The MySQL indexes are not being renamed at the same time as RENAME table despite the
CASCADE. Therefore it is probably better to just recreate the indexes instead.
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: 6543 <6543@obermui.de>
* Restore compatibility with SQLServer 2008 R2 in migrations
`ALTER TABLE DROP ... IF EXISTS ...` is only supported in SQL Server >16.
The `IF EXISTS` here is a belt-and-braces and does not need to be present. Therefore
can be dropped.
We need to figure out some way of restricting our SQL syntax against the minimum
version of SQL Server we will support.
My suspicion is that `ALTER DATABASE database_name SET COMPATIBILITY_LEVEL = 100` may
do that but there may be other side-effects so I am not whether to do that.
Signed-off-by: Andrew Thornton <art27@cantab.net>
* try just dropping the index only
Signed-off-by: Andrew Thornton <art27@cantab.net>
* use lowercase for system tables
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: Lauris BH <lauris@nix.lv>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
`models` does far too much. In particular it handles all `UserSignin`.
It shouldn't be responsible for calling LDAP, SMTP or PAM for signing in.
Therefore we should move this code out of `models`.
This code has to depend on `models` - therefore it belongs in `services`.
There is a package in `services` called `auth` and clearly this functionality belongs in there.
Plan:
- [x] Change `auth.Auth` to `auth.Method` - as they represent methods of authentication.
- [x] Move `models.UserSignIn` into `auth`
- [x] Move `models.ExternalUserLogin`
- [x] Move most of the `LoginVia*` methods to `auth` or subpackages
- [x] Move Resynchronize functionality to `auth`
- Involved some restructuring of `models/ssh_key.go` to reduce the size of this massive file and simplify its files.
- [x] Move the rest of the LDAP functionality in to the ldap subpackage
- [x] Re-factor the login sources to express an interfaces `auth.Source`?
- I've done this through some smaller interfaces Authenticator and Synchronizable - which would allow us to extend things in future
- [x] Now LDAP is out of models - need to think about modules/auth/ldap and I think all of that functionality might just be moveable
- [x] Similarly a lot Oauth2 functionality need not be in models too and should be moved to services/auth/source/oauth2
- [x] modules/auth/oauth2/oauth2.go uses xorm... This is naughty - probably need to move this into models.
- [x] models/oauth2.go - mostly should be in modules/auth/oauth2 or services/auth/source/oauth2
- [x] More simplifications of login_source.go may need to be done
- Allow wiring in of notify registration - *this can now easily be done - but I think we should do it in another PR* - see #16178
- More refactors...?
- OpenID should probably become an auth Method but I think that can be left for another PR
- Methods should also probably be cleaned up - again another PR I think.
- SSPI still needs more refactors.* Rename auth.Auth auth.Method
* Restructure ssh_key.go
- move functions from models/user.go that relate to ssh_key to ssh_key
- split ssh_key.go to try create clearer function domains for allow for
future refactors here.
Signed-off-by: Andrew Thornton <art27@cantab.net>
* 企业微信webhook
* 企业微信webhook
* 企业微信webhook
* Update templates/admin/hook_new.tmpl
Co-authored-by: a1012112796 <1012112796@qq.com>
* Update services/webhook/wechatwork.go
Co-authored-by: a1012112796 <1012112796@qq.com>
* 修善wechatwork
* 修善wechatwork
* fix
* Update locale_cs-CZ.ini
fix
* fix build
* fix
* fix build
* make webhooks.zh-cn.md
* delet unnecessary blank line
* delet unnecessary blank line
* 企业微信webhook
* 企业微信webhook
* 企业微信webhook
* Update templates/admin/hook_new.tmpl
Co-authored-by: a1012112796 <1012112796@qq.com>
* Update services/webhook/wechatwork.go
Co-authored-by: a1012112796 <1012112796@qq.com>
* 修善wechatwork
* 修善wechatwork
* fix
* fix build
* fix
* fix build
* make webhooks.zh-cn.md
* delet unnecessary blank line
* delet unnecessary blank line
* 企业微信webhook
* 企业微信webhook
* 企业微信webhook
* 企业微信webhook
* 企业微信webhook
* fix
* fix
* 企业微信webhook
* 企业微信webhook
* 企业微信webhook
* fix wechat
* fix wechat
* fix wechat
* fix wechat
* Fix invalid params and typo of email templates (#16394)
Signed-off-by: Meano <meanocat@gmail.com>
* Add LRU mem cache implementation (#16226)
The current default memory cache implementation is unbounded in size and number of
objects cached. This is hardly ideal.
