bpo-32996: The bulk of What's New in Python 3.7 #6978

elprans · 2018-05-18T22:40:27Z

This represents the bulk of the content for What's New in Python 3.7.

https://bugs.python.org/issue32996

This represents the bulk of the content for What's New in Python 3.7.

1st1 · 2018-05-18T22:42:16Z

I'll start reviewing this patch later today.

ned-deily · 2018-05-18T22:52:06Z

Thanks, @elprans! I'll be reviewing it this weekend as well.

1st1 · 2018-05-18T23:10:43Z

Doc/library/ensurepip.rst

@@ -78,9 +78,6 @@ options:

 Providing both of the script selection options will trigger an exception.

-.. versionchanged:: 3.7.0
-   The exit status is non-zero if the command fails.
-


Right, this was backported to 3.6 and 2.7.

1st1 · 2018-05-18T23:18:22Z

Doc/whatsnew/3.7.rst

+
+Improved conistency of exceptions raised by :func:`issubclass` when called
+with a non-class and an abstract base class as the first and second
+arguments, respectively.  (Contributed by Josh Bronson in :issue:`33018`.)


I'd also add a ref to optimizations section's paragraph that covers ABCMeta being rewritten in C.

Note the "conistency"

1st1 · 2018-05-18T23:21:46Z

Doc/whatsnew/3.7.rst

+(Contributed by Yury Selivanov in :issue:`32251`.)
+
+The new :func:`asyncio.get_running_loop` function returns the currently
+running loop, or ``None`` if no loop is running.  This is in contrast with


This needs to be fixed: if no loop is running it raises a RuntimeError.

1st1 · 2018-05-18T23:22:29Z

Doc/whatsnew/3.7.rst

+:meth:`Future.get_loop() <asyncio.Future.get_loop>` methods
+return the instance of the loop on which a task or a future were created.
+(Contributed by Yury Selivanov in :issue:`32415`.)
+:meth:`Server.get_loop() <asyncio.Server.get_loop>` allows to do the same for


Need a new line here.

1st1 · 2018-05-18T23:25:17Z

Doc/whatsnew/3.7.rst

+(Jelle Zijlstra with further improvements by INADA Naoki, Serhiy Storchaka,
+and Raymond Hettinger in :issue:`28638`.)
+
+Additionally, ``collections.namedtuple()`` now supports default values.


I'd remove "Additionally", feels a bit out of place.

1st1 · 2018-05-18T23:29:46Z

Doc/whatsnew/3.7.rst

+The new :func:`sys.get_coroutine_origin_tracking_depth` function returns
+the current coroutine origin tracking depth, as set by
+the new :func:`sys.set_coroutine_origin_tracking_depth`.
+(Contributed by Nathaniel J. Smith in :issue:`32591`.)


I'd also mention here that asyncio now uses this new mechanism, and that this new API is intended to replace sys.set_coroutine_wrapper() API, which is, btw, deprecated in 3.7.

1st1 · 2018-05-18T23:31:22Z

Doc/whatsnew/3.7.rst

+
+C API Changes
+=============
+


Need to mention PEP 567 related changes to the C API: https://www.python.org/dev/peps/pep-0567/#c-api

These C APIs seem to be undocumented, BTW.

I will add an entry together with the documentation for PEP 567 C API in a separate PR.

ilevkivskyi · 2018-05-18T23:44:45Z

Doc/whatsnew/3.7.rst

+dynamically as specified by :pep:`560`.
+(Contributed by Ivan Levkivskyi in :issue:`32717`.)
+
+


There are some perf improvements in typing module due to PEP 560, I would probably mention 6-7x faster import of typing, but there are other perf improvements. They are all in the same PR #4906

Perf improvements should be in the "Optimizations" section.

bbayles · 2018-05-19T01:52:14Z

Doc/whatsnew/3.7.rst

+ipaddress
+---------
+
+The new ``subnet_of`` and ``superset_of`` methods of


I think this ought to be supernet_of rather than superset_of, yes?

bbayles · 2018-05-19T01:55:58Z

Doc/whatsnew/3.7.rst

+
+:class:`~zipfile.ZipFile` now accepts the new *compresslevel* parameter to
+control the compression level.
+(Contributed by Stephen J. Turnbull. in :issue:`28102`.)


