TranscendComputing / StackStudio

The front end for private and public clouds.
Other
2 stars 0 forks source link

Remove all dead code #145

Closed lifeBCE closed 10 years ago

lifeBCE commented 10 years ago

There is a LOT of dead code everywhere throughout StackStudio. There should be a concerted effort to clean this up before moving on to any large efforts.

  1. Old references and logic pertaining to stackplace and more
  2. Unused view, etc files from old functionality or refactored functionality.
  3. Massive commented blocks, not sure if these were supposed to be revisited eventually or not
  4. Many XXX and TODO comments that should be addressed.
  5. Make some sense out of js/views/*

    5 is due to the fact that we only have 23 named routes pointing at 10 different views and we have about 90 different view files under js/views/*. I don't think that the other 80 view files are all sub-views (especially since we only have 30 collections and 35 models).

lifeBCE commented 10 years ago

As part of this effort, I see the fallback logic to local vendor libs being a prime candidate. This logic is reliant on Grunt curl downloading all of these vendor libs from CDN first in order to even be present in js/vendor but this wont happen if the user is offline which is what this logic is supposed to address. Otherwise we are required to obtain vendor lib, add to source control and track as part of application.

Either use case seems somewhat pointless since the user will not be able to reach any cloud instances for management purposes if they do not have network connectivity. IMO network connectivity is a requirement for this application and supporting a fallback to offline mode might be overkill.

huxoll commented 10 years ago

Actually, the curls are primarily for non-CDN resources; there are JS libs that aren't hosted on CDN, and we must pull them from github or some such thing, which we would clearly not do in production. It would also be dangerous to copy these into our repo, because then we'd lose the upstream improvements/fixes.

Once we require some curls, it's convenient to cache the JS locally. I, for example, have done StackStudio development on a plane, and you can at least run unit tests and simple web interactions while disconnected.

lifeBCE commented 10 years ago

Makes sense.

We currently have most of these committed to repo though but that is probably because they got pushed prior to .gitignore being updated. And they are overwritten each time by curl anyway.