This PR proposes creating a TwoQueue LRU cache as the underlying cache for Gitea.
The cache is limited by the number of objects stored in the cache (rather than size)
for simplicity. The default number of objects is 50000 - which is perhaps too small
as most of our objects cached are going to be much less than 1kB.
It may be worth considering using a different LRU implementation that actively limits
sizes or avoids GC - however, this is just a beginning implementation.
Signed-off-by: Andrew Thornton <art27@cantab.net>
* [skip ci] Updated translations via Crowdin
* Replace `plugins/docker` with `techknowlogick/drone-docker`in ci (#16407)
* plugins/docker -> techknowlogick/drone-docker
* It is multi-arch
* docs: rewrite email setup (#16404)
* Add intro for both the docs page and mailer methods
* Fix numbering level in SMTP section
* Recommends implicit TLS
Signed-off-by: Bagas Sanjaya <bagasdotme@gmail.com>
* Validate Issue Index before querying DB (#16406)
* Fix external renderer (#16401)
* fix external renderer
* use GBackground context as fallback
* no fallback, return error
Co-authored-by: Lauris BH <lauris@nix.lv>
* Add checkbox to delete pull branch after successful merge (#16049)
* Add checkbox to delete pull branch after successful merge
* Omit DeleteBranchAfterMerge field in json
* Log a warning instead of error when PR head branch deleted
* Add DefaultDeleteBranchAfterMerge to PullRequestConfig
* Add support for delete_branch_after_merge via API
* Fix for API: the branch should be deleted from the HEAD repo
If head and base repo are the same, reuse the already opened ctx.Repo.GitRepo
* Don't delegate to CleanupBranch, only reuse branch deletion code
CleanupBranch contains too much logic that has already been performed by the Merge
* Reuse gitrepo in MergePullRequest
Co-authored-by: Andrew Thornton <art27@cantab.net>
* [skip ci] Updated translations via Crowdin
* Detect encoding changes while parsing diff (#16330)
* Detect encoding changes while parsing diff
* Let branch/tag name be a valid ref to get CI status (#16400)
* fix #16384#
* refactor: move shared helper func to utils package
* extend Tests
* use ctx.Repo.GitRepo if not nil
* fix
* fix
* 企业微信webhook
* 企业微信webhook
* 企业微信webhook
* fix build
* fix build
* Apply suggestions from code review
Co-authored-by: a1012112796 <1012112796@qq.com>
Co-authored-by: myheavily <myheavily>
Co-authored-by: zhaoxin <gitea@fake.local>
Co-authored-by: Meano <Meano@foxmail.com>
Co-authored-by: zeripath <art27@cantab.net>
Co-authored-by: GiteaBot <teabot@gitea.io>
Co-authored-by: 6543 <6543@obermui.de>
Co-authored-by: Bagas Sanjaya <bagasdotme@gmail.com>
Co-authored-by: Norwin <noerw@users.noreply.github.com>
Co-authored-by: Lauris BH <lauris@nix.lv>
Co-authored-by: Jimmy Praet <jimmy.praet@telenet.be>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
* Add option to provide signed token to verify key ownership
Currently we will only allow a key to be matched to a user if it matches
an activated email address. This PR provides a different mechanism - if
the user provides a signature for automatically generated token (based
on the timestamp, user creation time, user ID, username and primary
email.
* Ensure verified keys can act for all active emails for the user
* Add code to mark keys as verified
* Slight UI adjustments
* Slight UI adjustments 2
* Simplify signature verification slightly
* fix postgres test
* add api routes
* handle swapped primary-keys
* Verify the no-reply address for verified keys
* Only add email addresses that are activated to keys
* Fix committer shortcut properly
* Restructure gpg_keys.go
* Use common Verification Token code
Signed-off-by: Andrew Thornton <art27@cantab.net>
This PR removes multiple unneeded fields from the `HookTask` struct and adds the two headers `X-Hub-Signature` and `X-Hub-Signature-256`.
## ⚠️ BREAKING ⚠️
* The `Secret` field is no longer passed as part of the payload.
* "Breaking" change (or fix?): The webhook history shows the real called url and not the url registered in the webhook (`deliver.go`@129).
