[][github-actions] [][github-actions-tests] [][dependabot-pulls] [][github-releases] [][precommit-config] [][project-license] [][gomod-file]
Download binary from releases or build with go build .
.
If you have api key, it free and will suggest to create a free account, create a file .go-wallhaven
at home dir
{
"API_KEY" : "<API_KEY>",
"DIR" : "<DIRECTORY YOU WANTED TO BE SAVE>"
}
This section is intended to help developers and contributors get a working copy of
go-wallhaven
on their end
1. Clone this repository
git clone https://github.com/coderj001/go-wallhaven
cd go-wallhaven
2. Install `golangci-lint`
Install golangci-lint
from the [official website][golangci-install] for your OS
This section will guide you to setup a fully-functional local copy of go-wallhaven
on your end and teach you how to use it! Make sure you have installed
[golangci-lint][golangci-install] before following this section!
Note: This section relies on the usage of [Makefile][makefile-official]. If you
can't (or don't) use Makefile, you can follow along by running the internal commands
from [go-wallhaven's
Makefile][makefile-file] (most of which are
OS-independent)!
To install all dependencies associated with go-wallhaven
, run the
command
make install
Code formatters format your code to match pre-decided conventions. To run automated code formatters, use the Makefile command
make codestyle
Linters are tools that analyze source code for possible errors. This includes typos, code formatting, syntax errors, calls to deprecated functions, potential security vulnerabilities, and more!
To run pre-configured linters, use the command
make lint
Tests in go-wallhaven
are classified as fast and slow - depending
on how quick they are to execute.
To selectively run tests from either test group, use the Makefile command
make fast-test
OR
make slow-test
Alternatively, to run the complete test-suite -- i.e. fast and slow tests at one go, use the command
make test
The test-suite
is simply a wrapper to run linters, stylecheckers and all tests
at once!
To run the test-suite, use the command
make test-suite
In simpler terms, running the test-suite is a combination of running linters
and all tests one after the other!
Tap for a list of Makefile commands
Command | Description | Prerequisites |
---|---|---|
help |
Generate help dialog listing all Makefile commands with description | NA |
install |
Fetch project dependencies | NA |
codestyle |
Run code-formatters | golangci-lint |
lint |
Check codestyle and run linters | golangci-lint |
test |
Run all tests | NA |
fast-tests |
Selectively run fast tests | NA |
slow-tests |
Selectively run slow tests | NA |
test-suite |
Check codestyle, run linters and all tests | golangci-lint |
run |
Run go-wallhaven | NA |
docker-gen |
Create production docker image for go-wallhaven | docker |
docker-debug |
Create debug-friendly docker image for go-wallhaven | docker |
clean-docker |
Remove docker image generated by docker-gen |
docker |
Optionally, to see a list of all Makefile commands, and a short description of what they do, you can simply run
make
Which is equivalent to;
make help
Both of which will list out all Makefile commands available, and a short description of what they do!
To generate binaries for multiple OS/architectures, simply run
bash build-script.sh
The command will generate binaries for Linux, Windows and Mac targetting multiple
architectures at once! The binaries, once generated will be stored in the bin
directory inside the project directory.
The binaries generated will be named in the format
go-wallhaven_<version>_<target-os>_<architecture>.<extension>
The <extension>
is optional. By default, version
is an empty string. A custom
version can be passed as an argument while running the script. As an example;
bash build-script.sh v1.2.1
An example of the files generated by the previous command will be;
go-wallhaven_v1.2.1_windows_x86_64.exe
Follow [Codecov Docs][codecov-docs] to activate Codecov for your project repository.
Once you've activated Codecov for your project, you should be able to see the
Repository Upload Token
. Copy this token, and add it as a secret to your Github
repository. Checkout [Creating secrets for a Repository][creating-secrets] for info
on how to add secrets on Github.
For the secret, the name of the secret should be "CODECOV_TOKEN
" (no spaces,
copy-paste the string as it is). The value of the secret would be the
Repository Upload Token
obtained from Codecov.
Save the secret, you should be able to a secret named CODECOV_TOKEN
in the
Settings > Secrets section of your project repository. If this field is visible,
you are done with setting up Codecov, and should be able to see code coverage reports
after the next run of CI pipeline!
To run go-wallhaven
in a docker container, read the instructions in
docker section.
To run go-wallhaven, use the command
make run
Additionally, you can pass any additional command-line arguments (if needed) as the
argument "q
". For example;
make run q="--help"
OR
make run q="--version"
You can check out a list of previous releases on the [Github Releases][github-releases] page.
What is Semantic Versioning?
Semantic versioning is a versioning scheme aimed at making software management easier. Following semantic versioning, version identifiers are divided into three parts;
<major>.<minor>.<patch>
MAJOR version when you make incompatible API changes [breaking changes]
MINOR version when you add functionality in a backwards compatible manner [more features]
PATCH version when you make backwards compatible bug fixes [bug fixes and stuff]
For a more detailed description, head over to [semver.org][semver-link]
[Release Drafter][release-drafter] automatically updates the release version as pull requests are merged.
Labels allowed;
major
: Affects the<major>
version number for semantic versioningminor
,patch
: Affects the<patch>
version number for semantic versioning
Whenever a pull request with one of these labels is merged to the master
branch,
the corresponding version number will be bumped by one digit!
Pull requests once merged, will be classified into categories by [release-drafter][release-drafter] based on pull request labels
This is managed by the [release-drafter.yml
][release-drafter-config] config file.
Label | Title in Releases |
---|---|
security |
π Security |
enhancement , feature , update |
π Updates |
bug , bugfix , fix |
π Bug Fixes |
documentation , docs |
π Documentation |
wip , in-progress , incomplete , partial , hotfix |
π§ Work in Progress |
dependencies , dependency |
π¦ Dependencies |
refactoring , refactor , tests , testing |
π§ͺ Tests and Refactor |
build , ci , pipeline |
π€ CI/CD and Pipelines |
The labels bug
, enhancement
, and documentation
are automatically created by Github
for repositories. [Dependabot][dependabot-link] will implicitly create the
dependencies
label with the first pull request raised by it.
The remaining labels can be created as needed!
- Limit the number of go worker should be user input
- Page input should be a range of pages (need to rewrite apifech function & others)
- Resolution shlould be fzf style multi-select (will try to apply on other fields) : :