Stop relying on SHA internals

[jakobnissen]

MD5 currently relies on SHA internals, including:

• Unexported, undocumented functions transform!, pad_remainder!, lrot, buffer_pointer, blocklen, short_blocklen, state_type and digestlen
• Unexported, undocumented type SHA_CTX

Since Julia stdlibs are about to be set free, this could get very bad if SHA ever begins to cut more releases.

[jw3126]

or maybe this package could move into SHA.jl?

[oxinabox]

Coping what i said over in JuliaLang/julia#51687 (comment):

For historical note: MD5.jl was written using those internals when both it and SHA.jl were external libraries with few users.
Including changes to SHA.jl specifically for supporting MD5.jl: JuliaCrypto/SHA.jl#49
Probably what should have happened then is we should have documented those internal abstractions as part of the public API.
Or we should have stuck MD5 into SHA.jl

TBH we didn't worry about anything like that then because it seemed unlikely that the internals of SHA.jl would ever change. And we were pretty close to correct given they didn't change for over 5 years.
I think its not terrible to assume they won't start changing rapidly once it is out of the stdlib.
So I am not to worried about addressing this.