This references the incorrect issue - it was actually 21417, and the attribution is also scrambled.

Fixed. Thanks for catching this.

1st1 · 2018-05-19T03:22:03Z

Also, I'd move PEP 545 description to the top of the document (where we have all other PEPs described).

@ned-deily It's a long weekend in Canada, so Elvis will be able to fix the comments here on Tuesday and I'll merge the PR.

carljm

Rendered the updated whatsnew in Sphinx and read through it in a browser. Looks good, modulo comments.

What a great Python release!

carljm · 2018-05-19T05:04:18Z

Doc/whatsnew/3.7.rst

@@ -53,20 +55,123 @@ For full details, see the :ref:`changelog <changelog>`.
   so it's worth checking back even after reading earlier versions.


Is this PR not the time to remove this "prerelease users" note yet?

carljm · 2018-05-19T05:11:50Z

Doc/whatsnew/3.7.rst

-that would act like a normal :meth:`__getattr__` method, except that it will
-be defined on module *instances*.
+Python 3.7 allows defining :meth:`__getattr__` on modules and will call
+it whenever a module attribute is accessed.


Slightly misleading; like object __getattr__, module __getattr__ is called only if the attribute is otherwise not found. So it's not called "whenever a module attribute is accessed." Suggested re-word: "...and will call it whenever a module attribute is otherwise not found."

carljm · 2018-05-19T05:17:41Z

Doc/whatsnew/3.7.rst

+method now supports now supports passing a :term:`path-like objects <path-like object>`
+as a socket path.  (Contributed by Yury Selivanov in :issue:`32066`.)
+
+The :func:`asyncio.windows_utils.socketpair`` function has been removed,


extra backtick after windows_utils.socketpair

carljm · 2018-05-19T05:26:57Z

Doc/whatsnew/3.7.rst

-:func:`~contextlib.asynccontextmanager` and
-:class:`~contextlib.AbstractAsyncContextManager` have been added. (Contributed
+The new :func:`~contextlib.nullcontext` is a simpler and faster alternative
+to :class:`~contextlib.ExitStack` for handling simple context managers.


I don't think that "handling simple context managers" is an accurate description of what nullcontext is good for. It won't help you handle or implement a simple context manager, it's useful only as a no-op context managers. Suggested re-word: "... a simpler and faster no-op context manager than ExitStack."

carljm · 2018-05-19T05:27:42Z

Doc/whatsnew/3.7.rst

 Alexander Mohr and Ilya Kulakov in :issue:`29302`.)

+
 cProfile
 --------

 :mod:`cProfile` command line now accepts ``-m module_name`` as an alternative


Nit: for consistency with other entries, this sentence should start with "The ".

carljm · 2018-05-19T05:50:42Z

Doc/whatsnew/3.7.rst

+now accepts negative *limit*, truncating the result to the
+``abs(limit)`` oldest frames.  To get the old behaviour, use
+the new *most_recent_first* argument to ``Traceback.format()``.
+(Contributed by Jesse Bakker. in :issue:`32121`.)


Remove period after "Bakker"

carljm · 2018-05-19T05:52:03Z

Doc/whatsnew/3.7.rst

@@ -837,6 +1630,13 @@ The initialization of the default warnings filters has changed as follows:
 (Contributed by Nick Coghlan and Victor Stinner in :issue:`20361`,
 :issue:`32043`, and :issue:`32230`)


Need a period to end this sentence

carljm · 2018-05-19T05:54:35Z

Doc/whatsnew/3.7.rst

+avoid creating bound method instances.
+(Contributed by Yury Selivanov and INADA Naoki in :issue:`26110`.)
+
+The :mod:`asyncio` received a number of notable optimizations for commonly


missing the word "module"

carljm · 2018-05-19T05:55:04Z

Doc/whatsnew/3.7.rst

+:meth:`dict.copy` is now up to 5.5 times faster.
+(Contributed by Yury Selivanov in :issue:`31179`.)
+
+:func:`hasattr` and :func:`getattr`` are now about 4 times faster when


extra backtick after getattr

carljm · 2018-05-19T05:56:28Z

Doc/whatsnew/3.7.rst

+The :meth:`~selectors.BaseSelector.modify` methods of classes
+:class:`selectors.EpollSelector`, :class:`selectors.PollSelector`
+and :class:`selectors.DevpollSelector` may be around 10% faster under
+heavy loads.  (Contributed by Giampaolo Rodola' in :issue:`30014`)


use correct accent in contributor's name instead of trailing apostrophe?

