While BREACH requires the following three factors to be effectively exploitable we should add another mitigation:
1. Application must support HTTP compression
2. Response most reflect user-controlled input
3. Response should contain sensitive data
Especially part 2 is with ownCloud not really given since user-input is usually only echoed if a CSRF token has been passed.
To reduce the risk even further it is however sensible to encrypt the CSRF token with a shared secret. Since this will change on every request an attack such as BREACH is not feasible anymore against the CSRF token at least.
This changeset removes the static class `OC_Request` and moves the functions either into `IRequest` which is accessible via `\OC::$server::->getRequest()` or into a separated `TrustedDomainHelper` class for some helper methods which should not be publicly exposed.
This changes only internal methods and nothing on the public API. Some public functions in `util.php` have been deprecated though in favour of the new non-static functions.
Unfortunately some part of this code uses things like `__DIR__` and thus is not completely unit-testable. Where tests where possible they ahve been added though.
Fixes https://github.com/owncloud/core/issues/13976 which was requested in https://github.com/owncloud/core/pull/13973#issuecomment-73492969
When `mod_unique_id` is enabled the ID generated by it will be used for logging. This allows for correlation of the Apache logs and the ownCloud logs.
Testplan:
- [ ] When `mod_unique_id` is enabled the request ID equals the one generated by `mod_unique_id`.
- [ ] When `mod_unique_id` is not available the request ID is a 20 character long random string
- [ ] The generated Id is stable over the lifespan of one request
Changeset looks a little bit larger since I had to adjust every unit test using the HTTP\Request class for proper DI.
Fixes https://github.com/owncloud/core/issues/13366
Thomas Müller
Lukas Reschke
Joas Schilling
Bernhard Posselt