While reviewing cpuprofiles of #10985, I noticed a bunch of slice-grows within this function.
Instead of `io.ReadAll` (which does not know the size in advance), use `fs.ReadFile`, which will perform correctly-sized allocations.
### Tests
This function is already covered by tests in `modules/assetfs/layered_test.go`
### Documentation
- [ ] I created a pull request [to the documentation](https://codeberg.org/forgejo/docs) to explain to Forgejo users how to use this change.
- [x] I did not document these changes and I do not expect someone else to do it.
### Release notes
- [ ] This change will be noticed by a Forgejo user or admin (feature, bug fix, performance, etc.). I suggest to include a release note for this change.
- [x] This change is not visible to a Forgejo user or admin (refactor, dependency upgrade, etc.). I think there is no need to add a release note for this change.
*The decision if the pull request will be shown in the release notes is up to the mergers / release team.*
The content of the `release-notes/<pull request number>.md` file will serve as the basis for the release notes. If the file does not exist, the title of the pull request will be used instead.
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/10987
Reviewed-by: Michael Kriese <michael.kriese@gmx.de>
Reviewed-by: Mathieu Fenniak <mfenniak@noreply.codeberg.org>
Co-authored-by: oliverpool <git@olivier.pfad.fr>
Co-committed-by: oliverpool <git@olivier.pfad.fr>
- This reverts behavior that was partially unintentionally introduced in forgejo/forgejo#8143, symbolic links were no longer followed (if they escaped the asset folder) for local assets.
- Having symbolic links for user-added files is, to my understanding, a ,common usecase for NixOS and would thus have symbolic links in the asset folders. Avoiding symbolic links is not easy.
- The previous code used `http.Dir`, we cannot use that as it's not of the same type. The equivalent is `os.DirFS`.
- Unit test to prevent this regression from happening again.
Reported-by: bloxx12 (Matrix).
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/8596
Reviewed-by: Earl Warren <earl-warren@noreply.codeberg.org>
Co-authored-by: Gusted <postmaster@gusted.xyz>
Co-committed-by: Gusted <postmaster@gusted.xyz>
- Replace the current vfsgen with our own bindata generator.
- zstd is used instead of gzip. This reduces the size of the resulting binary by 2MiB, the size of the bundled assets were thus reduced from 13MiB to 11MiB.
- If [the browser accepts zstd encoding](https://caniuse.com/zstd), then the compressed bytes can be served directly, otherwise it falls back to being compressed by gzip if it's not disabled via `[server].ENABLE_GZIP`
- The compression and decompression speed is roughly 4 times faster.
- The generated filesystem is now of type `fs.Fs` instead of `http.FileSystem`, this slightly simplifies the generated code and
handling of the assets.
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/8143
Co-authored-by: Gusted <postmaster@gusted.xyz>
Co-committed-by: Gusted <postmaster@gusted.xyz>
- Add a new `Seq` function to the `Set` type, this returns an iterator over the values.
- Convert some users of the `Values` method to allow for more optimal code.
Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/6676
Reviewed-by: Earl Warren <earl-warren@noreply.codeberg.org>
Co-authored-by: Gusted <postmaster@gusted.xyz>
Co-committed-by: Gusted <postmaster@gusted.xyz>
Before, in dev mode, there might be some error logs like:
```
2023/07/17 13:54:51 ...s/assetfs/layered.go:221:WatchLocalChanges() [E] Unable to watch directory .: lstat /data/work/gitea/custom/templates: no such file or directory
```
Because there is no "custom/templates" directory.
After: ignore such error, no such error message anymore.
Replace #25580Fix#19453
The problem was: when users set "GITEA__XXX__YYY" , the "install page"
doesn't respect it.
So, to make the result consistent and avoid surprising end users, now
the "install page" also writes the environment variables to the config
file.
And, to make things clear, there are enough messages on the UI to tell
users what will happen.
There are some necessary/related changes to `environment-to-ini.go`:
* The "--clear" flag is removed and it was incorrectly written there.
The "clear" operation should be done if INSTALL_LOCK=true
* The "--prefix" flag is removed because it's never used, never
documented and it only causes inconsistent behavior.
The idea is to use a Layered Asset File-system (modules/assetfs/layered.go)
For example: when there are 2 layers: "custom", "builtin", when access
to asset "my/page.tmpl", the Layered Asset File-system will first try to
use "custom" assets, if not found, then use "builtin" assets.
This approach will hugely simplify a lot of code, make them testable.
Other changes:
* Simplify the AssetsHandlerFunc code
* Simplify the `gitea embedded` sub-command code
---------
Co-authored-by: Jason Song <i@wolfogre.com>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
