It is fine to use MockVariableValue to change a setting such as:
defer test.MockVariableValue(&setting.Mirror.Enabled, true)()
But when testing for errors and mocking a function, multiple variants
of the functions will be used, not just one. MockProtect a function
will make sure that when the test fails it always restores a sane
version of the function. For instance:
defer test.MockProtect(&mirror_service.AddPushMirrorRemote)()
mirror_service.AddPushMirrorRemote = mockOne
do some tests that may fail
mirror_service.AddPushMirrorRemote = mockTwo
do more tests that may fail
- rewrite a lot of hints on install page
- make sure checkboxes don't hide useful information behind hover
This is good for compactness but makes first-time installation more painful than it should be. BTW, this was inherited from Gogs.
- update related translation keys (will require Weblate sync to merge)
- make sure string locations in en-US.ini make sense. Unfortunately, makes viewing changes harder, but I've attached screenshots
## Preview
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/3961
Reviewed-by: Otto <otto@codeberg.org>
We wanted to be able to use the IAM role provided by the EC2 instance
metadata in order to access S3 via the Minio configuration. To do this,
a new credentials chain is added that will check the following locations
for credentials when an access key is not provided. In priority order,
they are:
1. MINIO_ prefixed environment variables
2. AWS_ prefixed environment variables
3. a minio credentials file
4. an aws credentials file
5. EC2 instance metadata
(cherry picked from commit c0880e7695346997c6a93f05cd01634cb3ad03ee)
Conflicts:
docs/content/administration/config-cheat-sheet.en-us.md
does not exist in Forgejo
This PR split the `Board` into two parts. One is the struct has been
renamed to `Column` and the second we have a `Template Type`.
But to make it easier to review, this PR will not change the database
schemas, they are just renames. The database schema changes could be in
future PRs.
---------
Co-authored-by: silverwind <me@silverwind.io>
Co-authored-by: yp05327 <576951401@qq.com>
(cherry picked from commit 98751108b11dc748cc99230ca0fc1acfdf2c8929)
Conflicts:
docs/content/administration/config-cheat-sheet.en-us.md
docs/content/index.en-us.md
docs/content/installation/comparison.en-us.md
docs/content/usage/permissions.en-us.md
non existent files
options/locale/locale_en-US.ini
routers/web/web.go
templates/repo/header.tmpl
templates/repo/settings/options.tmpl
trivial context conflicts
Fix #25897
Fix #30322#29464 cannot handle some complex `if` conditions correctly because it
only checks `always()` literally. In fact, it's not easy to evaluate the
`if` condition on the Gitea side because evaluating it requires a series
of contexts. But act_runner is able to evaluate the `if` condition
before running the job (for more information, see
[`gitea/act`](517d11c671/pkg/runner/run_context.go (L739-L753)))
. So we can use act_runner to check the `if` condition.
In this PR, how to handle a blocked job depends on its `needs` and `if`:
- If not all jobs in `needs` completed successfully and the job's `if`
is empty, set the job status to `StatusSkipped`
- In other cases, the job status will be set to `StatusWaiting`, and
then act_runner will check the `if` condition and run the job if the
condition is met
(cherry picked from commit 31a0c4dfb4156a7b4d856cceae1e61c7fc1a4a1b)
This exception existed for both instance admins and repo admins
before ApplyToAdmins was introduced in
79b7089360.
It should have been kept for instance admins only because they are not
subject to permission checks.
Verify variations of branch protection that are in play when merging a
pull request as:
* instance admin
* repository admin / owner
* user with write permissions on the repository
In all cases the result is expected to be the same when merging
the pull request via:
* API
* web
Although the implementations are different.
* split into testPullMergeForm which can be called directly if
the caller wants to specify extra parameters.
* testPullMergeForm can expect something different than StatusOK
* http.StatusMethodNotAllowed can be expected: only retry if the
error message is "Please try again later"
* split into doAPIMergePullRequestForm which can be called directly if
the caller wants to specify extra parameters.
- make sure margins are all consistent and good, elements are not too close or too apart
- this also applies to "Show commit body" button
- remove unused code. The class `commit-status-link` doesn't exist in templates, nor I could find it on any related pages in case it's generated in runtime
## Preview
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/3948
Reviewed-by: Beowulf <beowulf@noreply.codeberg.org>
Use the same padding horizontally and vertically, so the views like readme look a bit nicer. Just slightly adjusted two values, nothing really test-able here.
## Motivation
I came to the conclusion that they should be the same myself, later I checked GitHub and it turned out to also use the same paddings. I would like to notice that the padding here (2em = 32px) is the same as GitHub uses too.
I find this as a logical UI change because the paddings are usually same on both axis across the UI (like on PR sidebar).
Also updated paddings for when the files are shown in profile, but copied the `1.5em` that GitHub uses. This, once again, makes sense, because the overview markdown isn't the primary content, or as primary as the readme on the repo is, taking the full usable width.
## Preview
https://codeberg.org/attachments/55f6685c-1978-410a-a17b-9fac91f0642e
---
https://codeberg.org/attachments/d9016a1c-13cf-4ea6-a8e4-2619d93f3560
## Note
`.non-diff-file-content .plain-text` is left untouched with `1em 2em`, because the plaintext seems to add it's own margins, so it would make it look worse.
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/3944
Reviewed-by: Otto <otto@codeberg.org>
Reviewed-by: Beowulf <beowulf@noreply.codeberg.org>
Remove CSS code that was made unused by some changes in Gitea. I was working on a layout change here but was bothered a bit by these. I dug a bit into the git history to find out how they were made unused but it's relatively uneasy.
- remove rule that was setting `width: 100%;`: the exactly same selector setting this exact value is duplicated below
- remove rules with `followers` in selectors: we don't use this class in templates (would be nice if someone double-checks)
- my editor forced EoF fix
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/3937
Reviewed-by: Beowulf <beowulf@noreply.codeberg.org>
