Use UTF-8 character set

[chrisimcevoy]

Hi, thanks for this great action! I have a hopefully small request please.

I am working on an ODBC library written in Python. My CI test matrix runs the test suite using different combinations of OS and DBMS.

On Windows with a PostgreSQL backend (created via this action), I get one failed test. It is a round-trip of inserting/reading a UTF-8 string to check that what is read matches what was inserted. The same test passes for all other combinations of OS and DBMS, but for Postgres on Windows I get this error back from the ODBC driver:

22P05ERROR: character with byte sequence 0xf0 0x9f 0x8c 0x9c in encoding "UTF8" has no equivalent in encoding "WIN1252"

This error occurs on the insert SQL statement. The issue is that the database has a character encoding of WIN1252 and the unusual UTF-8 character in the string cannot be converted to the server's encoding. (SO link)

Would it be possible for the database to be created with UTF-8 character set? That way, I would be able to insert and read back the string in this test.

I'm not an expert, but I think I am asking for something like this:

initdb -E UTF8

[ikalnytskyi]

Hey @chrisimcevoy,

Thanks for reporting the issue. I'll see what I can do. Let me check some docs regarding client/server encodings and encoding negotiation (if any).

[ikalnytskyi]

I have published ikalnytskyi/action-setup-postgres@v3 where encoding=UTF-8 and locale=en_US.UTF-8 are enforced on Linux, macOS and Windows runners.