Handle *.sol files that are invalid Solidity

[area]

I introduced a nice bug in #147 which this fixes. In order to remove constant from all functions for instrumentation to take place, including those imported from ./node_modules, I ran the (increasingly badly named...) preprocessor over all .sol files that we find in the tree. Unfortunately, if we're using copyNodeModules:true, then that includes the solidity-coverage repository, which includes SimpleError.sol, which deliberately has a syntax error in it! This caused an uncaught error to be thrown, and solidity-coverage to fail to run.

I did think about just skipping SimpleError.sol itself, but don't want to be in the position where we break if any other package installed uses .sol files that aren't Solidity. I therefore leave any *.sol files that fail to parse alone, and generate a warning in case this is unexpected by the user. SimpleError.sol is now added to the truffle mock in order to prevent a regression along these lines.

Fixes #155

[area]

Check I've not done something silly again here @cgewecke, but otherwise I'm happy for this to be merged.

EDIT: I think coverage changes are because these tests aren't run on CI.

[codecov-io]

Codecov Report

Merging #156 into master will decrease coverage by 0.5%.
The diff coverage is 93.33%.

@@            Coverage Diff             @@
##           master     #156      +/-   ##
==========================================
- Coverage    97.6%   97.09%   -0.51%     
==========================================
  Files           6        6              
  Lines         376      379       +3     
  Branches       86       86              
==========================================
+ Hits          367      368       +1     
- Misses          9       11       +2

Impacted Files	Coverage Δ
lib/preprocessor.js	94.73% <93.33%> (-5.27%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 194a8aa...a1a55ef. Read the comment docs.

[cgewecke]

@area Yes looks good! What a weird problem. Merging and printing.

Side note: at the moment we have lost our E2E tests because Zeppelin is not upgrading their deps to the fork. Will the colonySale run clean on 0.4.x? We need a new (or additional) baseline.

[area]

Sorry, I don't understand why the Zeppelin contracts don't work?

[cgewecke]

Because their tests check for invalid opcode so most of them crash.
[EDIT] - Link to their CI for 0.3.5 which runs clean on the PR the testrpc maintainer opened for them, upgrading to 6.0

[area]

Very odd that they're not upgrading, especially with the PR's sat right there...