This PR will fix the bug on migration v283 and also the CI hidden
behaviour.
The reason is on the Makefile
`GITEA_ROOT="$(CURDIR)" GITEA_CONF=tests/mysql.ini $(GO) test
$(GOTESTFLAGS) -tags='$(TEST_TAGS)' $(MIGRATE_TEST_PACKAGES)` will
return the error exit code.
But
`for pkg in $(shell $(GO) list
code.gitea.io/gitea/models/migrations/...); do \
GITEA_ROOT="$(CURDIR)" GITEA_CONF=tests/mysql.ini $(GO) test
$(GOTESTFLAGS) -tags '$(TEST_TAGS)' $$pkg; \
done`
Yarden Shoham [Thu, 7 Mar 2024 07:28:33 +0000 (09:28 +0200)]
Remove jQuery AJAX from the repo editor (#29636)
# Preview Tab
- Removed the jQuery AJAX call and replaced with our fetch wrapper
- Tested the preview tab functionality and it works as before
# Diff Tab
- Removed the jQuery AJAX call and replaced with htmx
- Tested the diff tab functionality and it works as before
## htmx Attributes
- `hx-post="{{.RepoLink}}..."`: make a POST request to the endpoint
- `hx-indicator=".tab[data-tab='diff']"`: attach the loading indicator
to the tab body
- `hx-target=".tab[data-tab='diff']"`: target the tab body for swapping
with the response
- `hx-swap="innerHTML"`: swap the target's inner HTML
- `hx-include="#edit_area"`: include the value of the textarea (content)
in the request body
- `hx-vals='{"context":"{{.BranchLink}}"}'`: include the context in the
request body
- `hx-params="context,content"`: include only these keys in the request
body
# Demo using `fetch` and `htmx` instead of jQuery AJAX
Earl Warren [Wed, 6 Mar 2024 16:10:06 +0000 (00:10 +0800)]
Add download URL for executable files (#28260)
Consider executable files as a valid case when returning a DownloadURL for them.
They are just regular files with the difference being the executable permission bit being set.
Rafael Heard [Wed, 6 Mar 2024 14:20:26 +0000 (09:20 -0500)]
Move all login and account creation page labels to be above inputs (#29432)
There are a few inconsistencies within Gitea and this PR addresses one
of them. This PR updates the sign-in page layout, including the register
and openID tabs, to match the layout of the settings pages
(/user/settings) for more consistency.
This PR updates the following routes:
`/user/login`
`/user/sign_up`
`/user/login/openid`
`/user/forgot_password`
`/user/link_account`
`/user/recover_account`
Lunny Xiao [Wed, 6 Mar 2024 12:17:19 +0000 (20:17 +0800)]
Cache repository default branch commit status to reduce query on commit status table (#29444)
After repository commit status has been introduced on dashaboard, the
most top SQL comes from `GetLatestCommitStatusForPairs`.
This PR adds a cache for the repository's default branch's latest
combined commit status. When a new commit status updated, the cache will
be marked as invalid.
wxiaoguang [Wed, 6 Mar 2024 11:50:39 +0000 (19:50 +0800)]
Avoid unexpected panic in graceful manager (#29629)
There is a fundamental design problem of the "manager" and the "wait
group".
If nothing has started, the "Wait" just panics: sync: WaitGroup is
reused before previous Wait has returned
There is no clear solution besides a complete rewriting of the "manager"
If there are some mistakes in the app.ini, end users would just see the
"panic", but not the real error messages. A real case: #27643
This PR is just a quick fix for the annoying panic problem.
Jason Song [Wed, 6 Mar 2024 09:56:04 +0000 (17:56 +0800)]
Detect broken git hooks (#29494)
Detect broken git hooks by checking if the commit id of branches in DB
is the same with the git repo.
It can help #29338 #28277 and maybe more issues.
Users could complain about actions, webhooks, and activities not
working, but they were not aware that it is caused by broken git hooks
unless they could see a warning.
Jason Song [Wed, 6 Mar 2024 08:47:52 +0000 (16:47 +0800)]
Sync branches to DB immediately when handle git hook calling (#29493)
Unlike other async processing in the queue, we should sync branches to
the DB immediately when handling git hook calling. If it fails, users
can see the error message in the output of the git command.
It can avoid potential inconsistency issues, and help #29494.
No known official action of GitHub makes use of this claim.
