Followup of #23876 according to my unreleased review demanding tooltips.
Additionally
- add a `muted` equivalent for buttons
- convert `switch to legacy` to an actual button
- enroll `switch to legacy` in the builtin pseudo focus cycle
- remove spaces between the buttons
The effect of the `muted` class is what you would expect: The button
loses all of its normal styling, and is defined only by its content instead.
This will help reduce a11y infractions in the future, as that was one of
the major points why people didn't use `<button>` tags and decided on a
bad fix (i.e. through `<div>`s) instead.
## Appearance
---------
Co-authored-by: silverwind <me@silverwind.io>
- Add placeholders and aria-label all input fields on these two pages
- Add margin before wiki change message
- Remove labels from release page, replacing them with aria-label
I've heard many reports of users getting scared when they see their own
email address for their own profile, as they believe that the email
field is also visible to other users. Currently, using Incognito mode
or going over the Settings is the only "reasonable" way to verify this
from the perspective of the user.
A locked padlock should be enough to indicate that the email is not
visible to anyone apart from the user and the admins. An unlocked
padlock is used if the email address is only shown to authenticated
users.
Some additional string-related changes in the Settings were introduced
as well to ensure consistency, and the comments in the relevant tests
were improved so as to allow for easier modifications in the future.
---
#### Screenshot (EDIT: Scroll down for more up-to-date screenshots)
***Please remove this section before merging.***
This lock should only appear if the email address is explicitly hidden
using the `Hide Email Address` setting. The change was originally tested
on top of and designed for the Forgejo fork, but I don't expect any
problems to arise from this and I don't think that a
documentation-related change is strictly necessary.
---------
Co-authored-by: silverwind <me@silverwind.io>
Right now the authors search dropdown might take a long time to load if
amount of authors is huge.
Example: (In the video below, there are about 10000 authors, and it
takes about 10 seconds to open the author dropdown)
https://user-images.githubusercontent.com/17645053/229422229-98aa9656-3439-4f8c-9f4e-83bd8e2a2557.mov
Possible improvements can be made, which will take 2 steps (Thanks to
@wolfogre for advice):
Step 1：
Backend: Add a new api, which returns a limit of 30 posters with matched
prefix.
Frontend: Change the search behavior from frontend search(fomantic
search) to backend search(when input is changed, send a request to get
authors matching the current search prefix)
Step 2：
Backend: Optimize the api in step 1 using indexer to support fuzzy
search.
This PR is implements the first step. The main changes:
1. Added api: `GET /{type:issues|pulls}/posters` , which return a limit
of 30 users with matched prefix (prefix sent as query). If
`DEFAULT_SHOW_FULL_NAME` in `custom/conf/app.ini` is set to true, will
also include fullnames fuzzy search.
2. Added a tooltip saying "Shows a maximum of 30 users" to the author
search dropdown
3. Change the search behavior from frontend search to backend search
After:
https://user-images.githubusercontent.com/17645053/229430960-f88fafd8-fd5d-4f84-9df2-2677539d5d08.mov
Fixes: https://github.com/go-gitea/gitea/issues/22586
---------
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Co-authored-by: silverwind <me@silverwind.io>
I neglected that the `NameKey` of `Unit` is not only for translation,
but also configuration. So it should be `repo.actions` to maintain
consistency.
## ⚠️ BREAKING ⚠️
If users already use `actions.actions` in `DISABLED_REPO_UNITS` or
`DEFAULT_REPO_UNITS`, it will be treated as an invalid unit key.
Follow #23633 and #23240Close#23814
Now we almost have a complete test set for Gitea's LocalStore.
This PR is still a quick fix for the legacy locale system (see the
TODOs), to resolve the problems fundamentally, it needs more work in the
future.
Closes#20955
This PR adds the possibility to disable blank Issues, when the Repo has
templates. This can be done by creating the file
`.gitea/issue_config.yaml` with the content `blank_issues_enabled` in
the Repo.
