Each branch must have a name in the following format with a proper prefix and number of Github's ticket it refers to:
prefix/000-branch-name
Allowed prefixes:
- feature
- bugfix
- docs
- test (when adding missing tests)
- improvement
- refactor
- chores (for those tasks that needs be done for infrastructure compliance, some policy change)
- ...
Each commit (except for merge commits) must have a name in the following format:
000 - description of what was changed
Each pull request must have a name in the following format where 000
represents a ticket number which this PR refers
to:
000(-1) - description of what was changed
When you'are creating more than one PR for a ticket you should use postfixes like -1
, -2
to define order of merging.
Each pull request must have a nice description of what was changed. The description should start with a list of all corresponding tickets it closes or refers to. Then it must have information about all necessary actions need to be taken during deployment like database migrations or additional configuration.
Pull requests should be as concise as possible to make them easier to review. If your PR gets bigger and verbose, consider slice it into a few smaller but fully functional ones.
There is a basic regression test of Lorre/Conseil in the it
folder.
The test needs a pre-filled database to run on, and defined using the usual .conf
files. It starts a conseil process and verifies the responses to predefined endpoints against expected json results.
You can run the regression calling
sbt it:run
If you need to pre-fill the database with current tezos data, pass-in as argument the name of a tezos network as configured in your usual .conf
file. E.g.
sbt "it:run staging"
Will look for a staging
network defined for the tezos platform in your configuration.
This will fetch a fixed number of blocks from the node, that will be used by the conseil test.
You can contribute by adding further test cases, which will be useful to verify that
- any refactor of Lorre's code will not break existing behaviour
- expected changes to the outputs from new features will be correctly handled