Close#16115Fixes#7788Fixes#11755
Co-authored-by: zeripath <art27@cantab.net>
* Add migrating message
Signed-off-by: Andrew Thornton <art27@cantab.net>
* simplify messenger
Signed-off-by: Andrew Thornton <art27@cantab.net>
* make messenger an interface
Signed-off-by: Andrew Thornton <art27@cantab.net>
* rename
Signed-off-by: Andrew Thornton <art27@cantab.net>
* prepare for merge
Signed-off-by: Andrew Thornton <art27@cantab.net>
* as per tech
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: 6543 <6543@obermui.de>
Unfortunately the v180 migration picked up a few non-standalone dependencies. This PR
forcibly copies the important parts back into the migration.
Fix#16150
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
* Add a new table issue_index to store the max issue index so that issue could be deleted with no duplicated index
* Fix pull index
* Add tests for concurrent creating issues
* Fix lint
* Fix tests
* Fix postgres test
* Add test for migration v180
* Rename wrong test file name
Co-authored-by: 6543 <6543@obermui.de>
Co-authored-by: Lauris BH <lauris@nix.lv>
* Always store primary email address into email_address table and also the state
* Add lower_email to not convert email to lower as what's added
* Fix fixture
* Fix tests
* Use BeforeInsert to save lower email
* Fix v180 migration
* fix tests
* Fix test
* Remove wrong submited codes
* Fix test
* Fix test
* Fix test
* Add test for v181 migration
* remove change user's email to lower
* Revert change on user's email column
* Fix lower email
* Fix test
* Fix test
* encrypt migration credentials in task persistence
Not sure this is the best approach, we could encrypt the entire
`PayloadContent` instead. Also instead of clearing individual fields in
payload content, we could just delete the task once it has
(successfully) finished..?
* remove credentials of past migrations
* only run DB migration for completed tasks
* fix binding
* add omitempty
* never serialize unencrypted credentials
* fix import order
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
Co-authored-by: zeripath <art27@cantab.net>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Unfortunately some old repositories can have tags with empty Tagger, Commit
or Author. Go-Git variants will always have empty values for these whereas
the native git variant leaves them at nil. The simplest solution is just to
always have these set to empty Signatures.
v156 migration also makes the incorrect assumption that these cannot be empty.
Therefore add some handling to this and add logging and adjust broken
logging elsewhere in this migration.
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
* Use single shared random string generation function
- Replace 3 functions that do the same with 1 shared one
- Use crypto/rand over math/rand for a stronger RNG
- Output only alphanumerical for URL compatibilty
Fixes: #15536
* use const string method
* Update modules/avatar/avatar.go
Co-authored-by: a1012112796 <1012112796@qq.com>
Co-authored-by: a1012112796 <1012112796@qq.com>
go panics otherwise with `panic: interface conversion: error is git.ErrNotExist, not *git.ErrNotExist`, thanks to Codeberg/Andi for reporting this.
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
It is possible that tag commits could be deleted or missing from repos. This causes
migration 156 to fail and breaks upgrade.
This PR simply logs the failure.
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: 6543 <6543@obermui.de>
* Refactored handleOAuth2SignIn in routers/user/auth.go
The function handleOAuth2SignIn was called twice but some code path could only
be reached by one of the invocations. Moved the unnecessary code path out of
handleOAuth2SignIn.
* Refactored user creation
There was common code to create a user and display the correct error message.
And after the creation the only user should be an admin and if enabled a
confirmation email should be sent. This common code is now abstracted into
two functions and a helper function to call both.
* Added auto-register for OAuth2 users
If enabled new OAuth2 users will be registered with their OAuth2 details.
The UserID, Name and Email fields from the gothUser are used.
Therefore the OpenID Connect provider needs additional scopes to return
the coresponding claims.
* Added error for missing fields in OAuth2 response
* Linking and auto linking on oauth2 registration
* Set default username source to nickname
* Add automatic oauth2 scopes for github and google
* Add hint to change the openid connect scopes if fields are missing
* Extend info about auto linking security risk
Co-authored-by: Viktor Kuzmin <kvaster@gmail.com>
Signed-off-by: Martin Michaelis <code@mgjm.de>
The Session table must have an Expiry field not a created_unix field - somehow
this migration adds the incorrect named field leading to #15445 reports.
Fix#15445
Signed-off-by: Andrew Thornton <art27@cantab.net>
* Implemented LFS client.
* Implemented scanning for pointer files.
* Implemented downloading of lfs files.
* Moved model-dependent code into services.
* Removed models dependency. Added TryReadPointerFromBuffer.
* Migrated code from service to module.
* Centralised storage creation.
* Removed dependency from models.
* Moved ContentStore into modules.
* Share structs between server and client.
* Moved method to services.
* Implemented lfs download on clone.
* Implemented LFS sync on clone and mirror update.
