gh-143055: Implementation of PEP 798 by adqm · Pull Request #143056 · python/cpython · GitHub

This PR is an implementation of PEP 798 (Unpacking in Comprehensions), reflecting the SC's required modification to genexp semantics from the original proposal.

Feedback is welcome! :)

Issue: Implement PEP 798 – Unpacking in Comprehensions #143055

📚 Documentation preview 📚: https://cpython-previews--143056.org.readthedocs.build/

* add error messages for dict unpacking in dict keys/values * change phrasing of existing error messages about using dict unpacking in a listcomp / genexp

Thanks! A few nitpicky comments, I didn't read past Lib/test yet.

Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com>

The C code all looks good. I'll run it through the buildbots to make sure though.

🤖 New build scheduled with the buildbot fleet by @JelleZijlstra for commit 81e86b3 🤖

Results will be shown at:

https://buildbot.python.org/all/#/grid?branch=refs%2Fpull%2F143056%2Fmerge

If you want to schedule another build, you need to add the 🔨 test-with-buildbots label again.

I'm assuming the plan is to document this in 3.15 "What's New" later?

@ZeroIntensity Thanks for the review!

I'm assuming the plan is to document this in 3.15 "What's New" later?

I haven't drafted a blurb for "What's New" yet, but I'm happy to do so. I can do that in a separate PR later, or I don't mind adding it to this one if that's better.

Probably easiest to add the What's New entry here too.

Will try to make a pass over the grammar next week if @lysnikolaou doesn't beat me to it

I had a look at the grammar. The changes look great in general. I left a couple of minor comments, but feel free to merge if the PR is not blocked otherwise.

Maybe @pablogsal wants to give it a look as well?

I'm pretty sure this alternative is never reached, since, in the second pass, it's encountered in expression before we get to if_expression and only in invalid rules otherwise that will overwrite the error either way.

Thanks for taking a look! Yep, I think you're right that we can't reach invalid_if_expression here, good catch. I'll push that change today (along with a what's new entry).

Changing the order here is probably only about error messages, right? If so, have you checked whether this affects parsing performance at all? To be clear, I'm okay to sacrifice a bit of performance for better error messages.

Yep, the reordering here was purely for error reporting. I hadn't done any performance testing before, but I did a quick, unscientific test this morning, doing some quick timings on ast.parse for a few things.

As might be expected, compared to the place I branched from, it looks like this branch is, in general, a little bit faster on comprehensions and a little bit slower on literal lists. It's a little hard to say how this works out on the whole, but the time for parsing the full standard library was always within ±2% every time I ran it. I can look more thoroughly if we're concerned about this, but this little test at least suggests that things haven't blown up on average.

Sample results:

Test	Reason	This Branch	Merge base	Change
[] (1e5 times)	empty list	0.5838	0.5539	~5% slower
[x, y, z, a, b, c] (1e5 times)	well-formed list	1.4051	1.3827	~2% slower
[1, 2, 3, 4, 5, 6 (1e5 times)	broken list	1.1750	1.1413	~3% slower
[x for x in y if z and a] (1e5 times)	well-formed comprehension	1.1642	1.2393	~6% faster
[x for x in (1e5 times)	broken comprehension	0.6568	0.7014	~6% faster
[**x for x in y] (1e5 times)	broken comprehension unpacking (new specific error message)	1.2675	1.0519	~20% slower
[x if x else y] (1e5 times)	good ifexp	1.0163	0.9890	~3% slower
[*x if x else y] (1e5 times)	broken ifexp (new specific error message)	1.3299	1.7870	~26% faster
full stdlib (5 times)		26.1641	26.2642	~equal

Going to "Update branch" to see if it fixes CI

Thanks @JelleZijlstra! Does "update branch" run a different set of tests? Unless I'm missing something, it looks like the test that's currently failing wasn't run on any of the earlier commits. Is that something I should be concerned about?

It's relatively new, so it wasn't on this PR before the update. I think it's an issue with main and not this PR, but we aren't blocked from merging by a failing CIFuzz anyway.

