lp:~cjwatson/storm/py3-strings

Branch merges

Propose for merging

No branches dependent on this one.

Merged into lp:storm at revision 526

Simon Poirier (community): Approve on 2019-09-16

Diff: 848 lines (+140/-103)

20 files modified

storm/cextensions.c (+9/-7)
storm/database.py (+3/-1)
storm/databases/postgres.py (+2/-5)
storm/databases/sqlite.py (+13/-9)
storm/expr.py (+26/-24)
storm/info.py (+2/-2)
storm/properties.py (+1/-0)
storm/references.py (+1/-1)
storm/sqlobject.py (+15/-7)
storm/tests/databases/base.py (+2/-1)
storm/tests/databases/postgres.py (+6/-2)
storm/tests/expr.py (+2/-2)
storm/tests/mocker.py (+1/-1)
storm/tests/sqlobject.py (+2/-2)
storm/tests/store/base.py (+4/-4)
storm/tests/tracer.py (+2/-1)
storm/tests/variables.py (+15/-11)
storm/tracer.py (+7/-2)
storm/tz.py (+2/-2)
storm/variables.py (+25/-19)

api

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

514. By Colin Watson on 2019-09-16

Simplify type-checking conditions in Compile.__call__.

513. By Colin Watson on 2019-08-11

Update string handling for Python 3.

This is inspired by work done by Thiago Bellini. I've taken a different
approach in a few areas which are useful to discuss briefly here (and some
of this should end up in release notes):

 * I generally wrote "bytes" rather than "six.binary_type", since it's
   shorter and works in all supported versions of Python.

 * In the SQLite backend, there's no need to use memoryview, because the
   sqlite3 module in Python 3 automatically converts between the SQLite BLOB
   type and bytes.

 * Some exception messages have changed slightly for clarity.

 * On Python 3, raw=True and token=True in storm.expr.Compile.__call__ only
   treats str specially, not bytes and str, because ultimately the compiler
   is assembling a text string to send to the database.

 * On Python 3, storm.tracer.BaseStatementTracer.connection_raw_execute
   renders text parameters using ascii() rather than by encoding to bytes
   and then calling repr(). While this does result in slightly different
   output from Python 2, it's normally more useful since the encoding is in
   terms of Unicode codepoints rather than UTF-8.

 * storm.sqlobject.AutoUnicodeVariable (and hence StringCol) explicitly
   documents that it only accepts text on Python 3, since native strings are
   already Unicode there so there's much less need for the porting
   affordance.

512. By Colin Watson on 2019-08-11

Remove special case for unicode query parameters to PostgreSQL.

These have been supported natively by psycopg2 since 1.99.3 or thereabouts,
so we don't need to handle them ourselves any more.

511. By Colin Watson on 2019-08-11

Be more explicit about string types in tests.

The distinction between '...' and b'...' matters in Python 3, so be explicit
where it matters in tests, in preparation for corresponding changes to Storm
itself.

510. By Thiago Bellini on 2019-08-09

Use Python 3-compatible raise syntax. [r=simpoir]

509. By Colin Watson on 2019-07-02

Wrap DB-API exceptions rather than injecting virtual base classes. [r=simpoir]

508. By Colin Watson on 2019-06-19

Rename double-underscore identifiers in C extensions. [r=simpoir]

507. By Colin Watson on 2019-06-07

Import from six.moves where appropriate. [r=simpoir]

506. By Thiago Bellini on 2019-06-07

Use Python 3-compatible exec functions. [r=cjwatson]

505. By Martin v. Löwis on 2019-06-07

Make cextensions module build with 3.x. [r=cjwatson]