Adds API endpoints to manage issue/PR dependencies
* `GET /repos/{owner}/{repo}/issues/{index}/blocks` List issues that are
blocked by this issue
* `POST /repos/{owner}/{repo}/issues/{index}/blocks` Block the issue
given in the body by the issue in path
* `DELETE /repos/{owner}/{repo}/issues/{index}/blocks` Unblock the issue
given in the body by the issue in path
* `GET /repos/{owner}/{repo}/issues/{index}/dependencies` List an
issue's dependencies
* `POST /repos/{owner}/{repo}/issues/{index}/dependencies` Create a new
issue dependencies
* `DELETE /repos/{owner}/{repo}/issues/{index}/dependencies` Remove an
issue dependency
Closes https://github.com/go-gitea/gitea/issues/15393Closes#22115
Co-authored-by: Andrew Thornton <art27@cantab.net>
Close#20553
There were already a lot of functions powered by JavaScript in Gitea.
Without JavaScript, the Gitea Web UI almost doesn't work (only some
static links work ....)
It use old en-US locales as reference, fill the old other locales with
new locales.
----
## More broken translations
Many translations are still broken. The reason is: at the last time
restoring the ini to crowdin, many semicolon are treated as comments.
Two kinds of broken strings:
### Some translations can be re-translated
<details>
```
skipping options/locale/locale_si-LK.ini org teams.add_nonexistent_repo
skipping options/locale/locale_tr-TR.ini repo commits.search.tooltip
skipping options/locale/locale_es-ES.ini repo settings.trust_model.committer.desc
skipping options/locale/locale_es-ES.ini admin dashboard.new_version_hint
skipping options/locale/locale_pt-PT.ini org teams.add_nonexistent_repo
skipping options/locale/locale_hu-HU.ini install require_sign_in_view_popup
skipping options/locale/locale_hu-HU.ini repo migrate.invalid_local_path
skipping options/locale/locale_id-ID.ini repo migrate.invalid_local_path
skipping options/locale/locale_id-ID.ini org teams.add_nonexistent_repo
skipping options/locale/locale_de-DE.ini repo settings.protect_protected_file_patterns_desc
```
</details>
So this PR also does some small changes on them, to trigger the
re-translation.
### The `locale_el-GR.ini` contains many broken tranlsations
I guess we should reset them from crowdin side, then translators can
re-translate them.
----
Update: in latest main, the strings have been fixed.
## TODO
Update: the el-GR translators have done great job and fixes these broken
translations.
<details>
Merge this PR ASAP and upload `locale_el-GR.ini` to crowdin to remove
broken strings.
Out-dated, fixed in main.
</details>
---------
Co-authored-by: delvh <dev.lh@web.de>
This PR follows #22599 and #23450
The major improvements:
1. The `aria-*.js` are totally transparent now, no need to call
`attachDropdownAria` explicitly anymore.
* It hooks the `$.fn.checkbox` and `$.fn.dropdown`, then our patch
works.
* It makes all dynamically generated checkbox/dropdown work with a11y
without any change
* eg: the `conversation.find('.dropdown').dropdown();` in `repo-diff.js`
2. Since it's totally transparent now, it could be easier to modify or
remove in the future.
3. It handles all selection labels as well (by onLabelCreate), so it
supports "multiple selection dropdown" now.
* It partially completes one of my TODOs: `TODO: multiple selection is
not supported yet.`
4. The code structure is clearer, code blocks are splitted into
different functions.
* The old `attachOneDropdownAria` was splitted into separate functions.
* It makes it easier to add more fine tunes in the future, and co-work
with contributors.
6. The code logic is similar as before, only two new parts:
1. the `ariaCheckboxFn` and `ariaDropdownFn` functions
2. the `onLabelCreate` and `updateSelectionLabel` functions
In `aria-dropdown.js` I had to mix jQuery and Vanilla JS somewhat, I
think the code is still understandable, otherwise the code would be much
more complex to read.
Thanks to fsologureng for the idea about "improving the 'delete icon'
with aria attributes".
If there is anything unclear or incorrect, feel free to ask and discuss,
or propose new PRs for it.
