- Use maintained fork https://github.com/golangci/misspell - Rename `mispell-check` to `lint-spell`, add `lint-spell-fix` - Run `lint-spell` in separate actions step - Lint more files, fix discovered issues - Remove inaccurate and outdated info in docs (we do not need GOPATH for tools anymore) Maybe later we can add more spellchecking tools, but I have not found any good ones yet.
13 KiB
date | title | slug | sidebar_position | toc | draft | aliases | menu | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2016-12-01T16:00:00+02:00 | Hacking on Gitea | hacking-on-gitea | 10 | false | false |
|
|
Hacking on Gitea
Quickstart
To get a quick working development environment you could use Gitpod.
Installing go
You should install go and set up your go
environment correctly.
Next, install Node.js with npm which is
required to build the JavaScript and CSS files. The minimum supported Node.js
version is @minNodeVersion@ and the latest LTS version is recommended.
Note: When executing make tasks that require external tools, like
make watch-backend
, Gitea will automatically download and build these as
necessary. To be able to use these you must have the "$GOPATH"/bin
directory
on the executable path. If you don't add the go bin directory to the
executable path you will have to manage this yourself.
Note 2: Go version @minGoVersion@ or higher is required.
Gitea uses gofmt
to format source code. However, the results of
gofmt
can differ by the version of go
. Therefore it is
recommended to install the version of Go that our continuous integration is
running. As of last update, the Go version should be @goVersion@.
To lint the template files, ensure Python and
Poetry are installed.
Installing Make
Gitea makes heavy use of Make to automate tasks and improve development. This
guide covers how to install Make.
On Linux
Install with the package manager.
On Ubuntu/Debian:
sudo apt-get install make
On Fedora/RHEL/CentOS:
sudo yum install make
On Windows
One of these three distributions of Make will run on Windows:
- Single binary build. Copy somewhere and add to
PATH
. - MinGW-w64 / MSYS2.
- MSYS2 is a collection of tools and libraries providing you with an easy-to-use environment for building, installing and running native Windows software, it includes MinGW-w64.
- In MingGW-w64, the binary is called
mingw32-make.exe
instead ofmake.exe
. Add thebin
folder toPATH
. - In MSYS2, you can use
make
directly. See MSYS2 Porting. - To compile Gitea with CGO_ENABLED (eg: SQLite3), you might need to use tdm-gcc instead of MSYS2 gcc, because MSYS2 gcc headers lack some Windows-only CRT functions like
_beginthread
.
- Chocolatey package. Run
choco install make
Note: If you are attempting to build using make with Windows Command Prompt, you may run into issues. The above prompts (Git bash, or MinGW) are recommended, however if you only have command prompt (or potentially PowerShell) you can set environment variables using the set command, e.g. set TAGS=bindata
.
Downloading and cloning the Gitea source code
The recommended method of obtaining the source code is by using git clone
.
git clone https://github.com/go-gitea/gitea
(Since the advent of go modules, it is no longer necessary to build go projects
from within the $GOPATH
, hence the go get
approach is no longer recommended.)
Forking Gitea
Download the main Gitea source code as above. Then, fork the
Gitea repository on GitHub,
and either switch the git remote origin for your fork or add your fork as another remote:
# Rename original Gitea origin to upstream
git remote rename origin upstream
git remote add origin "git@github.com:$GITHUB_USERNAME/gitea.git"
git fetch --all --prune
or:
# Add new remote for our fork
git remote add "$FORK_NAME" "git@github.com:$GITHUB_USERNAME/gitea.git"
git fetch --all --prune
To be able to create pull requests, the forked repository should be added as a remote
to the Gitea sources. Otherwise, changes can't be pushed.
Building Gitea (Basic)
Take a look at our
instructions
for building from source.
The simplest recommended way to build from source is:
TAGS="bindata sqlite sqlite_unlock_notify" make build
The build
target will execute both frontend
and backend
sub-targets. If the bindata
tag is present, the frontend files will be compiled into the binary. It is recommended to leave out the tag when doing frontend development so that changes will be reflected.
See make help
for all available make
targets. Also see .drone.yml
to see how our continuous integration works.
Building continuously
To run and continuously rebuild when source files change:
# for both frontend and backend
make watch
# or: watch frontend files (html/js/css) only
make watch-frontend
# or: watch backend files (go) only
make watch-backend
On macOS, watching all backend source files may hit the default open files limit which can be increased via ulimit -n 12288
for the current shell or in your shell startup file for all future shells.
Formatting, code analysis and spell check
Our continuous integration will reject PRs that fail the code linters (including format check, code analysis and spell check).
You should format your code:
make fmt
and lint the source code:
# lint both frontend and backend code
make lint
# lint only backend code
make lint-backend
Note: The results of gofmt
are dependent on the version of go
present.
You should run the same version of go that is on the continuous integration
server as mentioned above.
