This has multiple benefits:
- Respects the "Do not mock what you do not own" principle, instead we mock `log.raw` when necessary
- Lets us not re-assign `console.log`, which breaks as Mocha uses `console.log` as well
- Save and restore initial `log.raw` in test hooks (before/after), otherwise this would break Mocha/Chai
I used `npm run coverage` while *not* excluding the test folder to detect dead code in our test folder, it is actually pretty useful to do so (as a one-shot, not to do that in our config).
Only remaining unreached path is L40 in `test/plugins/auth/ldap.js`, but it does seem to me that it might be useful in case of failures, so I preferred to leave it there.
- Load up to 5 previews per message (to avoid abuse)
- Do not load multiple times the same URL
- Prepare preview containers per message instead of appending (to maintain correct order)
- Store an array of previews instead of a single preview in `Msg` objects
- Consolidate preview rendering for new messages and upon refresh/load history (when rendering entire channels)
- Update `parse` tests to reflect previous point
- Add test for multiple URLs
- Switch preview tests from `assert` API to `expect` API
Several ES6 additions are only available in strict mode. Example:
> SyntaxError: Block-scoped declarations (let, const, function, class) not yet supported outside strict mode
Strict mode was also enabled in a few of our files already, and it is a good thing to have anyway.