* Added form fields.
* Updated templates.
* Fixed condition.
* Use alternate endpoint.
* Added missing methods.
* Fixed typo and make linter happy.
* Detached pointer parser from gogit dependency.
* Fixed TestGetLFSRange test.
* Added context to support cancellation.
* Use ReadFull to probably read more data.
* Removed duplicated code from models.
* Moved scan implementation into pointer_scanner_nogogit.
* Changed method name.
* Added comments.
* Added more/specific log/error messages.
* Embedded lfs.Pointer into models.LFSMetaObject.
* Moved code from models to module.
* Moved code from models to module.
* Moved code from models to module.
* Reduced pointer usage.
* Embedded type.
* Use promoted fields.
* Fixed unexpected eof.
* Added unit tests.
* Implemented migration of local file paths.
* Show an error on invalid LFS endpoints.
* Hide settings if not used.
* Added LFS info to mirror struct.
* Fixed comment.
* Check LFS endpoint.
* Manage LFS settings from mirror page.
* Fixed selector.
* Adjusted selector.
* Added more tests.
* Added local filesystem migration test.
* Fixed typo.
* Reset settings.
* Added special windows path handling.
* Added unit test for HTTPClient.
* Added unit test for BasicTransferAdapter.
* Moved into util package.
* Test if LFS endpoint is allowed.
* Added support for git://
* Just use a static placeholder as the displayed url may be invalid.
* Reverted to original code.
* Added "Advanced Settings".
* Updated wording.
* Added discovery info link.
* Implemented suggestion.
* Fixed missing format parameter.
* Added Pointer.IsValid().
* Always remove model on error.
* Added suggestions.
* Use channel instead of array.
* Update routers/repo/migrate.go
* fmt
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: zeripath <art27@cantab.net>
* Create Proper Migration tests
Unfortunately our testing regime has so far meant that migrations do not
get proper testing.
This PR begins the process of creating migration tests for this.
* Add test for v176
* fix mssql drop db
Signed-off-by: Andrew Thornton <art27@cantab.net>
* Fix Migration 176 yet again
Whilst creating a test for v176 in the migrations_test PR
it has become clear that this was still wrong.
This is now fixed. Genuinely.
Signed-off-by: Andrew Thornton <art27@cantab.net>
* and fix repo transfer
Signed-off-by: Andrew Thornton <art27@cantab.net>
There is a serious issue with the v176 migration where there is a mistaken missing
label_id selection.
*introduced by #14912*
Signed-off-by: Andrew Thornton <art27@cantab.net>
* Never add labels not from this repository or organisation and remove org labels on transfer
Prevent the addition of labels from outside of the repository or
organisation and remove organisation labels on transfer.
Related #14908
Signed-off-by: Andrew Thornton <art27@cantab.net>
* switch to use sql
Signed-off-by: Andrew Thornton <art27@cantab.net>
* remove AS
Signed-off-by: Andrew Thornton <art27@cantab.net>
* subquery alias
Signed-off-by: Andrew Thornton <art27@cantab.net>
* Give me some AS?
Signed-off-by: Andrew Thornton <art27@cantab.net>
* double AS
Signed-off-by: Andrew Thornton <art27@cantab.net>
* try try again
Signed-off-by: Andrew Thornton <art27@cantab.net>
* once more around the merry go round
Signed-off-by: Andrew Thornton <art27@cantab.net>
* fix api problem
Signed-off-by: Andrew Thornton <art27@cantab.net>
* Add outside label consistency check into doctor
This PR adds another consistency check into doctor in order to detect
labels that have been added from outside of repositories and organisations
Fix#14908
Signed-off-by: Andrew Thornton <art27@cantab.net>
* fix migration
Signed-off-by: Andrew Thornton <art27@cantab.net>
* prep for merge
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: 6543 <6543@obermui.de>
Co-authored-by: Lauris BH <lauris@nix.lv>
* Fix postgres ID sequences broken by recreate-table
Unfortunately there is a subtle problem with recreatetable on postgres which
leads to the sequences not being renamed and not being left at 0.
Fix#14725
Signed-off-by: Andrew Thornton <art27@cantab.net>
* let us try information_schema instead
Signed-off-by: Andrew Thornton <art27@cantab.net>
* try again
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
Co-authored-by: 6543 <6543@obermui.de>
* chore: rewrite format.
* chore: update format
Signed-off-by: Bo-Yi Wu <appleboy.tw@gmail.com>
* chore: update format
Signed-off-by: Bo-Yi Wu <appleboy.tw@gmail.com>
* chore: Adjacent parameters with the same type should be grouped together
* chore: update format.
