cmd/dump.go uses os.Readdir to list the directory.
This is fine on its own, but TestAddRecursiveExclude in cmd/dump_test.go
depends on the order of the directory listing, which is where the issue
lays.
Directory listings using os.Readdir (lstat) don't actually guarantee an
order. They can differ due to a number of factors. Most notably the OS,
file system and settings.
As such, the test should not check the /order of the files/ added to the
archive, but instead simply check whether the archive /contains/ them.
So this is precisely what this commit does.
Note that only TestAddRecursiveExclude/File_inside_directory/No_exclude
has been observed to fail due to this, but all TestAddRecursiveExclude
subtests have been updated for consistency.
While trying to understand #1236, I was quite confused not to see the `Use Custom URLs` checkbox.
This checkbox disappeared in b95a893b22 (because `getElementById` does not expect a `#` as first char), fixed in 4e816e1326086ff0929c028f837f62ba1c747759.
After solving this, switching from `Nextcloud` to `OpenID Connect` triggered a JS error, which is addressed in 3efa4d836a300dc45b3ffece766b2b13539fc47c.
Manual testing:
- go to http://localhost:3000/admin/auths
- click on `Add authentication source`
- Choose `Authentication type`: `OAuth2`
- Choose `OAuth2 provider`: `Nextcloud`
- check that the `Use Custom URLs Instead of Default URLs` checkbox toggles the fields below
- let the checkbox be checked
- Switch the `OAuth2 provider` to `OpenID Connect`
- ensure that no JS error is shown
- Switch the `OAuth2 provider` to `Mastodon`
- check that the fields below `Use Custom URLs Instead of Default URLs` have the right defaults (mastodon.social)
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/4059
Reviewed-by: twenty-panda <twenty-panda@noreply.codeberg.org>
Reviewed-by: Earl Warren <earl-warren@noreply.codeberg.org>
Co-authored-by: oliverpool <git@olivier.pfad.fr>
Co-committed-by: oliverpool <git@olivier.pfad.fr>
Fix adopt repository has empty object name in database (#31333)
Fix #31330
Fix #31311
A workaround to fix the old database is to update object_format_name to
`sha1` if it's empty or null.
(cherry picked from commit 1968c2222dcf47ebd1697afb4e79a81e74702d31)
With tests services/repository/adopt_test.go
This adds org-mode readmes to the extensions that are prioritized when
searching for readme. Org-mode readmes come after markdown readmes in
terms of priority and before txt readmes.
Closes #4073
This removes the difference between high density images and other images regarding the pasting.
## Why
With this change, all images are clickable by default again. I don't think there is any problem regarding the img size because 1. it is the old behaviour, 2. the comment container already limits the size of the image.
## Alternatives
We can add an a-tag automatically when the user pastes an image. I do not prefer this because this adds a really long text (it's already bad with the img-tag) e.g.: `<a href="/attachments/28cf2254-13be-46c6-a433-efc77f556083" target="_blank"><img width="385" alt="grafik" src="/attachments/28cf2254-13be-46c6-a433-efc77f556083"></a>`
## Testing
1. Open an issue or pull request
2. Paste an image in the comment text box
3. The image should be pasted with valid Markdown syntax
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/3965
Reviewed-by: twenty-panda <twenty-panda@noreply.codeberg.org>
Reviewed-by: 0ko <0ko@noreply.codeberg.org>
Co-authored-by: Beowulf <beowulf@beocode.eu>
Co-committed-by: Beowulf <beowulf@beocode.eu>
More info in the linked PR.
---
Make positioning of the repo tabs make more sense. This is an isolated implementation for one of many changes discussed in the referenced issue, it will work good without the other changes too.
## Changes
- Actions are moved to the edge. This tab is the least relevant to both visitors and developers. The first don't really need it at all, the second only visit it when something goes unexpected (run did not happen or attached to the wrong event), or just to see the run queue to know when their actions is going to get processed. This is not a tab with always-relevant information.
- put Packages after releases. The Packages are like a download page for Releases, but for released packages instead of binaries/source code. It is relevant to Releases, so it should stay close, but it is secondary to Releases by importance. For example, because they don't actually contain release notes unlike Releases.
- the above makes Projects appear next to Issues and Pull requests which I think is nice as they're related.
## Preview
### v7
https://codeberg.org/attachments/c434e8fd-aaab-4c27-9071-2a3ba68ad4b7
### This PR
https://codeberg.org/attachments/74743c03-883e-40cf-8cb1-384d1d8cf63c
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/4139
Reviewed-by: Earl Warren <earl-warren@noreply.codeberg.org>
Reviewed-by: Otto <otto@codeberg.org>
Reviewed-by: Beowulf <beowulf@noreply.codeberg.org>
Add tag protection manage via rest API.
---------
Co-authored-by: Alexander Kogay <kogay.a@citilink.ru>
Co-authored-by: Giteabot <teabot@gitea.io>
(cherry picked from commit d4e4226c3cbfa62a6adf15f4466747468eb208c7)
Conflicts:
modules/structs/repo_tag.go
trivial context conflict
templates/swagger/v1_json.tmpl
fixed with make generate-swagger
Fix #31327
This is a quick patch to fix the bug.
Some parameters are using 0, some are using -1. I think it needs a
refactor to keep consistent. But that will be another PR.