This is how Giampaolo is spelling his name.

willingc

@elprans Great work on the What's New. ☀️ I've left a few comments about typos and spelling.

willingc · 2018-05-19T06:06:11Z

Doc/whatsnew/3.7.rst

+an easy and consistent way to enter the Python debugger.
+
+Built-in ``breakpoint()`` calls :func:`sys.breakpointhook`.  By default, this
+latter imports :mod:`pdb` and then calls ``pdb.set_trace()``, but by binding


willingc · 2018-05-19T06:07:59Z

Doc/whatsnew/3.7.rst

+Python 3.7 allows defining :meth:`__getattr__` on modules and will call
+it whenever a module attribute is accessed.
+
+A typical example of where this may be useful is attribute access deprecation


... attribute access, deprecation...

This actually means "deprecation warnings for attribute access". I'll reword for clarity.

willingc · 2018-05-19T06:10:32Z

Doc/whatsnew/3.7.rst

-provide nanosecond resolution: they return a number of nanoseconds as a Python
-``int``.
+The new functions are similar in function to the existing functions
+without the ``_ns`` suffix, but instead of returning fractional seconds,


...suffix. They differ by returning nanoseconds as integers, instead of fractional seconds.

willingc · 2018-05-19T06:15:43Z

Doc/whatsnew/3.7.rst

+abc
+---
+
+Improved conistency of exceptions raised by :func:`issubclass` when called


sp./ consistency

willingc · 2018-05-19T06:17:58Z

Doc/whatsnew/3.7.rst

-the user intermix options and positional arguments on the command line,
-as is possible in many unix commands.  It supports most but not all
-argparse features.  (Contributed by paul.j3 in :issue:`14191`.)
+The new :meth:`ArgumentParser.parse_intermixed_args() \


As done on line 589, it probably is more readable to avoid using \ and put the method link all on one line.

If it is just the backslash that worries you, perhaps you can remove it but leave the line break in

willingc · 2018-05-19T06:26:25Z

Doc/whatsnew/3.7.rst

+
+Support for directly ``await``-ing instances of :class:`asyncio.Lock` and
+other asyncio synchronization primitives has been deprecated.  An
+asynchronous context manager has to be used in order to acquire and release


~~has to~~ must

willingc · 2018-05-19T06:34:06Z

Doc/whatsnew/3.7.rst

+The new :func:`gc.freeze`` function allows freezing all objects tracked
+by the garbage collector and exclude them from future collections.
+This can be used before a POSIX ``fork()`` call to make the GC copy-on-write
+friendly or to speed up collection.  The new :func:`gc.unfreeze` functions


Looks like there is an extra period before "The new..."

willingc · 2018-05-19T06:56:54Z

Doc/whatsnew/3.7.rst

+===========================================
+
+Function :c:func:`PySlice_GetIndicesEx` is deprecated and replaced with
+a macro if ``Py_LIMITED_API`` is not set or set to the value between


... is not set, is set to a value in the range between 0x03050400 and 0x03060000 (not inclusive), or is 0x03060100 or higher.

willingc · 2018-05-19T06:58:27Z

Doc/whatsnew/3.7.rst

+
+* Direct instantiation of :class:`ssl.SSLSocket` and :class:`ssl.SSLObject`
+  objects is now prohibited. The constructors were never documented, tested,
+  or designed as public constructors. Users were suppose to use


sp. supposed

willingc · 2018-05-19T07:01:09Z

Doc/whatsnew/3.7.rst


-* ``repr`` for :exc:`BaseException` has changed not to include trailing comma
-  in the output. Mind that most exceptions are affected by this change.
+* ``repr`` for :exc:`BaseException` has changed not to include the trailing


~~has changed not to include~~ has changed to exclude

serhiy-storchaka

A have reviewed a small part of the PR and have many comments. Many changes should be in other sections, more brief, or not be added at all.

serhiy-storchaka · 2018-05-19T07:44:19Z

Doc/whatsnew/3.7.rst

+abc
+---
+
+Improved conistency of exceptions raised by :func:`issubclass` when called


I think this is not worth an entry in this section. If you with, it can be added in the section "Other Language Changes" or like.