Current releases throw an error when configure to use actions cache
```
| ERROR: failed to solve: failed to configure gha cache exporter: invalid token without access controls
| ##[error]buildx failed with: ERROR: failed to solve: failed to configure gha cache exporter: invalid token without access controls
```
silverwind [Tue, 5 Mar 2024 04:31:29 +0000 (05:31 +0100)]
Adjust tailwind content globs (#29596)
Tailwind content is not going to appear in `web_src/css`,
`web_src/fomantic` or `web_src/svg` or the JSON templates, so we don't
need to have tailwind scan these directories which will speed up the
build.
Denys Konovalov [Mon, 4 Mar 2024 20:49:21 +0000 (21:49 +0100)]
Fix projects mode bugs (#29593)
Fix for regressions introduced by #28805
Enabled projects on repos created before the PR weren't detected. Also,
the way projects mode was detected in settings didn't match the way it
was detected on permission check, which leads to confusion.
silverwind [Mon, 4 Mar 2024 20:24:12 +0000 (21:24 +0100)]
Regenerate fomantic lockfile and build it with our browserslist (#29560)
1. Make fomantic build use [our
browserslist](https://github.com/go-gitea/gitea/blob/e3524c63d6d42865ea8288af89b372544d35474b/package.json#L99).
I found no other way than to sed-replace into it's js, the normal
browserlist config files do not work. The effect of this change is the
removal of some uneeded CSS vendor prefixes.
2. Regenerate `web_src/fomantic/package-lock.json`, this might shut up
some security scanners.
Lunny Xiao [Mon, 4 Mar 2024 13:51:12 +0000 (21:51 +0800)]
Document that all unmerged feature PRs will be moved to next milestone when the feature freeze time comes (#29578)
Some contributors may be surprised when moving the feature PRs to the
next release when the feature freeze time comes. So this PR documents
the habits we have done for feature freeze so people expect the
maintainers' behaviors. We are sorry for disturbing you with the
milestones changes.
A feature freeze announcement should be published 2 or 3 weeks before
the feature freeze by maintainers.
wxiaoguang [Mon, 4 Mar 2024 12:02:45 +0000 (20:02 +0800)]
Remove unnecessary SanitizeHTML from code (#29575)
* "mail/issue/default.tmpl": the body is rendered by backend
`markdown.RenderString() HTML`, it has been already sanitized
* "repo/settings/webhook/base_list.tmpl": "Description" is prepared by
backend `ctx.Tr`, it doesn't need to be sanitized
KN4CK3R [Mon, 4 Mar 2024 08:16:03 +0000 (09:16 +0100)]
Add user blocking (#29028)
Fixes #17453
This PR adds the abbility to block a user from a personal account or
organization to restrict how the blocked user can interact with the
blocker. The docs explain what's the consequence of blocking a user.
Denys Konovalov [Mon, 4 Mar 2024 02:56:52 +0000 (03:56 +0100)]
Enable/disable owner and repo projects independently (#28805)
Part of #23318
Add menu in repo settings to allow for repo admin to decide not just if
projects are enabled or disabled per repo, but also which kind of
projects (repo-level/owner-level) are enabled. If repo projects
disabled, don't show the projects tab.
6543 [Mon, 4 Mar 2024 00:37:00 +0000 (01:37 +0100)]
Extend issue template yaml engine (#29274)
Add new option:
`visible`: witch can hide a specific field of the form or the created
content afterwards
It is a string array witch can contain `form` and `content`. If only
`form` is present, it wont show up in the created issue afterwards and
the other way around. By default it sets both except for markdown
As they are optional and github don't have any similar thing, it is non
breaking and also do not conflict with it.
With this you can:
- define "post issue creation" elements like a TODO list to track an
issue state
- make sure to have a checkbox that reminds the user to check for a
thing but dont have it in the created issue afterwards
- define markdown for the created issue (was the downside of using yaml
instead of md in the past)
- ...
## Demo
```yaml
name: New Contribution
description: External Contributor creating a pull
body:
- type: checkboxes
id: extern-todo
visible: [form]
attributes:
label: Contribution Guidelines
options:
- label: I checked there exist no similar feature to be extended
required: true
- label: I did read the CONTRIBUTION.MD
required: true
- type: checkboxes
id: intern-todo
visible: [content]
attributes:
label: Maintainer Check-List
options:
- label: Does this pull follow the KISS principe
- label: Checked if internal bord was notifyed
# ....
```
[Demo
Video](https://cloud.obermui.de/s/tm34fSAbJp9qw9z/download/vid-20240220-152751.mkv)
---
*Sponsored by Kithara Software GmbH*
---------
Co-authored-by: John Olheiser <john.olheiser@gmail.com> Co-authored-by: delvh <dev.lh@web.de>
Yarden Shoham [Sat, 2 Mar 2024 19:31:59 +0000 (21:31 +0200)]
Fix elipsis button not working if the last commit loading is deferred (#29544)
Before this change, if we had more than 200 entries being deferred in
loading, the entire table would get replaced thus losing any event
listeners attached to the elements within the table, such as the elipsis
button and commit list with tippy.