This PR adds support for reflogs on all repositories. It does this by
adding a global configuration entry.
Implements #14865
---------
Signed-off-by: Philip Peterson <philip.c.peterson@gmail.com>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
When there is an error creating a new openIDConnect authentication
source try to handle the error a little better.
Close#23283
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
## TLDR
* Fix the broken page / broken image problem when click "Install"
* Close#20089
* Fix the Password Hash Algorithm display problem for #22942
* Close#23183
* Close#23184
## Details
### The broken page / broken image problem when click "Install"
(Redirect failed after install gitea #23184)
Before: when click "install", all new requests will fail, because the
server has been restarted. Users just see a broken page with broken
images, sometimes the server is not ready but the user would have been
redirect to "/user/login" page, then the users see a new broken page
(connection refused or something wrong ...)
After: only check InstallLock=true for necessary handlers, and sleep for
a while before restarting the server, then the browser has enough time
to load the "post-install" page. And there is a script to check whether
"/user/login" is ready, the user will only be redirected to the login
page when the server is ready.
### During new instance setup make 'Gitea Base URL' filled from
window.location.origin #20089
If the "app_url" input contains `localhost` (the default value from
config), use current window's location href as the `app_url` (aka
ROOT_URL)
### Fix the Password Hash Algorithm display problem for "Provide the
ability to set password hash algorithm parameters #22942"
Before: the UI shows `pbkdf2$50000$50`
<details>
</details>
After: the UI shows `pbkdf2`
<details>
</details>
### GET data: net::ERR_INVALID_URL #23183
Cause by empty `data:` in `<link rel="manifest"
href="data:{{.ManifestData}}">`
---------
Co-authored-by: Jason Song <i@wolfogre.com>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
The locales of Gitea has been broken for long time, till now, it's still
not fully fixed.
One of the root problems is that the `ini` library is quite quirky and
the `update-locales` script doesn't work well for all cases.
This PR fixes the `update-locales` script to make it satisfy `ini`
library and the crowdin.
See the comments for more details.
The `locale_zh-CN.ini` is an example, it comes from crowdin and is
processed by the new `update-locales.sh`. Especially see the `feed_of`:
https://github.com/go-gitea/gitea/pull/23240/files#diff-321f6ca4eae1096eba230e93c4740f9903708afe8d79cf2e57f4299786c4528bR268
Extract from #11669 and enhancement to #22585 to support exclusive
scoped labels in label templates
* Move label template functionality to label module
* Fix handling of color codes
* Add Advanced label template
This includes pull requests that you approved, requested changes or
commented on. Currently such pull requests are not visible in any of the
filters on /pulls, while they may need further action like merging, or
prodding the author or reviewers.
Especially when working with a large team on a repository it's helpful
to get a full overview of pull requests that may need your attention,
without having to sift through the complete list.
The API to create tokens is missing the ability to set the required
scopes for tokens, and to show them on the API and on the UI.
This PR adds this functionality.
Signed-off-by: Andrew Thornton <art27@cantab.net>
Add a new "exclusive" option per label. This makes it so that when the
label is named `scope/name`, no other label with the same `scope/`
prefix can be set on an issue.
The scope is determined by the last occurence of `/`, so for example
`scope/alpha/name` and `scope/beta/name` are considered to be in
different scopes and can coexist.
Exclusive scopes are not enforced by any database rules, however they
are enforced when editing labels at the models level, automatically
removing any existing labels in the same scope when either attaching a
new label or replacing all labels.
In menus use a circle instead of checkbox to indicate they function as
radio buttons per scope. Issue filtering by label ensures that only a
single scoped label is selected at a time. Clicking with alt key can be
used to remove a scoped label, both when editing individual issues and
batch editing.
Label rendering refactor for consistency and code simplification:
* Labels now consistently have the same shape, emojis and tooltips
everywhere. This includes the label list and label assignment menus.
* In label list, show description below label same as label menus.
* Don't use exactly black/white text colors to look a bit nicer.