serhiy-storchaka · 2018-05-19T07:44:53Z

Doc/whatsnew/3.7.rst

+aifc
+----
+
+Improved exceptions raised for invalid number of channels and sample


This change is not worth be mentioned in this document at all.

serhiy-storchaka · 2018-05-19T07:48:29Z

Doc/whatsnew/3.7.rst

+intermixing options and positional arguments.
+(Contributed by paul.j3 in :issue:`14191`.)
+
+:mod:`argparse` subparsers are now required by default.  This matches the


This will better look in the section "Changes in the Python API".

serhiy-storchaka · 2018-05-19T07:51:18Z

Doc/whatsnew/3.7.rst

+ast
+---
+
+:meth:`ast.literal_eval()` is now more strict.  Addition and subtraction of


This will better look in the section "Changes in the Python API". But I'm not sure it is worth mentioning in this document at all. This change makes the implementation closer to the documentation and will affect too little code that use literal_eval() in non-purposed way.

serhiy-storchaka · 2018-05-19T07:54:10Z

Doc/whatsnew/3.7.rst

+asyncio
+-------
+
+The new :term:`provisional <provisional api>` :func:`asyncio.run` function can


This is a very long list. It would be better to write it in more compact form. See how long list of functions were introduced in previous releases.

Well, asyncio always have a lot of stuff in what's new. Every change here is important, please keep the section as is.

serhiy-storchaka · 2018-05-19T08:16:24Z

Doc/whatsnew/3.7.rst

+dictionary are now being implicitly converted to strings.
+(Contributed by James Tocknell. in :issue:`23835`.)
+
+The *filename* parameter of the


Not in this section.

serhiy-storchaka · 2018-05-19T08:19:14Z

Doc/whatsnew/3.7.rst

+csv
+---
+
+A single empty field is now always quoted when written into a CSV


Not in this section.

serhiy-storchaka · 2018-05-19T08:22:38Z

Doc/whatsnew/3.7.rst

+dbm
+---
+
+:func:`dbm.dumb` now supports reading read-only files and no longer writes the


Not in this section. This should be in two sections: a part in minor enhancements, and a part in deprecations.

:mod:`dbm.dumb` instead of :func:`dbm.dumb`.

serhiy-storchaka · 2018-05-19T08:23:16Z

Doc/whatsnew/3.7.rst

+(Contributed by Serhiy Storchaka in :issue:`28847`.)
+
+
+:func:`dbm.open` now encodes filenames using the filesystem encoding rather


This is a bug fix.

serhiy-storchaka · 2018-05-19T08:23:38Z

Doc/whatsnew/3.7.rst

+the new *depth* parameter.
+(Contributed by Serhiy Storchaka in :issue:`11822`.)
+
+Functions in the :mod:`dis` module now works with asynchronous generators and


Not in this section.

vadmium · 2018-05-19T03:17:08Z

Doc/whatsnew/3.7.rst

 (Contributed by Paul Ganssle in :issue:`15873`.)

+The :class:`tzinfo <datetime.tzinfo>` objects now supprt sub-minute offsets.


Maybe it would be more correct to say “The tzinfo class now supports . . .”, unless there are only specific instances with the support.

vadmium · 2018-05-19T03:18:49Z

Doc/whatsnew/3.7.rst

+:meth:`Future.get_loop() <asyncio.Future.get_loop>` methods
+return the instance of the loop on which a task or a future were created.
+(Contributed by Yury Selivanov in :issue:`32415`.)
+:meth:`Server.get_loop() <asyncio.Server.get_loop>` allows to do the same for


As far as I know allows to is awkward and incorrect grammar in English, though it crops up often. In this case I suggest something like “Server.get_loop allows doing the same . . .”, or even “Server.get_loop does the same”.

vadmium · 2018-05-19T03:21:11Z

Doc/whatsnew/3.7.rst

+---
+
+A single empty field is now always quoted when written into a CSV
+file.  This allows to distinguish an empty row from a row consisting of


This allows distinguishing . . .

vadmium · 2018-05-19T03:24:37Z

Doc/whatsnew/3.7.rst

+--
+
+The new :func:`gc.freeze`` function allows freezing all objects tracked
+by the garbage collector and exclude them from future collections.


