Support Python 3.11 #7143
Labels
component: infrastructure
Issues related to development scripts, CI, servers, etc.
priority: 0 - high
Issues which are currently the primary focus.
Comments
The-Compiler
added
component: infrastructure
The-Compiler
added a commit
that referenced
this issue
Apr 28, 2022
Still reflected in commit status See #7143
twigleingrid
pushed a commit
to twigleingrid/qutebrowser
that referenced
this issue
May 13, 2022
Still reflected in commit status See qutebrowser#7143
|
Gave this another try with the 3.11 branch (i.e. after Python 3.11.0b1) on
And reported/reproduced some upstream bugs:
CPython patches: From ea116aaa1c9904149befdd8a45a985d550cad2aa Mon Sep 17 00:00:00 2001
From: Florian Bruhin <me@the-compiler.org>
Date: Thu, 26 May 2022 14:40:23 +0200
Subject: [PATCH] Disable broken asserts
---
Python/compile.c | 2 +-
Python/pystate.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/Python/compile.c b/Python/compile.c
index cc0d76e038..42507532a5 100644
--- a/Python/compile.c
+++ b/Python/compile.c
@@ -7531,7 +7531,7 @@ write_location_info_long_form(struct assembler* a, struct instr* i, int length)
assert(length > 0 && length <= 8);
write_location_first_byte(a, PY_CODE_LOCATION_INFO_LONG, length);
write_location_signed_varint(a, i->i_lineno - a->a_lineno);
- assert(i->i_end_lineno >= i->i_lineno);
+ //assert(i->i_end_lineno >= i->i_lineno);
write_location_varint(a, i->i_end_lineno - i->i_lineno);
write_location_varint(a, i->i_col_offset+1);
write_location_varint(a, i->i_end_col_offset+1);
diff --git a/Python/pystate.c b/Python/pystate.c
index 3e28a6ab69..c801cdcd8b 100644
--- a/Python/pystate.c
+++ b/Python/pystate.c
@@ -2215,7 +2215,7 @@ _PyThreadState_PopFrame(PyThreadState *tstate, _PyInterpreterFrame * frame)
}
else {
assert(tstate->datastack_top);
- assert(tstate->datastack_top >= base);
+ //assert(tstate->datastack_top >= base);
tstate->datastack_top = base;
}
}
--
2.36.1
This all results in: when running: It seems like mostly timeout issues (perhaps due to the thread issue above?) and this: Should probably retry with a non-debug Python and the |
|
Tried with a release build of Python 3.11b2 with python/cpython#93397 on top. The only remaining blocker with that is python/cpython#93252 I think. |
|
With python/cpython#94708 and python/cpython#93895 I believe we can finally get back to this once 3.11.0b4 is released. Not sure what the exact plan is for that, see Mailman 3 [Release] Status of Python 3.11 release - Python-Dev - python.org. It looks like there are some new failing tests though, due to some kind of coverage/tracing changes: and others: Also, the "Bad file descriptor" issue in cheroot still seems to be around: in various somewhat random tests: |
|
3.11.0b4 is out now. The coverage issue is nedbat/coveragepy#1419 which has been fixed in nedbat/coveragepy@0d64498 - and indeed our tests work fine with that. That means only the edit: Indeed, also happens with a Python 3.10 built with Probably time to re-enable Python 3.11 on CI and close this! 🎉 (at least when there has been a coverage.py release) |
Merged
|
Oh, looks like GitHub Actions is not quite ready yet, probably waiting for actions/python-versions#177. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Python 3.11 is planned for October 3rd, and the first beta will be released soon (May 6th). Thus, I made a first attempt at running the tests on 3.11.0a7:
Unfortunately, I couldn't really get things to work fully. Some of the tests in
tests/end2end/features/test_hints_bdd.pyandtests/end2end/features/test_tabs_bdd.pyseem to be flaky in strange ways. We should probably have a closer look, but for now, I should probably focus on Qt 6 first.The text was updated successfully, but these errors were encountered: