[pypy-commit] [Git][pypy/pypy][branch/hpy-0.0.2] 8 commits: WIP: kill state.handles, since it is too confusing compared to the various

[Antonio Cuni]

Antonio Cuni pushed to branch branch/hpy-0.0.2 at PyPy / pypy


Commits:
2fada973 by Antonio Cuni at 2021-04-15T13:51:06+02:00
WIP: kill state.handles, since it is too confusing compared to the various
self.handles and it's too easy to use it by mistake.

Instead:

  1. introduce state.u_handles and state.d_handles, so that it's clear that
     they are for universal and debug handles

  2. the API functions receive directly the proper HandleManager instead of
     receiving the state. So, normal API functions get u_handles and DEBUG API
     functions get d_handles

test_noop passes, but all the rest is broken. The next commit will do a big
search&replace

--HG--
branch : hpy-0.0.2

- - - - -
ddca4620 by Antonio Cuni at 2021-05-01T10:14:52+02:00
WIP: search&replace to introduce the new handles argument to all existing API functions

--HG--
branch : hpy-0.0.2

- - - - -
8aace204 by Antonio Cuni at 2021-05-03T11:09:39+02:00
WIP: hg merge cfe54c8091a7, to pass (space, handles, ...) instead of (space, state, ...) to API functions.

This was a refactoring which I had lying around in my repo but was not pushed
before, so the commits diverged. See the commit message of cfe54c8091a7 for
more details.

--HG--
branch : hpy-0.0.2

- - - - -
101233e7 by Antonio Cuni at 2021-05-03T11:15:20+02:00
WIP: fix HPyType_FromSpec: now we no longer need to explicitly pass debug=True or debug=False, because 'handles' automatically contains the appropriate HandleManager

--HG--
branch : hpy-0.0.2

- - - - -
67183ce8 by Antonio Cuni at 2021-05-03T11:23:45+02:00
WIP: handles is no longer a module but an instance of HandleManager

this is an automatica search&replace to kill all '... import handles' and turn
'state.handles.*' into 'handles.*', because 'handles' is now a HandleManager
which is passed around

--HG--
branch : hpy-0.0.2

- - - - -
c391a4d1 by Antonio Cuni at 2021-05-03T11:26:29+02:00
WIP: rename _hpy_universal.handles into handlemanager, to avoid confusion with 'handles' which in most places is now an instance of HandleManager

--HG--
branch : hpy-0.0.2

- - - - -
c5b5e0cd by Antonio Cuni at 2021-05-03T11:33:41+02:00
WIP: more progress towards fixing test_hpytype.py, by passing around 'handles'
instead of relying on 'state'

--HG--
branch : hpy-0.0.2

- - - - -
4211456d by Antonio Cuni at 2021-05-03T11:41:43+02:00
bah, I still had handles.pyc laying around so there were tons of '... import
handles' which worked by mistaked. Ensure to kill all the leftovers

--HG--
branch : hpy-0.0.2

- - - - -


26 changed files:

- pypy/module/_hpy_universal/apiset.py
- pypy/module/_hpy_universal/bridge.py
- pypy/module/_hpy_universal/handles.py → pypy/module/_hpy_universal/handlemanager.py
- pypy/module/_hpy_universal/interp_builder.py
- pypy/module/_hpy_universal/interp_bytes.py
- pypy/module/_hpy_universal/interp_call.py
- pypy/module/_hpy_universal/interp_cpy_compat.py
- pypy/module/_hpy_universal/interp_descr.py
- pypy/module/_hpy_universal/interp_dict.py
- pypy/module/_hpy_universal/interp_err.py
- pypy/module/_hpy_universal/interp_extfunc.py
- pypy/module/_hpy_universal/interp_float.py
- pypy/module/_hpy_universal/interp_hpy.py
- pypy/module/_hpy_universal/interp_list.py
- pypy/module/_hpy_universal/interp_long.py
- pypy/module/_hpy_universal/interp_module.py
- pypy/module/_hpy_universal/interp_number.py
- pypy/module/_hpy_universal/interp_object.py
- pypy/module/_hpy_universal/interp_slot.py
- pypy/module/_hpy_universal/interp_tracker.py
- pypy/module/_hpy_universal/interp_tuple.py
- pypy/module/_hpy_universal/interp_type.py
- pypy/module/_hpy_universal/interp_unicode.py
- pypy/module/_hpy_universal/state.py
- pypy/module/_hpy_universal/test/test_apiset.py
- pypy/module/_hpy_universal/test/test_handles.py → pypy/module/_hpy_universal/test/test_handlemanager.py


View it on Heptapod: https://foss.heptapod.net/pypy/pypy/-/compare/376d23a976247942052acaf79fbde42f95ddd2f0...4211456d494b166d8f3816e9f8f966ad254c7aa6

-- 
View it on Heptapod: https://foss.heptapod.net/pypy/pypy/-/compare/376d23a976247942052acaf79fbde42f95ddd2f0...4211456d494b166d8f3816e9f8f966ad254c7aa6
You're receiving this email because of your account on foss.heptapod.net.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.python.org/pipermail/pypy-commit/attachments/20210503/09b0dd90/attachment-0001.html>