This simplifies the interface to just expose the methods from the underlying connection that we need. The addition of gen.go makes creating the generated mocks easy go generate.
This fixes a usage of uuid.New() which can potentially panic. Instead we use a uuid.NewRandom() which also generates a UUID V4 instead of a UUID V1. In addition all uuid.NewUUID() calls have been replaced by uuid.NewRandom().
* feat(oidc): oauth2 discovery and endpoint rename
This implements the oauth2 authorization server discovery document, adds tests to the discovery documents, implements an efficiency upgrade to these docs, and renames some endpoints to be uniform.
This implements Webauthn. Old devices can be used to authenticate via the appid compatibility layer which should be automatic. New devices will be registered via Webauthn, and devices which do not support FIDO2 will no longer be able to be registered. At this time it does not fully support multiple devices (backend does, frontend doesn't allow registration of additional devices). Does not support passwordless.
This enhances the existing time.Duration parser to allow multiple units, and implements a decode hook which can be used by koanf to decode string/integers into time.Durations as applicable.
Implements Proof Key for Code Exchange for OpenID Connect Authorization Code Flow. By default this is enabled for the public client type and requires the S256 challenge method.
Closes#2921
This adds an access-control command that checks the policy enforcement for a given criteria using a configuration file and refactors the configuration validation command to include all configuration sources.
This unifies the methods to obtain the X-Forwarded-* header values and provides logical fallbacks. In addition, so we can ensure this functionality extends to the templated files we've converted the ServeTemplatedFile method into a function that operates as a middlewares.RequestHandler.
Fixes#2765
This adds a smart delay on reset password attempts to prevent username enumeration. Additionally utilizes crypto rand instead of math rand. It also moves the timing delay functionality into its own handler func.
Adds encryption to the U2F public keys. While the public keys cannot be used to authenticate, only to validate someone is authenticated, if a rogue operator changed these in the database they may be able to bypass 2FA. This prevents that.
* change all instances (file names and docs) of "config.template.yml" to "configuration.template.yml" so its consistent with the expectations of the Dockerfile
* Keep config.template.yml named as is
Co-authored-by: Amir Zarrinkafsh <nightah@me.com>
* Update index.html
* revert filename changes and add a note about docker
* refactor: apply suggestions from code review
Co-authored-by: Amir Zarrinkafsh <nightah@me.com>
This replaces the standard duo_devices upsert with a PostgreSQL specific one and ensures the u2f_devices upsert uses the new unique key for the ON CONFLICT check.
This utilizes the referrer query parameters instead of current request query parameters for logging the requested URI and method. Minor performance improvements to header peek/sets.
This is a fix to the authentication_logs schema that prevents the VARCHAR from being insufficient for HTTP request methods such as PATCH, DELETE, OPTIONS, CONNECT.
Allow users to configure the TOTP Algorithm and Digits. This should be used with caution as many TOTP applications do not support it. Some will also fail to notify the user that there is an issue. i.e. if the algorithm in the QR code is sha512, they continue to generate one time passwords with sha1. In addition this drastically refactors TOTP in general to be more user friendly by not forcing them to register a new device if the administrator changes the period (or algorithm).
Fixes#1226.
Allow users to select and save the preferred duo device and method, depending on availability in the duo account. A default enrollment URL is provided and adjusted if returned by the duo API. This allows auto-enrollment if enabled by the administrator.
Closes#594. Closes#1039.
This change addresses an issue with the usage of the full sender configuration option in the MAIL FROM SMTP command. If a user includes a name in the sender this shouldn't be sent in the MAIL FROM command, instead we should extract it and use just the email portion.
Fixes#2571
This change makes it so only metadata about tokens is stored. Tokens can still be resigned due to conversion methods that convert from the JWT type to the database type. This should be more efficient and should mean we don't have to encrypt tokens or token info in the database at least for now.
This adds additional logging to the authentication logs such as type, remote IP, request method, redirect URL, and if the attempt was done during a ban. This also means we log attempts that occur when the attempt was blocked by the regulator for record keeping purposes, as well as record 2FA attempts which can be used to inform admins and later to regulate based on other factors.
Fixes#116, Fixes#1293.
This adds an AES-GCM 256bit encryption layer for storage for sensitive items. This is only TOTP secrets for the time being but this may be expanded later. This will require a configuration change as per https://www.authelia.com/docs/configuration/migration.html#4330.
Closes#682
This is a massive overhaul to the SQL Storage for Authelia. It facilitates a whole heap of utility commands to help manage the database, primary keys, ensures all database requests use a context for cancellations, and paves the way for a few other PR's which improve the database.
Fixes#1337
* feat: customizable static assets
This change provides the means to override specific assets from the embedded Go FS with files situated on disk.
We only allow overriding the following files currently:
* favicon.ico
* logo.png
* refactor(server): make logo string a const
* refactor(suites): override favicon and use ntp3 in traefik2 suite
* test(suites): test logo override in traefik2 suite
* test(suites): test asset override fallback in traefik suite
Closes#1630.
This adjusts the session ID generator making it use it's own random function rather than using one from the utils lib. This allows us to utilize crypto/rand or math/rand interchangeably. Additionally refactor the utils.RandomString func.
* refactor(suites): replace selenium with go-rod
This change replaces [tebeka/selenium](https://github.com/tebeka/selenium) with [go-rod](https://github.com/go-rod/rod).
We no longer have a chromedriver/external driver dependency to utilise Selenium as we instead utilise the Chrome Dev Protocol to communicate with the browser.
Rod [documents](https://go-rod.github.io/#/why-rod) benefits of choosing the library as opposed to the available alternatives.
This changes the validation of salts for sha512 to be done by the upstream API rather than locally. This allows the salts used in Linux to be utilized with Authelia provided the hash is a sha512 hash.
Co-authored-by: Amir Zarrinkafsh <nightah@me.com>
* feat(web): replace cra with vite
* fix: add istanbul
* fix: add jest
* fix: inject env vars
* fix: replicate cra output directories
* fix: post-frontend build for go templating
* fix: dynamic publicpath
* fix(web): import resolution with aliases for .module.css files
* refactor(server): baseurl var
* refactor(web): drop babel-jest for esbuild-jest
* refactor(web): add inline sourcemap for coverage bundle
* build(deps): update web deps
* build(deps): downgrade vite-plugin-istanbul to 2.2.0
98bf77dbaa is a breaking change that means production mode builds can no longer be instrumented.
* refactor(web): match frontend name and version
* refactor(web): drop cra readme
* ci(codecov): utilise new codecov uploader for coverage
The codecov bash uploader is being [deprecated](https://docs.codecov.com/docs/about-the-codecov-bash-uploader).
This utilises the new uploader which is recommended.
* ci(codecov): adjust file search path and name uploads
* fix(suites): coverage paths for codecov
* refactor: replace sqlite library with the original cgo variant
This change reverts our SQLite library back to the original for compatibility and performance reasons now that we always package with CGO.
* fix: cgo and build flags
* fix: gcc requirement
* fix: gcc in dev workflow too