Working on JS and CSS
Frontend development should follow Guidelines for Frontend Development
To build with frontend resources, either use the watch-frontend
target mentioned above or just build once:
make build && ./gitea
Before committing, make sure the linters pass:
make lint-frontend
Configuring local ElasticSearch instance
Start local ElasticSearch instance using docker:
mkdir -p $(pwd)/data/elasticsearch
sudo chown -R 1000:1000 $(pwd)/data/elasticsearch
docker run --rm --memory="4g" -p 127.0.0.1:9200:9200 -p 127.0.0.1:9300:9300 -e "discovery.type=single-node" -v "$(pwd)/data/elasticsearch:/usr/share/elasticsearch/data" docker.elastic.co/elasticsearch/elasticsearch:7.16.3
Configure app.ini
:
[indexer]
ISSUE_INDEXER_TYPE = elasticsearch
ISSUE_INDEXER_CONN_STR = http://elastic:changeme@localhost:9200
REPO_INDEXER_ENABLED = true
REPO_INDEXER_TYPE = elasticsearch
REPO_INDEXER_CONN_STR = http://elastic:changeme@localhost:9200
Building and adding SVGs
SVG icons are built using the make svg
target which compiles the icon sources defined in build/generate-svg.js
into the output directory public/assets/img/svg
. Custom icons can be added in the web_src/svg
directory.
Building the Logo
The PNG and SVG versions of the Gitea logo are built from a single SVG source file assets/logo.svg
using the TAGS="gitea" make generate-images
target. To run it, Node.js and npm must be available.
The same process can also be used to generate custom logo PNGs from a SVG source file by updating assets/logo.svg
and running make generate-images
. Omitting the gitea
tag will update only the user-designated logo files.
Updating the API
When creating new API routes or modifying existing API routes, you MUST
update and/or create Swagger
documentation for these using go-swagger comments.
The structure of these comments is described in the specification.
If you want more information about the Swagger structure, you can look at the
Swagger 2.0 Documentation
or compare with a previous PR adding a new API endpoint, e.g. PR #5483
You should be careful not to break the API for downstream users which depend
on a stable API. In general, this means additions are acceptable, but deletions
or fundamental changes to the API will be rejected.
Once you have created or changed an API endpoint, please regenerate the Swagger
documentation using:
make generate-swagger
You should validate your generated Swagger file:
make swagger-validate
You should commit the changed swagger JSON file. The continuous integration
server will check that this has been done using:
make swagger-check
Note: Please note you should use the Swagger 2.0 documentation, not the
OpenAPI 3 documentation.
Creating new configuration options
When creating new configuration options, it is not enough to add them to the
modules/setting
files. You should add information to custom/conf/app.ini
and to the
configuration cheat sheet
found in docs/content/doc/administer/config-cheat-sheet.en-us.md
Changing the logo
When changing the Gitea logo SVG, you will need to run and commit the results
of:
make generate-images
This will create the necessary Gitea favicon and others.
Database Migrations
If you make breaking changes to any of the database persisted structs in the
models/
directory, you will need to make a new migration. These can be found
in models/migrations/
. You can ensure that your migrations work for the main
database types using:
make test-sqlite-migration # with SQLite switched for the appropriate database
Testing
There are two types of test run by Gitea: Unit tests and Integration Tests.
Unit Tests
Unit tests are covered by *_test.go
in go test
system.
You can set the environment variable GITEA_UNIT_TESTS_LOG_SQL=1
to display all SQL statements when running the tests in verbose mode (i.e. when GOTESTFLAGS=-v
is set).
TAGS="bindata sqlite sqlite_unlock_notify" make test # Runs the unit tests
Integration Tests
Unit tests will not and cannot completely test Gitea alone. Therefore, we
have written integration tests; however, these are database dependent.
TAGS="bindata sqlite sqlite_unlock_notify" make build test-sqlite
will run the integration tests in an SQLite environment. Integration tests
require git lfs
to be installed. Other database tests are available but
may need adjustment to the local environment.
Take a look at tests/integration/README.md
for more information and how to run a single test.
Testing for a PR
Our continuous integration will test the code passes its unit tests and that
all supported databases will pass integration test in a Docker environment.
Migration from several recent versions of Gitea will also be tested.
Please submit your PR with additional tests and integration tests as
appropriate.
Documentation for the website
Documentation for the website is found in docs/
. If you change this you
can test your changes to ensure that they pass continuous integration using:
# from the docs directory within Gitea
make trans-copy clean build
You will require a copy of Hugo to run this task. Please
note: this may generate a number of untracked Git objects, which will need to
be cleaned up.
Visual Studio Code
A launch.json
and tasks.json
are provided within contrib/ide/vscode
for
Visual Studio Code. Look at
contrib/ide/README.md
for more information.
GoLand
Clicking the Run Application
arrow on the function func main()
in /main.go
can quickly start a debuggable Gitea instance.
The Output Directory
in Run/Debug Configuration
MUST be set to the
gitea project directory (which contains main.go
and go.mod
),
otherwise, the started instance's working directory is a GoLand's temporary directory
and prevents Gitea from loading dynamic resources (eg: templates) in a development environment.
To run unit tests with SQLite in GoLand, set -tags sqlite,sqlite_unlock_notify
in Go tool arguments
of Run/Debug Configuration
.
Submitting PRs
Once you're happy with your changes, push them up and open a pull request. It
is recommended that you allow Gitea Managers and Owners to modify your PR
branches as we will need to update it to main before merging and/or may be
able to help fix issues directly.
Any PR requires two approvals from the Gitea maintainers and needs to pass the
continuous integration. Take a look at our
CONTRIBUTING.md
document.
If you need more help pop on to Discord #Develop
and chat there.
That's it! You are ready to hack on Gitea.