Added a --debug-sql option for tests and runtests.py which outputs the
SQL logger for failing tests. When combined with --verbosity=2, it also
outputs the SQL for passing tests.

Thanks to Berker, Tim, Markus, Shai, Josh and Anssi for review and
discussion.

It was mainly for MySQL on Python 3, but now the current
recommended MySQL driver for Python 3 (mysqlclient) does support
binary fields, it is unneeded. Refs #20377.

This reverts commit 832b4a57.
We officially don't support MySQL on Python 3.2. Refs #20380.

Recommending Template(template_code) was dumb. Described alternatives.

Accounted for multiple template engines and made a few small fixes.

Accounted for multiple template engines and made a few small fixes.

This makes room for a more general introduction about templating.

Updated some links to point to the new location, but kept those that
didn't talk specifically about the DTL.

Thanks to Collin Anderson for the review and suggestions and Tim for the
final review.

Added support for PostgreSQL range types to contrib.postgres.

- 5 new model fields
- 4 new form fields
- New validators
- Uses psycopg2's range type implementation in python

These refactorings making overriding some text based lookup names on
other fields (specifically `contains`) much cleaner. It also removes a
bunch of duplication in the contrib.postgres lookups.

Thanks Peter Schmidt for the report and the initial patch.
Thanks to ​Oktay Sancak for writing the original failing test and
Alvin Savoy for supporting contributing back to the community.

Thanks Markus Holtermann and Tim Graham for the review.

Thanks Loïc Bistuer and Tim Graham for the discussion and review.

Refactored compiler SELECT, GROUP BY and ORDER BY generation.
While there, also refactored select_related() implementation
(get_cached_row() and get_klass_info() are now gone!).

Made get_db_converters() method work on expressions instead of
internal_type. This allows the backend converters to target
specific expressions if need be.

Added query.context, this can be used to set per-query state.

Also changed the signature of database converters. They now accept
context as an argument.

This reverts a change made in 40c60efe
which was incorrect and caused CacheKeyWarnings.

This fixes a regression caused by f7c287fc. Thanks Markus Holtermann
for identifying the regression.

`verbose_name_raw` return value comes from `force_text()` anyway.