* make repo as "pending transfer" if on transfer start doer has no right to create repo in new destination
* if new pending transfer ocured, create UI & Mail notifications
* Create Xorm session provider
This PR creates a Xorm session provider which creates
the appropriate Session table for macaron/session.
Fix#7137
Signed-off-by: Andrew Thornton <art27@cantab.net>
* extraneous l
Signed-off-by: Andrew Thornton <art27@cantab.net>
* fix lint
Signed-off-by: Andrew Thornton <art27@cantab.net>
* use key instead of ID to be compatible with go-macaron/session
Signed-off-by: Andrew Thornton <art27@cantab.net>
* And change the migration too.
Signed-off-by: Andrew Thornton <art27@cantab.net>
* Update spacing of imports
Co-authored-by: 6543 <6543@obermui.de>
* Update modules/session/xorm.go
Co-authored-by: techknowlogick <matti@mdranta.net>
* add xorm provider to the virtual provider
Signed-off-by: Andrew Thornton <art27@cantab.net>
* prep for master merge
* prep for merge master
* As per @lunny
* move migration out of the way
* Move to call this db session as per @lunny
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: 6543 <6543@obermui.de>
Co-authored-by: techknowlogick <matti@mdranta.net>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Migrations currently uses the default Xorm mapper which is
not the same as the mapper Gitea actually uses.
This means that there is a difference between the struct
parsing and mapping to database tables in migrations as
compared to normal Sync2.
This was the cause for the catastrophic problem in v168 -
untagged fields are not mapped in the same way in migrations
as compared to outside of migrations.
This is also likely the cause of some weird subtle failures
in other migrations as any untagged field may not be being
mapped exactly the same way.
This PR suggests that we ensure that the mapper is set at
the start of the migrations code - but also enforces a strict
clean mapper between each migration.
Signed-off-by: Andrew Thornton <art27@cantab.net>
* Fix mig 141
* Add Migration to fix it
* update null values to false first
* Alter Table if posible
* use dropTableColumns instead of recreateTable
* MySQL use Alter
* Postgres use Alter
* Update models/migrations/v167.go
* Apply suggestions from code review
* use 2x add col & 2x update & 2x drop col
* let sqlite be the only issue
* use recreate since it just WORKS
* Add redirect for user
* Add redirect for orgs
* Add user redirect test
* Appease linter
* Add comment to DeleteUserRedirect function
* Fix locale changes
* Fix GetUserByParams
* Fix orgAssignment
* Remove debug logging
* Add redirect prompt
* Dont Export DeleteUserRedirect & only use it within a session
* Unexport newUserRedirect
* cleanup
* Fix & Dedub API code
* Format Template
* Add Migration & rm dublicat
* Refactor: unexport newRepoRedirect() & rm dedub del exec
* if this fails we'll need to re-rename the user directory
Co-authored-by: 6543 <6543@obermui.de>
Co-authored-by: zeripath <art27@cantab.net>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
* move SaltGeneration into HashPasswort and rename it to what it does
* Migration: Where Password is Valid with Empty String delete it
* prohibit empty password hash
* let SetPassword("") unset pwd stuff
* Fix wrong type on hooktask to convert typ from char(16) to varchar(16)
* Fix bugs
* Improve code
* Use different trim function for MSSQL
* Fix bug
* Removed wrong changed line
* Removed wrong changed line
* Fix nullable
* Fix lint
* Ignore sqlite on migration
* Fix mssql modify column failure
* Move modifyColumn to migrations.go so that other migrate function could use it
This is "minimal" in the sense that only the Authorization Code Flow
from OpenID Connect Core is implemented. No discovery, no configuration
endpoint, and no user scope management.
OpenID Connect is an extension to the (already implemented) OAuth 2.0
protocol, and essentially an `id_token` JWT is added to the access token
endpoint response when using the Authorization Code Flow. I also added
support for the "nonce" field since it is required to be used in the
id_token if the client decides to include it in its initial request.
In order to enable this extension an OAuth 2.0 scope containing
"openid" is needed. Other OAuth 2.0 requests should not be impacted by
this change.
This minimal implementation is enough to enable single sign-on (SSO)
for other sites, e.g. by using something like `mod_auth_openidc` to
only allow access to a CI server if a user has logged into Gitea.
Fixes: #1310
Co-authored-by: 6543 <6543@obermui.de>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Co-authored-by: zeripath <art27@cantab.net>
