[docs] SemLock acquire() keyword arg 'blocking' is invalid (issue 23484)

[python at discontinuity.net]

Hi Berker --

Unrelated to this but thanks also for finishing off issue23400 today.


Davin




https://bugs.python.org/review/23484/diff/14347/Doc/library/multiprocessing.rst
File Doc/library/multiprocessing.rst (right):

https://bugs.python.org/review/23484/diff/14347/Doc/library/multiprocessing.rst#newcode1199
Doc/library/multiprocessing.rst:1199: .. versionchanged:: 2.7
Looking back through the prior branches, it turns out the timeout
parameter was introduced in 2.6 (same branch where multiprocessing
itself was introduced) and it is present in 3.0 upwards too. 
Unfortunately I trusted or misinterpreted the docs that suggested
otherwise -- makes your reacting to this an especially good catch.

I think the right thing to do is to remove the versionchanged line
altogether since both 'block' and 'timeout' have been there pretty much
since the beginning.  I'll make that quick change and update the patch
files.

On 2015/04/02 18:00:10, davin wrote:
> On 2015/04/01 17:36:10, berkerpeksag wrote:
> > If the timeout parameter was added in 2.7 and 3.0, we can remove
this
> directive
> > on 3.4 and default branches since it's there for the beginning of
Python 3.
> > 
> > On the other hand, for example, if it was added in 2.7 and 3.2, we
should
> change
> > it to .. versionchanged:: 3.2 in Python 3 documentation.
> 
> Makes sense -- I will double-check when it got added to Python 3 and
make that
> change.
> 
> My thinking was that bringing attention to this, even if it has been
there for a
> while, is a good thing because this is the first time it's been
reflected in the
> documentation.  But if your immediate reaction is that this looks
weird (coming
> from 2.7 or 3.0), then my idea was the wrong way to go.

https://bugs.python.org/review/23484/