* Simplify text color computation. There is no point computing luminance
in linear color space, as this is a perceptual problem and sRGB is
closer to perceptually linear.
* Increase height of label assignment menus to show more labels. Showing
only 3-4 labels at a time leads to a lot of scrolling.
* Render all labels with a new RenderLabel template helper function.
Label creation and editing in multiline modal menu:
* Change label creation to open a modal menu like label editing.
* Change menu layout to place name, description and colors on separate
lines.
* Don't color cancel button red in label editing modal menu.
* Align text to the left in model menu for better readability and
consistent with settings layout elsewhere.
Custom exclusive scoped label rendering:
* Display scoped label prefix and suffix with slightly darker and
lighter background color respectively, and a slanted edge between them
similar to the `/` symbol.
* In menus exclusive labels are grouped with a divider line.
---------
Co-authored-by: Yarden Shoham <hrsi88@gmail.com>
Co-authored-by: Lauris BH <lauris@nix.lv>
The "Reference in new issue" option shows up in a menu when looking at
pull requests. All the other options there follow the "Title case":
* Copy Link
* Quote Reply
* Edit
This patch makes sure this option also follow the Title case.
Screenshot of how it looks without this patch:
Co-authored-by: Dalai Felinto <dalai@blender.org>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Add setting to allow edits by maintainers by default, to avoid having to
often ask contributors to enable this.
This also reorganizes the pull request settings UI to improve clarity.
It was unclear which checkbox options were there to control available
merge styles and which merge styles they correspond to.
Now there is a "Merge Styles" label followed by the merge style options
with the same name as in other menus. The remaining checkboxes were
moved to the bottom, ordered rougly by typical order of operations.
---------
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Original Issue: https://github.com/go-gitea/gitea/issues/22102
This addition would be a big benefit for design and art teams using the
issue tracking.
The preview will be the latest "image type" attachments on an issue-
simple, and allows for automatic updates of the cover image as issue
progress is made!
This would make Gitea competitive with Trello... wouldn't it be amazing
to say goodbye to Atlassian products? Ha.
First image is the most recent, the SQL will fetch up to 5 latest images
(URL string).
All images supported by browsers plus upcoming formats: *.avif *.bmp
*.gif *.jpg *.jpeg *.jxl *.png *.svg *.webp
The CSS will try to center-align images until it cannot, then it will
left align with overflow hidden. Single images get to be slightly
larger!
Tested so far on: Chrome, Firefox, Android Chrome, Android Firefox.
Current revision with light and dark themes:
---------
Co-authored-by: Jason Song <i@wolfogre.com>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Co-authored-by: delvh <dev.lh@web.de>
Fixes#19555
Test-Instructions:
https://github.com/go-gitea/gitea/pull/21441#issuecomment-1419438000
This PR implements the mapping of user groups provided by OIDC providers
to orgs teams in Gitea. The main part is a refactoring of the existing
LDAP code to make it usable from different providers.
Refactorings:
- Moved the router auth code from module to service because of import
cycles
- Changed some model methods to take a `Context` parameter
- Moved the mapping code from LDAP to a common location
I've tested it with Keycloak but other providers should work too. The
JSON mapping format is the same as for LDAP.
---------
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
This PR fixes two problems. One is when filter repository issues, only
repository level projects are listed. Another is if you list open
issues, only open projects will be displayed in filter options and if
you list closed issues, only closed projects will be displayed in filter
options.
In this PR, both repository level and org/user level projects will be
displayed in filter, and both open and closed projects will be listed as
filter items.
---------
Co-authored-by: John Olheiser <john.olheiser@gmail.com>
Co-authored-by: zeripath <art27@cantab.net>
Co-authored-by: delvh <dev.lh@web.de>
The error reported when a user passes a private ssh key as their ssh
public key is not very nice.
This PR improves this slightly.
Ref #22693
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: delvh <dev.lh@web.de>
Fixes#22183
Replaces #22187
This PR adds secrets for users. I refactored the files for organizations
and repos to use the same logic and templates. I splitted the secrets
from deploy keys again and reverted the fix from #22187.