With this change we remove the previous javascript code that replaced
the table and use htmx to replace the table.
htmx attributes added:
- `hx-indicator="tr.notready td.message span"`: attach the loading
spinner to the files whose last commit is still being loaded
- `hx-trigger="load"` trigger the request-replace behavior as soon as
possible
- `hx-swap="morph"`: use the idiomorph morphing algorithm, this is the
thing that makes it so the elipsis button event listener is kept during
the replacement, fixing the bug because we don't actually replace the
table, only modifying it
- `hx-post="{{.LastCommitLoaderURL}}"`: make a post request to this url
to get the table with all of the commit information
As part of this change I removed the handling of partial replacement in
the case we have less than 200 "not ready" files. The first reason is
that I couldn't make htmx replace only a subset of returned elements,
the second reason is that we have a cache implemented in the backend
already so the only cost added is that we query the cache a few times
(which is sure to be populated due to the initial request), and the last
reason is that since the last refactor of this functionality that
removed jQuery we don't properly send the "not ready" entries as the
backend expects `FormData` with `f[]` and we send a JSON with `f` so we
always query for all rows anyway.
wxiaoguang [Sat, 2 Mar 2024 16:07:54 +0000 (00:07 +0800)]
Fix queue worker incorrectly stopped when there are still more items in the queue (#29532)
Without `case <-t.C`, the workers would stop incorrectly, the test won't
pass. For the worse case, there might be only one running worker
processing the queue items for long time because other workers are
stopped. The root cause is related to the logic of doDispatchBatchToWorker.
It isn't a serious problem at the moment, so keep it as-is.
Lunny Xiao [Sat, 2 Mar 2024 14:03:39 +0000 (22:03 +0800)]
Fix a bug returning 404 when display a single tag with no release (#29466)
Partially caused by #29149
When use
```go
releases, err := getReleaseInfos(ctx, &repo_model.FindReleasesOptions{
ListOptions: db.ListOptions{Page: 1, PageSize: 1},
RepoID: ctx.Repo.Repository.ID,
TagNames: []string{ctx.Params("*")},
// only show draft releases for users who can write, read-only users shouldn't see draft releases.
IncludeDrafts: writeAccess,
})
```
replace
```go
release, err := repo_model.GetRelease(ctx, ctx.Repo.Repository.ID, ctx.Params("*"))
```
It missed `IncludeTags: true,`. That means this bug will be occupied only when the release is a tag.
This PR will fix
- Get the right tag record when it's not a release
- Display correct tag tab but not release tag when it's a tag.
- The button will bring the tag name to the new page when it's a single tag page
- the new page will automatically hide the release target inputbox when the tag name is pre filled. This should be backport to v1.21.
charles [Sat, 2 Mar 2024 13:38:34 +0000 (21:38 +0800)]
Add a check for when the command is canceled by the program on Window… (#29538)
Close #29509
Windows, unlike Linux, does not have signal-specified exit codes.
Therefore, we should add a Windows-specific check for Windows. If we
don't do this, the logs will always show a failed status, even though
the command actually works correctly.
If you check the Go source code in exec_windows.go, you will see that it
always returns exit code 1.
wxiaoguang [Sat, 2 Mar 2024 12:02:34 +0000 (20:02 +0800)]
Fix incorrect subpath in links (#29535)
* `$referenceUrl`: it is constructed by "Issue.Link", which already has
the "AppSubURL"
* `window.location.href`: AppSubURL could be empty string, so it needs
the trailing slash
ChristopherHX [Sat, 2 Mar 2024 09:12:17 +0000 (10:12 +0100)]
Actions Artifacts v4 backend (#28965)
Fixes #28853
Needs both https://gitea.com/gitea/act_runner/pulls/473 and
https://gitea.com/gitea/act_runner/pulls/471 on the runner side and
patched `actions/upload-artifact@v4` / `actions/download-artifact@v4`,
like `christopherhx/gitea-upload-artifact@v4` and
`christopherhx/gitea-download-artifact@v4`, to not return errors due to
GHES not beeing supported yet.
Yarden Shoham [Sat, 2 Mar 2024 08:48:14 +0000 (10:48 +0200)]
Remove jQuery AJAX from common global functions (#29528)
- Removed all jQuery AJAX calls and replaced with our fetch wrapper
- Tested the locale change functionality and it works as before
- Tested the delete button functionality and it works as before
wxiaoguang [Fri, 1 Mar 2024 16:46:02 +0000 (00:46 +0800)]
Fix issue & comment history bugs (#29525)
* Follow #17746: `HasIssueContentHistory` should use expr builder to
make sure zero value (0) be respected.
