PARQUET-1300: [C++] Implement encrypted Parquet read and write support

[thamht4190]

I've just move from patch from apache/parquet-cpp PR #475 with @wesm's great help. Will update CMakeLists.txt and example for encryption later.

[wesm]

Plan to get the examples migrated this week. Will also review this patch when I can.

@ggershinsky@majetideepak there may have been some unaddressed comments from the parquet-cpp PR, so some additional review may be necessary

[nguoithichkhampha]

@wesm@ggershinsky@majetideepak . Can you help us review this code ? We are planing to use this feature on our product. We want to make sure the file which is generated by C++ code can decrypt by java code.

[ggershinsky]

@nguoithichkhampha The encryption design is currently being voted by the community. The process seems to be converging, but it is not over yet. Once the design is signed off, we'll resume the work on Java and C++ versions. Today, the C++ lags behind a bit, as there were recent changes in the spec and Java code (eg support for legacy readers). I keep the track of these - after the voting is finished, I'll get in touch with @thamht4190 and other C++ community members to get this going.

[nguoithichkhampha]

@ggershinsky , thanks for quick reply. Can you please tell me the moment will have final design for encryption ?

[ggershinsky]

@nguoithichkhampha , no problem, our timezones are not that different :).
Certainly, I'll inform you the moment the encryption design is formally signed off by the community.

[majetideepak]

This PR needs to be rebased. @thamht4190 do you want to do this?

[thamht4190]

Yes, I'll do it @majetideepak

[ggershinsky]

@thamht4190 I've updated the crypto package to match the signed-off specification. Please check out the changes at
#3520
IV changes are not visible, but writing the buffer size to ciphertext will affect you code. Basically,
when writing/reading encrypted Thrift structures, no need write/read their length anymore.

[thamht4190]

Thanks @ggershinsky. I've just pushed my code change to adapt your change at #3520.
I also rebased as well @majetideepak. Please check it out.

[ggershinsky]

@thamht4190 , thanks for handling the rebase.
@wesm@majetideepak , no more open issues, except for the mingw in one config with double read ; we're working on debugging.

[thamht4190]

@wes@majetideepak since this pull has grown too much, should I close it and open a fresh one from the same branch?

[wesm]

Yes, can you open a new PR so I can review again with less GitHub UI lag? Thanks

[thamht4190]

Thanks @wesm. Will do it now.