⚠️⚠️⚠️ Buildbot failure ⚠️⚠️⚠️

Hi! The buildbot AMD64 FreeBSD14 3.x (tier-3) has failed when building commit ccbe41e.

What do you need to do:

Don't panic.
Check the buildbot page in the devguide if you don't know what the buildbots are or how they work.
Go to the page of the buildbot that failed (https://buildbot.python.org/#/builders/1232/builds/7885) and take a look at the build logs.
Check if the failure is related to this commit (ccbe41e) or if it is a false positive.
If the failure is related to this commit, please, reflect that on the issue and make a new Pull Request with a fix.

You can take a look at the buildbot page here:

https://buildbot.python.org/#/builders/1232/builds/7885

Summary of the results of the build (if available):

==

Click to see traceback logs

Traceback (most recent call last):
  File "/home/buildbot/buildarea/3.x.opsec-fbsd14/build/Lib/test/support/__init__.py", line 947, in gc_collect
    gc.collect()
    ~~~~~~~~~~^^
ResourceWarning: unclosed <socket.socket fd=8, family=2, type=1, proto=6, laddr=('127.0.0.1', 50174), raddr=('127.0.0.1', 54992)>
Warning -- Unraisable exception
Exception ignored while calling deallocator <function _SelectorTransport.__del__ at 0x1c7f3879be90>:
Traceback (most recent call last):
  File "/home/buildbot/buildarea/3.x.opsec-fbsd14/build/Lib/asyncio/selector_events.py", line 873, in __del__
    _warn(f"unclosed transport {self!r}", ResourceWarning, source=self)
    ~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ResourceWarning: unclosed transport <_SelectorSocketTransport closing fd=8>
k


Traceback (most recent call last):
  File "/home/buildbot/buildarea/3.x.opsec-fbsd14/build/Lib/test/support/__init__.py", line 947, in gc_collect
    gc.collect()
    ~~~~~~~~~~^^
ResourceWarning: unclosed <socket.socket fd=8, family=2, type=1, proto=6, laddr=('127.0.0.1', 60632), raddr=('127.0.0.1', 51955)>
Warning -- Unraisable exception
Exception ignored while calling deallocator <function _SelectorTransport.__del__ at 0x1bd099e02990>:
Traceback (most recent call last):
  File "/home/buildbot/buildarea/3.x.opsec-fbsd14/build/Lib/asyncio/selector_events.py", line 873, in __del__
    _warn(f"unclosed transport {self!r}", ResourceWarning, source=self)
    ~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ResourceWarning: unclosed transport <_SelectorSocketTransport closing fd=8>
k

Looks unrelated

Exciting to see this merged! Thanks for the help, everyone :)

If issues with the code do come up, I'm happy to help fix them.

Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com>

adqm added 8 commits December 20, 2025 18:24

tests for pep798

0288001

initial implementation of pep798 basics

a7ea132

update error messages

c23496e

* add error messages for dict unpacking in dict keys/values * change phrasing of existing error messages about using dict unpacking in a listcomp / genexp

improve error reporting for invalid starred expressions

49405ac

fix for affected test case in test_encodings

26841f9

add news blurb

3ebe2f6

first pass at adding documentation for comprehension unpacking

0713bbc

code formatting

0393552

adqm requested review from AA-Turner, Eclips4, JelleZijlstra, iritkatriel, isidentical, lysnikolaou, markshannon, pablogsal, tomasr8 and willingc as code owners December 21, 2025 23:46

bedevere-app Bot added the awaiting review label Dec 21, 2025

bedevere-app Bot mentioned this pull request Dec 21, 2025

Implement PEP 798 – Unpacking in Comprehensions #143055

Closed

JelleZijlstra reviewed Dec 22, 2025

View reviewed changes

Comment thread Doc/reference/expressions.rst Outdated

Comment thread Doc/reference/expressions.rst Outdated

Comment thread Doc/tutorial/datastructures.rst Outdated

Comment thread Grammar/python.gram Outdated

Comment thread Grammar/python.gram

adqm and others added 2 commits December 22, 2025 14:24

Apply suggestions from code review

fa16f82

Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com>

rearrange dict comprehension reference into a bulleted list

81e86b3

JelleZijlstra approved these changes Dec 22, 2025

View reviewed changes

bedevere-app Bot added awaiting merge and removed awaiting review labels Dec 22, 2025

