Conversation
|
Will you consider where to stick this in the course outline, as well as add this into .travis.yml? |
|
|
||
| Check...for better or worse. | ||
|
|
||
| So are parameterized modules worth it? You decide! No newline at end of file |
There was a problem hiding this comment.
This seems too negative. The notion of parameterized modules is necessary, at the very least, to remove code duplication. Now, the current module system could be better (GaloisInc/cryptol#815).
You know, considering that the module system upgrade proposal seeks to remove the backtick imports, it might be worth while to rework this lab by removing them. Would that be possible?
There was a problem hiding this comment.
Somebody smarter than me might could. I didn't see a way to DRYly instantiate BlockCipher, add intermediate parameters for TWINE, and have TWINE_80 and TWINE_128 export them without introducing layers of backtick imports. The reworked module system should enable specs to add intermediate parameters more directly. In the meantime, others are welcome to simplify the current modules.
I suppose I should write some sort of narrative around this...