gh-137808: use argument clinic for _thread.lock and _thread.RLock

[kumaraditya303]

• Issue: use argument clinic for _thread.lock and _thread.RLock #137808

[serhiy-storchaka]

I am surprised that this change did not affect test_inspec test for _thread signatures. Could you please show the diff for the pydoc output for _thread?

[kumaraditya303]

Could you please show the diff for the pydoc output for _thread?

41c41
<      |  __exit__(self, /, *exc_info)
---
>      |  __exit__(self, exc_type=None, exc_value=None, exc_tb=None, /)
48c48,50
<      |      Lock the lock.  Without argument, this blocks if the lock is already
---
>      |      Lock the lock.
>      |
>      |      Without argument, this blocks if the lock is already
65c67,69
<      |      Release the lock, allowing another thread that is blocked waiting for
---
>      |      Release the lock.
>      |
>      |      Allows another thread that is blocked waiting for
84c88
<      |  __exit__(self, /, *exc_info)
---
>      |  __exit__(self, exc_type=None, exc_value=None, exc_tb=None, /)
91c95,97
<      |      Lock the lock.  `blocking` indicates whether we should wait
---
>      |      Lock the lock.
>      |
>      |      `blocking` indicates whether we should wait
105,106d110
<      |      locked()
<      |
110c114,116
<      |      Release the lock, allowing another thread that is blocked waiting for
---
>      |      Release the lock.
>      |
>      |      Allows another thread that is blocked waiting for
142c148
<      |  __exit__(self, /, *exc_info)
---
>      |  __exit__(self, exc_type=None, exc_value=None, exc_tb=None, /)
149c155,157
<      |      Lock the lock.  Without argument, this blocks if the lock is already
---
>      |      Lock the lock.
>      |
>      |      Without argument, this blocks if the lock is already
166c174,176
<      |      Release the lock, allowing another thread that is blocked waiting for
---
>      |      Release the lock.
>      |
>      |      Allows another thread that is blocked waiting for

[serhiy-storchaka]

What about the acquire() methods? Should not they now have machine readable signatures?

Oh, and pydoc does not show proivate methods for module and class, you need to do direct request for _thread.RLock._acquire_restore.

[kumaraditya303]

What about the acquire() methods? Should not they now have machine readable signatures?

Help on method descriptor acquire in _thread.RLock:

_thread.RLock.acquire = acquire(self, /, blocking=True, timeout=-1) unbound _thread.RLock method
    Lock the lock.

Help on method descriptor acquire in _thread.lock:

_thread.lock.acquire = acquire(self, /, blocking=True, timeout=-1) unbound _thread.lock method
    Lock the lock.

The signatures looks fine to me for acquire methods, what do you mean by "machine readable signatures"? are these two signatures incorrect?

[serhiy-storchaka]

Aha, this is because signatures were already encoded in the docstrings! Now Argument Clinic will generate identical signatures, this is why there was not difference in the pydoc output.

[serhiy-storchaka]

LGTM. 👍

[ZeroIntensity]

LGTM as well