JelleZijlstra added the 🔨 test-with-buildbots Test PR w/ buildbots; report in status section label Dec 22, 2025

bedevere-bot removed the 🔨 test-with-buildbots Test PR w/ buildbots; report in status section label Dec 22, 2025

ZeroIntensity approved these changes Dec 23, 2025

View reviewed changes

Comment thread Doc/reference/expressions.rst Outdated

Comment thread Doc/reference/expressions.rst Outdated

adqm added 2 commits December 23, 2025 10:32

0 -> zero

74df6b9

use .. versionchanged:: next instead of 3.15

23538a8

docs: fix generator_expression grammar rule

fcf129f

dylwil3 mentioned this pull request Jan 6, 2026

Allow Python 3.15 as valid target-version value in preview astral-sh/ruff#22419

Merged

lysnikolaou approved these changes Jan 28, 2026

View reviewed changes

adqm added 2 commits January 29, 2026 08:29

remove redundant grammar rule

acaf3b1

add "what's new" blurb

b5ae690

Merge branch 'main' into pep798

7c16250

JelleZijlstra merged commit ccbe41e into python:main Jan 31, 2026
48 of 50 checks passed

bedevere-app Bot removed the awaiting merge label Jan 31, 2026

thunder-coding pushed a commit to thunder-coding/cpython that referenced this pull request Feb 15, 2026

pythongh-143055 : Implementation of PEP 798 (python#143056 )

7bfd8a4

Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com>

+              if_expression[expr_ty]:
+                  | a=disjunction 'if' b=disjunction 'else' c=expression { _PyAST_IfExp(b, a, c, EXTRA) }
+                  | invalid_if_expression

+                  | &'(' (genexp | tuple | group)
+                  | &'[' (listcomp | list)
+                  | &'{' (dictcomp | setcomp | dict | set)

gh-143055: Implementation of PEP 798 by adqm · Pull Request #143056 · python/cpython · GitHub

Repository navigation

gh-143055: Implementation of PEP 798 - #143056

gh-143055: Implementation of PEP 798#143056JelleZijlstra merged 16 commits intopython:mainpython/cpython:mainfrom adqm:pep798adqm/cpython:pep798Copy head branch name to clipboard

Conversation

adqm commented Dec 21, 2025 • edited by github-actions Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

JelleZijlstra left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

JelleZijlstra left a comment

Choose a reason for hiding this comment

Uh oh!

bedevere-bot commented Dec 22, 2025

Uh oh!

ZeroIntensity left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

adqm commented Dec 23, 2025

Uh oh!

JelleZijlstra commented Dec 27, 2025

Uh oh!

pablogsal commented Dec 27, 2025

Uh oh!

lysnikolaou left a comment

Choose a reason for hiding this comment

Uh oh!

lysnikolaou Jan 27, 2026

Choose a reason for hiding this comment

Uh oh!

adqm Jan 29, 2026

Choose a reason for hiding this comment

Uh oh!

lysnikolaou Jan 28, 2026

Choose a reason for hiding this comment

Uh oh!

adqm Jan 29, 2026

Choose a reason for hiding this comment

Uh oh!

JelleZijlstra commented Jan 30, 2026

Uh oh!

adqm commented Jan 30, 2026

Uh oh!

ZeroIntensity commented Jan 30, 2026

Uh oh!

Uh oh!

bedevere-bot commented Jan 31, 2026

⚠️⚠️⚠️ Buildbot failure ⚠️⚠️⚠️

Uh oh!

JelleZijlstra commented Jan 31, 2026

Uh oh!

adqm commented Feb 1, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

gh-143055: Implementation of PEP 798#143056
JelleZijlstra merged 16 commits into
python:mainfrom
adqm:pep798

adqm commented Dec 21, 2025 •
edited by github-actions Bot
Loading