. . . and excluding them . . .

vadmium · 2018-05-19T03:37:56Z

Doc/whatsnew/3.7.rst

-   The improved host name check requires an OpenSSL 1.0.2 or 1.1 compatible
-   libssl. OpenSSL 0.9.8 and 1.0.1 are no longer supported. LibreSSL is
+   The improved host name check requires OpenSSL 1.0.2 or 1.1 compatible
+   libssl.  OpenSSL 0.9.8 and 1.0.1 are no longer supported.  LibreSSL is


I think it makes more sense with “an” than without. Another way of writing it could be “. . . requires a libssl implementation compatible with Open SSL’s 1.0.2 or 1.1 APIs.”

But after reading further down, it says one of these versions is required to build (and presumably use) the SSL module in general. Which seems to make this paragraph redundant.

vadmium · 2018-05-19T03:56:22Z

Doc/whatsnew/3.7.rst

+:func:`timeit.timeit` autorange now uses a single loop iteration if
+the benchmark takes less than 10 seconds, instead of 10 iterations.
+``python3 -m timeit -s 'import time' 'time.sleep(1)'`` now takes 4 seconds
+instead of 40 seconds.


This rule doesn’t make sense. Maybe it should be to use a single iteration if a benchmark is slower than some value (0.2 s?). Does running sleep(11) also do a single iteration?

vadmium · 2018-05-19T06:56:43Z

Doc/whatsnew/3.7.rst

+method can be used to replace the logger stream after handler creation.
+(Contributed by Vinay Sajip in :issue:`30522`.)
+
+It is now possible to specify keyword arguments to handler constructor in


Perhaps this should say handler constructors?

vadmium · 2018-05-19T08:55:57Z

Doc/whatsnew/3.7.rst

+
+The :func:`math.erf` and :func:`math.erfc` functions now use the (faster)
+C library implementation on most platforms.
+(Contributed by Serhiy Storchaka in :issue:`26121`.)


Duplicate entry, although I prefer your version :)

vadmium · 2018-05-19T09:19:43Z

Doc/whatsnew/3.7.rst

-* Due to internal changes in :mod:`socket` you won't be able to
-  :func:`socket.fromshare` a socket :func:`~socket.socket.share`-ed in older
-  Python versions.
+* Due to internal changes in :mod:`socket` calling :func:`socket.fromshare`


Needs a comma: “Due to changes in socket, calling fromshare . . .”

1st1 · 2018-05-19T13:04:22Z

@serhiy-storchaka @ned-deily Serhiy, I disagree with how you want to rewrite the asyncio section. As asyncio maintainer I like the what's new in 3.7 asyncio section, and find it similar in its size to 3.6 and 3.5. A lot of changes that you marked as "not in this section" or "not worth mentioning" are actually very subtle and must be communicated to asyncio users. I'm +1 to drop docs for aifc modules and alike as you suggested, but the asyncio section should stay as is in my opinion.

ned-deily · 2018-05-19T13:29:38Z

@1st1 @serhiy-storchaka Regarding the asyncio section, in general I think the level of detail is OK. There are a few entries that perhaps could be shortened a bit and the section would benefit by turning the items into a bullet list as was done in the 3.6 docs. I think it's better to have too much detail here than too little. Let's go with @1st1's instincts and see how it all looks after @elprans does further editing. We can always remove or add stuff later if necessary.

serhiy-storchaka · 2018-05-19T15:47:17Z

If other core developer (especially the maintainer of the module) want to include some information I have no objections.

elprans · 2018-05-19T16:33:31Z

Thanks everyone for the reviews! I updated the PR with fixes.

serhiy-storchaka · 2018-05-19T15:50:10Z

Doc/whatsnew/3.7.rst

+equivalent to CR.
+(Contributed by Bo Bayles in :issue:`32304`.)
+
+The unused ``install_misc`` command has been removed.


