* Removing some other fake apps
* More FakeApp removed
* Removing entirely FakeApp
* Fixing some tests
* Fixing get Cluster id from get plugin status
* Fixing failing tests
* Fixing tests
* Fixing test initialization for web
* Fixing InitServer for server tests
* Fixing InitServer for server tests
* Reverting go.sum and go.mod
* Removing unneded HTMLTemplates function in App layer
* Moving back some functions to its old place to easy the review
* Moving back some functions to its old place to easy the review
* Using the last struct2interface version
* Generating store layers
* Fixing merge problems
* Addressing PR comments
* Small fix
* Fixing app tests build
* Fixing tests
* fixing tests
* Fix tests
* Fixing tests
* Fixing tests
* Fixing tests
* Moving license to server struct
* Adding some fixes to the test compilation
* Fixing cluster and some jobs initialization
* Fixing some license tests compilation problems
* Fixing recursive cache invalidation
* Regenerating app layers
* Fix test compilation
Co-authored-by: mattermod <mattermod@users.noreply.github.com>
* [MM-24146] Add unix socket listener for mmctl local mode (#14296)
* add unix socket listener for mmctl local mode
* add a constant for local-mode socket path
* reflect review comments
* [MM-24401] Base approach for Local Mode (#14333)
* add unix socket listener for mmctl local mode
* First working PoC
* Adds the channel list endpoint
* Add team list endpoint
* Add a LocalClient to the api test helper and start local mode
* Add helper to test with both SystemAdmin and Local clients
* Add some docs
* Adds TestForAllClients test helper
* Incorporating @ashishbhate's proposal for adding test names to the helpers
* Fix init errors after merge
* Adds create channel tests
* Always init local mode to allow for enabling-disabling it via config
* Check the RemoteAddr of the request before marking session as local
* Mark the request as errored if it's local and the origin is remote
* Set the socket permissions to read/write when initialising
* Fix linter
* Replace RemoteAddr check to ditch connections with the IP:PORT shape
Co-authored-by: Ibrahim Serdar Acikgoz <serdaracikgoz86@gmail.com>
* Fix translations order
* [MM-24832] Migrate plugin endpoints to local mode (#14543)
* [MM-24832] Migrate plugin endpoints to local mode
* Fix client reference in helper
* [MM-24776] Migrate config endpoints to local mode (#14544)
* [MM-24776] Migrate get config endpoint to local mode
* [MM-24777] Migrate update config endpoint to local mode
* Fix update config to bypass RestrictSystemAdmin flag
* Add patchConfig endpoint
* MM-24774/MM-24755: local mode for addLicense and removeLicense (#14491)
Automatic Merge
Co-authored-by: Ibrahim Serdar Acikgoz <serdaracikgoz86@gmail.com>
Co-authored-by: Ashish Bhate <bhate.ashish@gmail.com>
* MM-25005: Remove HttpRequestsDuration bucket
The ApiTimesHistograms already captures what HttpRequestsDuration does.
And it's more granular than that. So there's no need to duplicate stuff.
Let's remove this one.
* Updating resource class to xlarge
Co-authored-by: mattermod <mattermod@users.noreply.github.com>
Co-authored-by: Jason Deland <jaydeland@gmail.com>
* ResponseWriter wrapper to get status code
For our metrics, we need the status code returned by a request
so this wrapper includes a new method StatusCode() that includes
the desired code
* Shadow the responsewriter variable in the handlers
In order to avoid confusion to people deciding what variable to use.
I've also changed the tests to reflect this change and added a new
one that checks the Flush method works
* MM-22057: Limit incoming request bodies
Set the max request body size to be equal to MaxFileSize.
Ideally, non-file request bodies should be smaller than file request bodies,
but we don't have a clean way to identify all file upload handlers.
There shouldn't be any valid request which exceeds the max file upload size.
So this is a safe global limit to apply.
* Fix tests
Co-authored-by: mattermod <mattermod@users.noreply.github.com>
* Consistent license message for all the go files
* Fixing the last set of unconsistencies with the license headers
* Addressing PR review comments
* Fixing busy.go and busy_test.go license header
* MM-8607: add ability to turn off non-critical services under load
* server busy invalid param unit tests
* MM-8607: rename server busy endpoints
* MM-8607: handle case where App not initialized
* MM-8607: additional unit test cases per feedback.
* MM-8607: use decorator to check isbusy when adding endpoint route
* MM-8607: rename endpoints, use struct for json
* Update api4/system.go
Fix misspelled log output
Co-Authored-By: Saturnino Abril <saturnino.abril@gmail.com>
* MM-8607: fix i18n order; max seconds for server busy expiry
* Add config setting to explicitly define which IP headers are trusted
* fix variable shadowing
* Optimize code flow; Add Ratelimit test for header set
* Extend Ratelimit tests
* Add additional unit tests
* Structured logging
* MM-14194: fix subpath csp directive until server restart
The SiteURL organically doesn't take effect until server restart, but in v5.8, the required CSP directive would change immediately. If changing from one subpath to another, the webapp would effectively be bricked until a server restart.
Avoid this by determining the CSP directive when the static handler is created.
* simplify access to config
* MM-14052: fix subpath yet again
The server now emits a script-src directive that overrides the root.html
rewrite. Fix this by emitting the requisite sha-256 hash server-side as
well as rewriting root.html. We can't remove the root.html rewrite,
since the assets may be on a CDN instead and we use the same code path
to rewrite them (on demand).
Prior to this change, going from / -> /subpath -> / would leave changes
in root.html: the Content-Security-Policy header would still have the
sha-256 hash, and the inline script would still override the publicPath
but to the default subpath value. To avoid sending down a sha-256 hash
server-side when no subpath is required, change this to fully strip out
the subpath changes. This is the only unit test change, as the existing
coverage proves the algorithm still works.
* fix subpath concatenation in test
path.Join isn't meant to work with a URL + path, and my test was effectively working with the subpath "/localhost:8065/subpath" instead of just "/subpath". The CI servers presumably caught this due to a different configuration than my local development.
* CSRF Token Implementation + Tests
Remove debug statements
Implement requested changes
* Fix non-cookie authentication methods stripping auth data from requests
* Fail when CSRF cookie is not returned as part of login
Config Checks at StartUp Part1
Config Checks; Tests for TLS Server
HSTS header implementation + tests
make gofmt happy with new go version...
make gofmt happy with new go version #2...
fix logic bug
fix typo
Fix unnecessary code block
* fix context's SiteURLHeader to use the subpath
This fixes various commands such as /open and /groupmsg, as well as
OAuth related redirects.
* MM-11246: also leverage GetSiteHeader() in completeSaml
* factor out GetSubpathFromConfig
* mv web/subpath.go to utils/subpath.go
* serve up web, api and ws on /subpath if configured
* pass config to utils.RenderWeb(App)?Error
This allows the methods to extract the configured subpath and redirect
to the appropriate `/subpath/error` handler.
* ensure GetSubpathFromConfig returns trailing slashes deterministically
* fix error 404 handling
* redirect /subpath to /subpath/
This is necessary for the static handler to match, otherwise none of the
registered routes find anything. This also makes it no longer necessary
to add trailing slashes in the root router.
* Refactor context out of API packages
* Update function names per feedback
* Move webhook handlers to web and fix web tests
* Move more webhook tests out of api package
* Fix static handler