* Add "doer" check to make sure `canSoftDeleteContentHistory` only be
called by sign-in users.
silverwind [Fri, 1 Mar 2024 16:12:21 +0000 (17:12 +0100)]
Fix incorrect diff expander for deletion of last lines in a file (#29501)
Fixes: https://github.com/go-gitea/gitea/issues/29498
I don't quite understand this code, but this change does seem to fix the
issue and I tested a number of diffs with it and saw no issue. The
function gets such value if last line is an addition:
sillyguodong [Fri, 1 Mar 2024 13:23:53 +0000 (21:23 +0800)]
Set pre-step status to `skipped` if job is skipped (#29489)
close #27496
1. Set pre-step (Set up job) status to `skipped` if job is skipped.
2. Apart from pre-step, the other steps should also be set to `skipped`.
The status of other steps are reported from the runner side. This will
be completed by this PR: https://gitea.com/gitea/act_runner/pulls/500
wxiaoguang [Fri, 1 Mar 2024 10:56:29 +0000 (18:56 +0800)]
Refactor the "attachments" sub-template data key to RenderedContent (#29517)
The value passed into "attachments" sub-template is from
"RedneredContent", so use the same name for consistent. And it makes
readers easy to know its data type.
wxiaoguang [Fri, 1 Mar 2024 07:11:51 +0000 (15:11 +0800)]
Refactor some Str2html code (#29397)
This PR touches the most interesting part of the "template refactoring".
1. Unclear variable type. Especially for "web/feed/convert.go":
sometimes it uses text, sometimes it uses HTML.
2. Assign text content to "RenderedContent" field, for example: `
project.RenderedContent = project.Description` in web/org/projects.go
3. Assign rendered content to text field, for example: `r.Note =
rendered content` in web/repo/release.go
4. (possible) Incorrectly calling `{{Str2html
.PackageDescriptor.Metadata.ReleaseNotes}}` in
package/content/nuget.tmpl, I guess the name Str2html misleads
developers to use it to "render string to html", but it only sanitizes.
if ReleaseNotes really contains HTML, then this is not a problem.
silverwind [Thu, 29 Feb 2024 05:11:11 +0000 (06:11 +0100)]
Lighten text colors on dark theme for increased contrast (#29481)
Improve contrast by lightening the text colors in dark theme by around
35%. Additionally, share some variables that had the same or similar
color, which will ease future theme creation.
silverwind [Wed, 28 Feb 2024 22:20:53 +0000 (23:20 +0100)]
Fix/Improve `processWindowErrorEvent` (#29407)
- `e.error` can be undefined in some cases which would raise an error
inside this error handler, fixed that.
- The displayed message mentions looking into the console, but in my
case of error from `ResizeObserver` there was nothing there, so add this
logging. I think this logging was once there but got lost during
refactoring.
silverwind [Wed, 28 Feb 2024 20:26:12 +0000 (21:26 +0100)]
Apply compact padding to small buttons with svg icons (#29471)
The buttons on the repo release tab were larger in height than on other
tabs because one of them contained the RSS icon which stretched the
button height by 3px. Workaround this problem by applying the "compact"
padding to any such button. They are within 0.4px in height now to
non-icon buttons.
Before:
<img width="406" alt="Screenshot 2024-02-28 at 15 30 23"
src="https://github.com/go-gitea/gitea/assets/115237/805bb93a-6fe4-40a0-82d1-03001bee8ecf">
After:
<img width="407" alt="Screenshot 2024-02-28 at 15 38 43"
src="https://github.com/go-gitea/gitea/assets/115237/27707588-890f-4852-ab08-105a57eda880">
For comparison, button on issue tab:
<img width="452" alt="Screenshot 2024-02-28 at 15 31 46"
src="https://github.com/go-gitea/gitea/assets/115237/74ac13d5-d016-49ba-9dd9-40ed32a748e9">
charles [Wed, 28 Feb 2024 20:23:49 +0000 (04:23 +0800)]
Fix counter display number incorrectly displayed on the page (#29448)
issue : #28239
The counter number script uses the 'checkbox' attribute to determine
whether an item is selected or not.
However, the input event only increments the counter value, and when
more items are displayed, it does not update all previously loaded
items.
As a result, the display becomes incorrect because it triggers the
update counter script, but checkboxes that are selected without the
'checked' attribute are not counted