(cherry picked from commit e4abaff7ffbbc5acd3aa668a9c458fbdf76f9573)
The PR replaces all `goldmark/util.BytesToReadOnlyString` with
`util.UnsafeBytesToString`, `goldmark/util.StringToReadOnlyBytes` with
`util.UnsafeStringToBytes`. This removes one `TODO`.
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
(cherry picked from commit 1761459ebc7eb6d432eced093b4583425a5c5d4b)
fix a bug from #30490
`prs.GetIssueIDs()` will also be used in other places, e.g.
`InvalidateCodeComments`
so we should not add `if pr.Issue == nil` in it, or if `pr.Issue` is
already loaded, you will not get the issueID in the results list and
this is not an expected result.
So this will caused a bug:
before calling `InvalidateCodeComments`, all `pr.Issues` in `prs` are
loaded, so `issueIDs` in this function will always be `[]`.
(cherry picked from commit e61e9a36b7117bab2cb122a95d606a86527ed45d)
have repo OrderBy definitions defined in one place and use a single type
for OrderBy database options
(cherry picked from commit bb04311b0b5b7a28f94c4bc409db1c4a04bcef17)
Fix a hash render problem like `<hash>: xxxxx` which is usually used in
release notes.
(cherry picked from commit 7115dce773e3021b3538ae360c4e7344d5bbf45b)
When using the MinIO storage driver for Actions Artifacts, we found that
the chunked artifact required significantly more memory usage to both
upload and merge than the local storage driver. This seems to be related
to hardcoding a value of `-1` for the size to the MinIO client [which
has a warning about memory usage in the respective
docs](https://pkg.go.dev/github.com/minio/minio-go/v7#Client.PutObject).
Specifying the size in both the upload and merge case reduces memory
usage of the MinIO client.
Co-authored-by: Kyle D <kdumontnu@gmail.com>
(cherry picked from commit 45dbeb5600d1f552c0134721fe49e8fd1099b5a4)
Fix #31330
Fix #31311
A workaround to fix the old database is to update object_format_name to
`sha1` if it's empty or null.
(cherry picked from commit 1968c2222dcf47ebd1697afb4e79a81e74702d31)
Enable [unparam](https://github.com/mvdan/unparam) linter.
Often I could not tell the intention why param is unused, so I put
`//nolint` for those cases like webhook request creation functions never
using `ctx`.
---------
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Co-authored-by: delvh <dev.lh@web.de>
(cherry picked from commit fc2d75f86d77b022ece848acf2581c14ef21d43b)
Conflicts:
modules/setting/config_env.go
modules/storage/azureblob.go
services/webhook/dingtalk.go
services/webhook/discord.go
services/webhook/feishu.go
services/webhook/matrix.go
services/webhook/msteams.go
services/webhook/packagist.go
services/webhook/slack.go
services/webhook/telegram.go
services/webhook/wechatwork.go
run make lint-go and fix Forgejo specific warnings
closes #3855
unlike #3854, this implementation uses a generic details html tag and a bit of tailwind magic...
---
## Maintainers Note
- previously tailwind classes of the form `[-a-zA-Z:0-9_.]` was disabled, however they were enabled since they were required for the `group-open:` classes
---
## Manual Testing
1. Visit the code search results after submitting a valid query for repo (if indexer disabled) or repo, user, explore (if indexer enabled)
2. Verify thst
1. the results are unfloded/open by default
2. the chevron points down when open and right when closed
<video src="/attachments/5a55c56f-6159-4422-ab80-962e0121e7d2" title="fold2.mp4" controls></video>
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/4134
Reviewed-by: Earl Warren <earl-warren@noreply.codeberg.org>
Reviewed-by: Beowulf <beowulf@noreply.codeberg.org>
Co-authored-by: Shiny Nematoda <snematoda.751k2@aleeas.com>
Co-committed-by: Shiny Nematoda <snematoda.751k2@aleeas.com>
This overrides handling of Tab and Enter keys in the ComboMarkdownEditor, so that:
* Pressing Tab prepending 4 spaces to the line under cursor, or all the lines in the selection.
* Pressing Shift+Tab removes up to 4 spaces.
* Pressing Enter repeats any indentation and a "repeatable" prefix (list or blockquote) from the current line.
Since Tab "capture" can interfere with keyboard navigation, it's only done if there was any previous input in the textarea or if it was selected with a pointer. Additionally, presing Esc will make the textarea lose focus, resuming tab navigation. This seems adequate to me, but I might be wrong.
Had to use the "deprecated" execCommand method, since anything else I tried messes up the undo history. There's a fallback for when (if?) it's actually removed.
Only tested in desktop Firefox and Chrome so far.
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/4072
Reviewed-by: Earl Warren <earl-warren@noreply.codeberg.org>
Reviewed-by: Beowulf <beowulf@noreply.codeberg.org>
Co-authored-by: Danko Aleksejevs <danko@very.lv>
Co-committed-by: Danko Aleksejevs <danko@very.lv>
use proper http time format than replacing with GMT in time.RFC1123 =)
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/4132
Reviewed-by: Earl Warren <earl-warren@noreply.codeberg.org>
Co-authored-by: Shiny Nematoda <snematoda.751k2@aleeas.com>
Co-committed-by: Shiny Nematoda <snematoda.751k2@aleeas.com>
When a comment was added and saved with the preview tab active, the
preview tab is still active, when the comment is edited again.
This adds a "hacky" solution, but it works and is simple.
Every time the edit is "started" and the editor already exists, the tab
with the edit text field gets clicked to activate it.
Fixes #1334
