Skip to content

Latest commit

 

History

History
56 lines (50 loc) · 2.33 KB

File metadata and controls

56 lines (50 loc) · 2.33 KB

The everything list

Dev env setup

Nix package manager

There is a startup *nix file, need to get the versions the same 
as the cabal file* How to do that?

Chef (we probably need both chef and nix because liferay is probably chef friendly)

Chef and nix seem to be competing for the same space* Though nix is probably more 
functional therefore suitable for the current application*

Install ruby -- Done

  • Setup java -- Done
  • Setup maven -- Done
  • Setup ant -- Done
  • Setup scala -- Done
  • Setup spray -- TBD
  • Setup liferay : 62 branch and master
  • Gridgain
  • Gridgain enable liferay installation
  • Postgres
  • Eclipse -- Done
  • Tmux -- Done
  • Setup liferay for automated tests
  • Implement security cookbook recepies
  • Install OAUTH provider
  • Grails 224
  • Sublime haskell -- Done
  • Sublime git -- Done
  • Enable erp CI using travis
  • Enable osgi for liferay
  • Enable osgi in general?
  • Install haxe ?? or should it be spray?
  • Install and get familiar with xmonad -- This can be sort of tmux?
  • Get better at using the ghci: how to look at the type signature of methods not exposed in the module
  • As a corollary, improve the interface in the current application* Keep a lot of things private
  • Really really add features to the model* We need to start processing messages for the model*
  • Closure compiler - understand code -- A bit**but that is a lot of code*

Dev tasks (should probably go in the readme*md)

  • Add ErpHelper module to handle imports for the project in one place
  • Replace JSON messages (a bit clunky) with show instances for each data constructor* Make this configurable to be able to switch it between raw messages and data messages*
  • Checkin a reference test output that can be reliable across different runs - this would most likely need a split between the server and the client threads and only single threaded tests to prevent interleaving of output?? -- The principle being that if the output has changed, then there is a either a bug fix so the change should equal the fix or there is a new feature* In all other circumstances, the output needs to be identical across all tests
  • Move the current thead into snap
  • Implement product rest services*
  • Make lenses for manipulating the model*
  • RabbitMQ or ZeroMQ for messaging infrastructure support*
  • Distributed haskell for distributing the application*