Cache size data by commit hash, and enable direct commit comparison via CLI

[ItsNickBarry]

Example usage: hh size-contracts --compare 103541870332c9fb9125fad3684fcee613f13960

[ItsNickBarry]

Another possibility would be "named" outputs. Multiple such outputs could be cached.

For example:

hh size-contracts --name asdf
# later...
hh size --compare asdf

Using --compare with a git commit hash could still work here (as long as nobody uses a git commit hash as a --name).

[ItsNickBarry]

Make --compare a variadic argument, so that it works like git diff:

• no arguments compares uncommitted changes to current commit
• 1 argument compares uncommitted changes to given reference
• 2 arguments compares 2 references

[ItsNickBarry]

Checking out git references is not enough; also have to make sure dependencies from that reference are installed.

Maybe skip the checkout process and only allow git references that have cached size data. Include something like git reflog to list valid references.

Make sure not to cache size data generated in an unclean working directory.