go?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This file will invoke models/migrate.go when migrate git/wiki data.

labels = append(labels, convertGithubLabel(l))
}
// FIXME: This API missing reactions, we may need another extra request to get reactions
/*var reactions *Reactions
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please remove the commented code.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done.


if opts.PullRequests {
log.Trace("migrating pull requests and comments")
prs, err := downloader.GetPullRequests(0, 1000000)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What does this magic number do? I'd guess it's a limit of some kind, but what if there are > 1000000? Will it fail? Will it just not downlod anything > 1000000?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

resolved.

@@ -979,6 +979,23 @@ function initRepository() {
}
}

var toggleMigrations = function(){
console.log("toggleMigrations")
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please remove this console.log

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done.

@@ -979,6 +979,23 @@ function initRepository() {
}
}

var toggleMigrations = function(){
console.log("toggleMigrations")
if (!$('#mirror').is(":checked") && $('#auth_username').val().length > 0 && ($('#clone_addr').val().startsWith("https://github.com") || $('#clone_addr').val().startsWith("http://github.com"))) {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Won't this only be triggered when a user loads the page and the clone url is already in the clone field? IMHO this should listen on input.

@@ -263,11 +290,12 @@ func MigratePost(ctx *context.Context, form auth.MigrateRepoForm) {
// remoteAddr may contain credentials, so we sanitize it
err = util.URLSanitizedError(err, remoteAddr)

if repo != nil {
// TODO: remove
/*if repo != nil {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please remove this commented code.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done.

@ashimokawa
Copy link
Contributor

@lunny
It seems to work, nice work. 👍

One minor issue I noticed is that wrong github passwords gives just a big error 500 screen, that should not happen.

@lunny
Copy link
Member Author

lunny commented Apr 13, 2019

@kolaente When you check mirror on the UI, all the migration items will be hidden. That means issues/milestones/... will not be mirrored.

@lunny
Copy link
Member Author

lunny commented Apr 13, 2019

@ashimokawa yes, I will fix that.

@lunny
Copy link
Member Author

lunny commented Apr 13, 2019

@ashimokawa auth error handled.

@ashimokawa
Copy link
Contributor

@lunny
Great progress!

What about introducing a new database column for comments "original_author" or something where "github/ashimokawa" is stored for example?

Then if it is confirmed later that ashimokawa signed up on the gitea instance where the repo was moved to, the admin could reassign the comments to the real local user ashimokawa.
This would also eliminate the need to store the original author as text inside the issue comment text itself, and it would be possible to put some nice "imported from github" widget above or below the comment instead, which is clickable and leads the the github user page as long as it has not been reassigned by the local admin to the local user.

@lunny
Copy link
Member Author

lunny commented Apr 14, 2019

@ashimokawa I will not do that change since it's not possible to let all repository related users register in this Gitea instance and you cannot mapping them in fact. It will not result in difficult on maintain all the issues and pull requests even if the posters are not the original authors.

@ashimokawa
Copy link
Contributor

@lunny
I get your point, still I think my proposal makes sense - not to propery assign all of the comments, but it would be great for a core team of a project, which moved to its own gitea instance where the admin knows who is who to be able to reassign authors properly. That is what my proposal was for.

@lunny
Copy link
Member Author

lunny commented Apr 14, 2019

@ashimokawa that could be next step for a private repository because I think there are many work to do. Currently I'll focus on public repositories.

@lunny lunny added the type/changelog Adds the changelog for a new Gitea version label Apr 14, 2019
@lunny
Copy link
Member Author

lunny commented May 5, 2019

@jolheiser done.

Copy link
Member

@techknowlogick techknowlogick left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good, my one concern is that with a large project migration API calls might hit rate limit quota, but I think adding in a GitHub token for auth to increase API rate limit can be done in another PR.

@GiteaBot GiteaBot added lgtm/done This PR has enough approvals to get merged. There are no important open reservations anymore. and removed lgtm/need 1 This PR needs approval from one additional maintainer to be merged. labels May 6, 2019
@lunny
Copy link
Member Author

lunny commented May 7, 2019

@techknowlogick that's already supported on this PR. you could put your token on username inputbox and then that's it.

@lunny lunny merged commit 08069dc into go-gitea:master May 7, 2019
@techknowlogick
Copy link
Member

@lunny oh that's perfect! I missed that. Thanks for telling me :)

@strk
Copy link
Member

strk commented Aug 7, 2019

The token-on-username thing is not documented anywhere, or did I miss it ?

@lunny
Copy link
Member Author

lunny commented Aug 7, 2019

It is mentioned on the tip above the input boxes.

@strk
Copy link
Member

strk commented Aug 7, 2019

Also, from gitea.com when selecting "This repo will be a mirror" the options to also sync issues/labels/releases etc. disappears. Is that not supported ?

@strk
Copy link
Member

strk commented Aug 7, 2019

The migration target page now became a 500 error on https://gitea.com/QGIS/QGIS ...

@lunny
Copy link
Member Author

lunny commented Aug 7, 2019

mirroring is not supported now. When it's migrating, it maybe failed on UI but it may still running on backend. That's what my new PR wants to resolve. See #6200

@6543 6543 mentioned this pull request Oct 25, 2019
24 tasks
@go-gitea go-gitea locked and limited conversation to collaborators Nov 24, 2020
@delvh delvh removed the type/changelog Adds the changelog for a new Gitea version label Oct 7, 2023
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
lgtm/done This PR has enough approvals to get merged. There are no important open reservations anymore. type/feature Completely new functionality. Can only be merged if feature freeze is not active.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

10 participants