In the "Removed" section.

serhiy-storchaka · 2018-05-19T15:50:42Z

Doc/whatsnew/3.7.rst

+and ``platforms`` fields are not specified as a list or a string.
+(Contributed by Berker Peksag in :issue:`19610`.)
+
+The ``upload`` command no longer tries to change CR end-of-line characters


This is a bug fix.

serhiy-storchaka · 2018-05-19T15:55:22Z

Doc/whatsnew/3.7.rst

+
+:class:`distutils.core.setup` now warns if the ``classifiers``, ``keywords``
+and ``platforms`` fields are not specified as a list or a string.
+(Contributed by Berker Peksag in :issue:`19610`.)


and Neil Schemenauer

serhiy-storchaka · 2018-05-19T16:03:20Z

Doc/whatsnew/3.7.rst

+therefore included in source distributions.
+(Contributed by Ryan Gonzalez in :issue:`11913`.)
+
+:class:`distutils.core.setup` now warns if the ``classifiers``, ``keywords``


In "Changes in the Python API".

serhiy-storchaka · 2018-05-19T16:06:48Z

Doc/whatsnew/3.7.rst

+enum members.
+(Contributed by Ethan Furman in :issue:`31801`.)
+
+In Python 3.8, attempting to check for non-Enum objects in :class:`Enum`


In "Deprecated"

serhiy-storchaka · 2018-05-19T16:26:22Z

Doc/whatsnew/3.7.rst

@@ -652,103 +1288,189 @@ can be set within the scope of a group.
 ``'^$'`` or ``(?=-)`` that matches an empty string.
 (Contributed by Serhiy Storchaka in :issue:`25054`.)

+Regular expressions compiled with the :const:`re.LOCALE` flag no longer


"Changes in the Python API"

serhiy-storchaka · 2018-05-19T16:28:43Z

Doc/whatsnew/3.7.rst

+when the compiled regular expression is used.
+(Contributed by Serhiy Storchaka in :issue:`30215`.)
+
+:exc:`FutureWarning` is now emitted if a regular expression contains


"Changes in the Python API".

serhiy-storchaka · 2018-05-19T16:42:58Z

Doc/whatsnew/3.7.rst

+timeit
+------
+
+:func:`timeit.timeit` autorange now uses a single loop iteration if


Actually many other changes were made in this issue.

serhiy-storchaka · 2018-05-19T16:44:40Z

Doc/whatsnew/3.7.rst

+types
+-----
+
+The new :class:`~types.WrapperDescriptorType`,


Merge the following two entries into a single paragraph.

serhiy-storchaka · 2018-05-19T16:45:45Z

Doc/whatsnew/3.7.rst

+dynamically as specified by :pep:`560`.
+(Contributed by Ivan Levkivskyi in :issue:`32717`.)
+
+


Perf improvements should be in the "Optimizations" section.

bbayles · 2018-05-19T16:54:18Z

Doc/whatsnew/3.7.rst

+
+:class:`~zipfile.ZipFile` now accepts the new *compresslevel* parameter to
+control the compression level.
+(Contributed by Bo Bayle in :issue:`21417`.)


Whoops, looks like my last name is misspelled here?

Ah, sorry. Will fix.

elprans · 2018-05-19T17:23:26Z

@serhiy-storchaka I agree that placing depreciations and removals together makes sense, but what is the value in splitting standard library improvements into “New” and “Changed”?

serhiy-storchaka · 2018-05-19T16:55:01Z

Doc/whatsnew/3.7.rst



 binascii
 --------

 The :func:`~binascii.b2a_uu` function now accepts an optional *backtick*
-keyword argument.  When it's true, zeros are represented by ``'`'``
+keyword argument.  When it's ``True``, zeros are represented by ``'`'``


Please keep "true". Any boolean true value switches on this feature, not just literal True.

It is common to write "if is true" for input arguments, but "return True" for output value.

serhiy-storchaka · 2018-05-19T17:08:16Z

Doc/whatsnew/3.7.rst

+* The unused :mod:`distutils ```install_misc`` command has been removed.
+  (Contributed by Eric N. Vander Weele in :issue:`29218`.)
+
+* The ``fpectl`` module has been removed.  It was never enabled by


It may be worth to add a separate section for this. Seems this is the first module removal since 3.0.

serhiy-storchaka · 2018-05-19T17:52:55Z

Doc/whatsnew/3.7.rst

+C library implementation on most platforms.
+(Contributed by Serhiy Storchaka in :issue:`26121`.)
+
+Many functions and methods in the :mod:`abc` module have been reimplemented


This change are already documented above.

serhiy-storchaka · 2018-05-19T17:58:32Z

Doc/whatsnew/3.7.rst

+(Contributed by Jelle Zijlstra with further improvements by INADA Naoki,
+Serhiy Storchaka, and Raymond Hettinger in :issue:`28638`.)
+
+:meth:`date.fromordinal` and :meth:`date.fromtimestamp` are now up to


