asyncpg v0.22.0

[elprans]

A new asyncpg release is here.

Notable additions include Python 3.9 support, support for recently added
PostgreSQL types like jsonpath, and last but not least, vastly
improved executemany() performance. Importantly, executemany() is
also now atomic, which means that either all iterations succeed, or
none at all, whereas previously partial results would have remained in
place, unless executemany() was called in a transaction.

There is also the usual assortment of improvements and bugfixes, see the
details below.

This is the last release of asyncpg that supports Python 3.5, which has
reached EOL last September.

Improvements

• Vastly speedup executemany by batching protocol messages (Batch executemany #295)
  (by @fantix in 690048d for Batch executemany #295)

• Allow using custom Record class
  (by @elprans in db4f1a6 for Add typings #577)

• Add Python 3.9 support (Fix compilation under Python 3.9+ #610)
  (by @elprans in c05d726 for Fix compilation under Python 3.9+ #610)

• Prefer SSL connections by default (Prefer SSL connections by default #660)
  (by @elprans in 16183aa for Prefer SSL connections by default #660)

• Add codecs for a bunch of new builtin types (Add codecs for a bunch of new builtin types #665)
  (by @elprans in b53f038 for Add codecs for a bunch of new builtin types #665)

• Expose Pool as asyncpg.Pool (Added Pool to public API #669)
  (by @rugleb in 0e0eb8d for Added Pool to public API #669)

• Avoid unnecessary overhead during connection reset (Avoid unnecessary pg_listening_channels queries on connection release #648)
  (by @kitogo in ff5da5f for Avoid unnecessary pg_listening_channels queries on connection release #648)

Fixes

• Add a workaround for bpo-37658
  (by @elprans in 2bac166 for #21894)

• Fix wrong default transaction isolation level (Fix wrong default transaction isolation level #622)
  (by @fantix in 4a627d5 for Fix wrong default transaction isolation level #622)

• Fix set_type_codec() to accept standard SQL type names (Fix set_type_codec() to accept standard SQL type names #619)
  (by @elprans in 68b40cb for Fix set_type_codec() to accept standard SQL type names #619)

• Ignore custom data codec for internal introspection (Ignore custom data codec for internal introspection #618)
  (by @fantix in e064f59 for Ignore custom data codec for internal introspection #618)

• Fix null/NULL quoting in array text encoder (Fix null/NULL quoting in array text encoder #627)
  (by @fantix in 92aa806 for Fix null/NULL quoting in array text encoder #627)

• Fix link in connect docstring (fix link in connect docstring #653)
  (by @samuelcolvin in 8b313bd for fix link in connect docstring #653)

• Make asyncpg work with pyinstaller (fix(pgproto): bump version to fix pyinstaller #651)
  (by @Atem18 in 5ddabb1 for fix(pgproto): bump version to fix pyinstaller #651)

• Fix possible AttributeError exception in ConnectionSettings (Fix AttributeError exception #632)
  (by @petriborg in 0d23182 for Fix AttributeError exception #632)

• Prohibit custom codecs on domains
  (by @elprans in 50f964f for Custom Codec doesn't work with DOMAIN types #457)

• Raise proper error on anonymous composite input (tuple arguments) (Raise proper error on anonymous composite input (tuple arguments) #664)
  (by @elprans in 7252dbe for Raise proper error on anonymous composite input (tuple arguments) #664)

• Fix incorrect application of custom codecs in some cases (Untangle custom codec confusion #662)
  (by @elprans in 50f65fb for Untangle custom codec confusion #662)

[Atem18]

@elprans Do you think it's ok to merge it or should the failing build be fixed ?

I am asking because it would be cool to have a new release on pypi for the Pyinstaller support instead of relying on master via pip.

Thanks in advance.

[elprans]

@Atem18 the failure in the aarch64 build is a real bug and, ideally, should be fixed before the release. I'll try to get to doing so over the coming weekend.

[Atem18]

@elprans Got it, thanks a lot and keep up the good work !