mariocj89 / hubsync Goto Github PK

Instead of mocking the github api one by one we should download all responses and update them as needed within the tests. This wont break the tests when we add new features to the script in the future.

On the first run (if no config file is found), guide the user over all the options he have and create a config file for him interactively.

See #7 as might be useful for this task

The config file is the main configurable of the script, all options should be documented.

Add an option to allow the user to automatically create forks for all the repos in organizations belongs to and add a remote for it.

Basic configuration should be store within a config file.
Like the root of the local workspace (instead of the current path)

We need a CI and PR validation tool

Update remote forks using hubsync
We can use the github api to remove stale branches and keep up to date the user forks.

pygit2 seems to have a nicer api than the git api this project is using

Add an option to the config file to provide a custom certificate

We need support for custom certificates as many enterprise github instances might be behind a proxy or use custom certs

Store Org/Repo configuration and default actions within the config file ( #2 ) so users dont get prompt for configured options

It might be more interesting to use yaml than a config file as it has proper support for structures like lists

This prevents for example being able to provide a list of commands (pre/post) at the moment users need to chain all their commands.

github api key is available within your local config, it should be read from it
(configurable through the main config file. See #2 )

When installing hubsync and running it:
Traceback (most recent call last):
File "/home/mcorcherojim/.local/bin/hubsync.py", line 8, in
import hubsync.github
File "/home17/mcorcherojim/.local/bin/hubsync.py", line 8, in
import hubsync.github
ImportError: No module named github

Issue is that it does not import the package hubsync but the python module (the script)

A possible solution is to move the script to a bin/scripts folder and remote the .py from it.

Allow users to set hubsync for multiple github instances (multiple enterprise or single enterprise and github for example) This requires config for multiple users, api key and api url

Add an option to configure hubsync to checkout all repos within the workspace root.
Would be better if it can be configured per org.

In brief the idea is to get

my_ws
|-- org1
    |-- repo1
    |-- repo2
| -- org2
    |-- repo3

as:

my_ws
|-- repo1
|-- repo2
|-- repo3

Provide within the config file a way to run shell commands when syncing repo/org/workspace

When checking out all the repos of an organization or an user, we need to be able to choose whether we want to clone locally not only the source repos but also the forks (No by default)

Users repositories can be included in an organization with the user's name

See http://click.pocoo.org/5/