---------
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
EDIT: The main change of this PR was resolved by #22599. This
complements that PR for some cases without label and complicated layout
to be added.
NOTE: Contributed by @Forgejo.
Currently only a single project like milestone, not multiple like
labels.
Implements #14298
Code by @brechtvl
---------
Co-authored-by: Brecht Van Lommel <brecht@blender.org>
Added ARIA navigation landmark to navigation bar and aria label for both
nav bar and footer.
Contributed by @forgejo.
---------
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
On activating local accounts, the error message didn't differentiate
between using a wrong or expired token, or a wrong password. The result
could already be obtained from the behaviour (different screens were
presented), but the error message was misleading and lead to confusion
for new users on Codeberg with Forgejo.
Now, entering a wrong password for a valid token prints a different
error message.
The problem was introduced in 0f14f69e60.
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
This PR adds the support for scopes of access tokens, mimicking the
design of GitHub OAuth scopes.
The changes of the core logic are in `models/auth` that `AccessToken`
struct will have a `Scope` field. The normalized (no duplication of
scope), comma-separated scope string will be stored in `access_token`
table in the database.
In `services/auth`, the scope will be stored in context, which will be
used by `reqToken` middleware in API calls. Only OAuth2 tokens will have
granular token scopes, while others like BasicAuth will default to scope
`all`.
A large amount of work happens in `routers/api/v1/api.go` and the
corresponding `tests/integration` tests, that is adding necessary scopes
to each of the API calls as they fit.
- [x] Add `Scope` field to `AccessToken`
- [x] Add access control to all API endpoints
- [x] Update frontend & backend for when creating tokens
- [x] Add a database migration for `scope` column (enable 'all' access
to past tokens)
I'm aiming to complete it before Gitea 1.19 release.
Fixes#4300
Addition to #22056
This PR adds a hint to mail text if replies are supported.
I can't tell if the text structure is supported in every language. Maybe
we need to put the whole line in the translation file and use
parameters.
This PR adds a task to the cron service to allow garbage collection of
LFS meta objects. As repositories may have a large number of
LFSMetaObjects, an updated column is added to this table and it is used
to perform a generational GC to attempt to reduce the amount of work.
(There may need to be a bit more work here but this is probably enough
for the moment.)
Fix#7045
Signed-off-by: Andrew Thornton <art27@cantab.net>
This PR introduce glob match for protected branch name. The separator is
`/` and you can use `*` matching non-separator chars and use `**` across
separator.
It also supports input an exist or non-exist branch name as matching
condition and branch name condition has high priority than glob rule.
Should fix#2529 and #15705
screenshots
<img width="1160" alt="image"
src="https://user-images.githubusercontent.com/81045/205651179-ebb5492a-4ade-4bb4-a13c-965e8c927063.png">
Co-authored-by: zeripath <art27@cantab.net>
Fixes#19091
Add Feed for Releases and Tags, can be accessed through
`reponame/releases.rss`, `reponame/releases.atom`, `reponame/tags.rss`,
and `reponame/tags.atom`
Signed-off-by: Reo <reo_999@proton.me>
This PR adds a button to allow quickly clearing the merge message of a
PR. The button will remove everything but the git trailers.
I found myself often pruning the commit message before merging,
especially for PRs generated by renovate - renovate puts a very long and
detailed comment with the full changelog in each PR it opens. This
clutters the commit message. However, I want to explicitly preserve the
git commit trailers. Doing this manually works, but having a button is a
lot easier.
Screenshot:
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Co-authored-by: delvh <dev.lh@web.de>
When I'm trying to use a user which has no repo numbers limit to create
a repo for a org which has reached the limit, it says "You have already
reached your limit of 5 repositories."
That's confusing. "I" haven't reached the limit, the owner has.