I suggest: Alternate :class:`datetime.date` constructors :meth:`~datetime.date.fromordinal` and :meth:`~datetime.date.fromtimestamp` ...

serhiy-storchaka · 2018-05-19T18:02:59Z

Doc/whatsnew/3.7.rst

-
-* Trace hooks may now opt in to receiving ``opcode`` events from the interpreter
-  by setting the new ``f_trace_opcodes`` attribute to :const:`True` on the frame
+* Trace hooks may now opt out of receiving the ``line`` and ``opcode`` events


"line" may now be opted out, but "opcode" may now be opted in.

serhiy-storchaka · 2018-05-19T18:13:47Z

Doc/whatsnew/3.7.rst

- Using non-integer value for selecting a plural form in :mod:`gettext` is
-  now deprecated.  It never correctly worked. (Contributed by Serhiy Storchaka
-  in :issue:`28692`.)
+aifc


Is it worth to add separate subsections for modules? Most contain only a single entry. This makes the text less compact.

serhiy-storchaka · 2018-05-19T18:17:58Z

Doc/whatsnew/3.7.rst

@@ -1157,6 +2077,13 @@ Changes in Python behavior
  exceptions.
  (Contributed by Yury Selivanov in :issue:`32670`.)

+* :keyword:`async` and :keyword:`await` names are now reserved keywords.


Please move this at the first place, as it is the most visible breaking change. Would be nice to add recommendations for solving problems caused by this change.

serhiy-storchaka · 2018-05-19T18:20:44Z

Doc/whatsnew/3.7.rst

@@ -1214,7 +2137,7 @@ Changes in the Python API
  (Contributed by Serhiy Storchaka in :issue:`29192`.)

 * ``Module``, ``FunctionDef``, ``AsyncFunctionDef``, and


I suggest to move this entry to the beginning of the sections, as it breaks several third-party projects.

serhiy-storchaka · 2018-05-19T18:22:07Z

Doc/whatsnew/3.7.rst

@@ -1289,6 +2213,10 @@ Changes in the Python API
  of escaping all characters other than ASCII letters, numbers, and ``'_'``.
  (Contributed by Serhiy Storchaka in :issue:`29995`.)

+* :func:`os.listdir` and :func:`os.scandir` now emit file name as bytes when


This is a bug fix.

serhiy-storchaka · 2018-05-19T18:24:59Z

Doc/whatsnew/3.7.rst

+  dictionary are now being implicitly converted to strings.
+  (Contributed by James Tocknell in :issue:`23835`.)
+
+* :mod:`csv` will now always quote empty fields.  This allows distinguishing


This is a bug fix.

serhiy-storchaka · 2018-05-19T18:38:48Z

New things can not break you code. Changed things can break the code that depends on the old behavior. Sometimes the boundary is not well determined.

Remove redundant content, reclassify some entries, fix typos and address other review feedback.

elprans · 2018-05-20T03:11:13Z

Changed things can break the code

Yes, and this is why we have the "Porting to Python 3.7" section, which begins with:

"This section lists previously described changes and other bugfixes
that may require changes to your code."

(emphasis mine)

I think optimizing the What's New document for readability is more important than trying to make it short. Thus, repeating change descriptions in different sections is OK.

miss-islington · 2018-05-20T03:15:09Z

Thanks @elprans for the PR, and @1st1 for merging it 🌮🎉.. I'm working now to backport this PR to: 3.7.
🐍🍒⛏🤖

miss-islington · 2018-05-20T03:16:17Z

Sorry, @elprans and @1st1, I could not cleanly backport this to 3.7 due to a conflict.
Please backport using cherry_picker on command line.
cherry_picker 63536bd286097e770909052052a21804a5e09b66 3.7

1st1 · 2018-05-20T03:16:27Z

The PR is merged now. All edits and suggestions will be addressed in separate PRs. Thanks, Elvis!

(cherry picked from commit 63536bd) Co-authored-by: Elvis Pranskevichus <elvis@magic.io>

bedevere-bot · 2018-05-20T03:34:08Z

GH-6998 is a backport of this pull request to the 3.7 branch.

