[WIP] Improve gp_hyperloglog codes and add tests for it.

[higuoxing]

This patch helps remove redundant base64 encoding and decoding functions in gp_hyperloglog.c and add test cases for gp_hyperloglog.

1. Postgres upstream provides the implementation of base64 encoding and decoding utilities. The difference between Postgres's base64 and the original base64 is that the original encoding function obeys the 76-character per-line rule (it inserts a \n to the encoded string every 76-character). I think it doesn't matter whether to start a new line every 76-characters and it's good to remove redundant codes.

2. Test cases for gp_hyperloglog are added in this patch (I don't know why there isn't any test for it).

Here are some reminders before you submit the pull request

• [x] Add tests for the change
• [x] Pass make installcheck

[kainwen]

Looks Good. Why WIP?

[kainwen]

Hi, interesting PR, any plan to make progress?

[higuoxing]

Hi, interesting PR, any plan to make progress?

Yes. My concern is that whether it's ok to print out the base64 string without obeying the 76 characters/line rule. Both of them are legal.

https://en.wikipedia.org/wiki/Base64