<img width="828" alt="xnip_2022-11-10_11-57-45"
src="https://user-images.githubusercontent.com/9418365/200997290-d0819e40-fb10-4c37-917c-167e2070b4f9.png">
Co-authored-by: Lauris BH <lauris@nix.lv>
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
Close https://github.com/go-gitea/gitea/issues/21640
Before: Gitea can create users like ".xxx" or "x..y", which is not
ideal, it's already a consensus that dot filenames have special
meanings, and `a..b` is a confusing name when doing cross repo compare.
After: stricter
Co-authored-by: Jason Song <i@wolfogre.com>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Co-authored-by: delvh <dev.lh@web.de>
_This is a different approach to #20267, I took the liberty of adapting
some parts, see below_
## Context
In some cases, a weebhook endpoint requires some kind of authentication.
The usual way is by sending a static `Authorization` header, with a
given token. For instance:
- Matrix expects a `Bearer <token>` (already implemented, by storing the
header cleartext in the metadata - which is buggy on retry #19872)
- TeamCity #18667
- Gitea instances #20267
- SourceHut https://man.sr.ht/graphql.md#authentication-strategies (this
is my actual personal need :)
## Proposed solution
Add a dedicated encrypt column to the webhook table (instead of storing
it as meta as proposed in #20267), so that it gets available for all
present and future hook types (especially the custom ones #19307).
This would also solve the buggy matrix retry #19872.
As a first step, I would recommend focusing on the backend logic and
improve the frontend at a later stage. For now the UI is a simple
`Authorization` field (which could be later customized with `Bearer` and
`Basic` switches):
The header name is hard-coded, since I couldn't fine any usecase
justifying otherwise.
## Questions
- What do you think of this approach? @justusbunsi @Gusted @silverwind
- ~~How are the migrations generated? Do I have to manually create a new
file, or is there a command for that?~~
- ~~I started adding it to the API: should I complete it or should I
drop it? (I don't know how much the API is actually used)~~
## Done as well:
- add a migration for the existing matrix webhooks and remove the
`Authorization` logic there
_Closes #19872_
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Co-authored-by: Gusted <williamzijl7@hotmail.com>
Co-authored-by: delvh <dev.lh@web.de>
I found myself wondering whether a PR I scheduled for automerge was
actually merged. It was, but I didn't receive a mail notification for it
- that makes sense considering I am the doer and usually don't want to
receive such notifications. But ideally I want to receive a notification
when a PR was merged because I scheduled it for automerge.
This PR implements exactly that.
The implementation works, but I wonder if there's a way to avoid passing
the "This PR was automerged" state down so much. I tried solving this
via the database (checking if there's an automerge scheduled for this PR
when sending the notification) but that did not work reliably, probably
because sending the notification happens async and the entry might have
already been deleted. My implementation might be the most
straightforward but maybe not the most elegant.
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: Lauris BH <lauris@nix.lv>
Co-authored-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
The OAuth spec [defines two types of
client](https://datatracker.ietf.org/doc/html/rfc6749#section-2.1),
confidential and public. Previously Gitea assumed all clients to be
confidential.
> OAuth defines two client types, based on their ability to authenticate
securely with the authorization server (i.e., ability to
> maintain the confidentiality of their client credentials):
>
> confidential
> Clients capable of maintaining the confidentiality of their
credentials (e.g., client implemented on a secure server with
> restricted access to the client credentials), or capable of secure
client authentication using other means.
>
> **public
> Clients incapable of maintaining the confidentiality of their
credentials (e.g., clients executing on the device used by the resource
owner, such as an installed native application or a web browser-based
application), and incapable of secure client authentication via any
other means.**
>
> The client type designation is based on the authorization server's
definition of secure authentication and its acceptable exposure levels
of client credentials. The authorization server SHOULD NOT make
assumptions about the client type.
https://datatracker.ietf.org/doc/html/rfc8252#section-8.4
> Authorization servers MUST record the client type in the client
registration details in order to identify and process requests
accordingly.
Require PKCE for public clients:
https://datatracker.ietf.org/doc/html/rfc8252#section-8.1
> Authorization servers SHOULD reject authorization requests from native
apps that don't use PKCE by returning an error message
Fixes#21299
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