…6998) (cherry picked from commit 63536bd) Co-authored-by: Elvis Pranskevichus <elvis@magic.io>

bpo-32996: The bulk of What's New in Python 3.7

18f45b9

This represents the bulk of the content for What's New in Python 3.7.

elprans requested review from 1st1 and asvetlov as code owners May 18, 2018 22:40

the-knights-who-say-ni added the CLA signed label May 18, 2018

bedevere-bot added the awaiting review label May 18, 2018

1st1 added skip news needs backport to 3.7 labels May 18, 2018

1st1 removed the request for review from asvetlov May 18, 2018 22:41

1st1 self-assigned this May 18, 2018

1st1 reviewed May 18, 2018

View reviewed changes

ilevkivskyi reviewed May 18, 2018

View reviewed changes

bbayles reviewed May 19, 2018

View reviewed changes

carljm reviewed May 19, 2018

View reviewed changes

willingc reviewed May 19, 2018

View reviewed changes

serhiy-storchaka reviewed May 19, 2018

View reviewed changes

vadmium reviewed May 19, 2018

View reviewed changes

elprans force-pushed the whatsnew-37 branch from d82b087 to ceeafce Compare May 19, 2018 16:38

serhiy-storchaka reviewed May 19, 2018

View reviewed changes

bbayles reviewed May 19, 2018

View reviewed changes

serhiy-storchaka reviewed May 19, 2018

View reviewed changes

bpo-32996: More What's New in Python 3.7 edits

4c08eac

Remove redundant content, reclassify some entries, fix typos and address other review feedback.

elprans force-pushed the whatsnew-37 branch from ceeafce to 4c08eac Compare May 20, 2018 03:00

1st1 merged commit 63536bd into python:master May 20, 2018

bedevere-bot removed the awaiting review label May 20, 2018

elprans added a commit to elprans/cpython that referenced this pull request May 20, 2018

[3.7] bpo-32996: The bulk of What's New in Python 3.7 (pythonGH-6978).

6c84583

(cherry picked from commit 63536bd) Co-authored-by: Elvis Pranskevichus <elvis@magic.io>

bedevere-bot removed the needs backport to 3.7 label May 20, 2018

1st1 pushed a commit that referenced this pull request May 20, 2018

[3.7] bpo-32996: The bulk of What's New in Python 3.7 (GH-6978). (GH-…

15f3d0c

…6998) (cherry picked from commit 63536bd) Co-authored-by: Elvis Pranskevichus <elvis@magic.io>

hroncok mentioned this pull request May 23, 2018

bpo-33514: Document PEP 492 changes in 3.7 changelog #6857

Closed

		dynamically as specified by :pep:`560`.
		(Contributed by Ivan Levkivskyi in :issue:`32717`.)

		@@ -53,20 +55,123 @@ For full details, see the :ref:`changelog <changelog>`.
		so it's worth checking back even after reading earlier versions.

		@@ -837,6 +1630,13 @@ The initialization of the default warnings filters has changed as follows:
		(Contributed by Nick Coghlan and Victor Stinner in :issue:`20361`,
		:issue:`32043`, and :issue:`32230`)

		(Contributed by Serhiy Storchaka in :issue:`28847`.)


		:func:`dbm.open` now encodes filenames using the filesystem encoding rather

		(Contributed by Paul Ganssle in :issue:`15873`.)

		The :class:`tzinfo <datetime.tzinfo>` objects now supprt sub-minute offsets.

		@@ -1214,7 +2137,7 @@ Changes in the Python API
		(Contributed by Serhiy Storchaka in :issue:`29192`.)

		* ``Module``, ``FunctionDef``, ``AsyncFunctionDef``, and

bpo-32996: The bulk of What's New in Python 3.7 #6978

bpo-32996: The bulk of What's New in Python 3.7 #6978

Conversation

elprans commented May 18, 2018 • edited by bedevere-bot Loading

1st1 commented May 18, 2018

ned-deily commented May 18, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

1st1 commented May 19, 2018

carljm left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

willingc left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

serhiy-storchaka left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

1st1 commented May 19, 2018 • edited Loading

ned-deily commented May 19, 2018

serhiy-storchaka commented May 19, 2018

elprans commented May 19, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

elprans commented May 18, 2018 •

edited by bedevere-bot

Loading

1st1 commented May 19, 2018 •

edited

Loading