{{ message }}
gh-157170: Restore use_env() after test_use_prescr_screen in test_curses - #157171
Merged
serhiy-storchaka merged 2 commits intoSep 9, 2026
Merged
Conversation
…st_curses curses.use_env() sets a process-wide ncurses default, not a property of the screen it is called on. test_use_prescr_screen turned it off and never turned it back on, so every newterm() in the rest of the process took the screen size from terminfo alone. For a terminfo entry with no size, such as "linux" on the CI runners, that makes newterm() return NULL: test_curses failed with 138 errors in the parallel phase of every Ubuntu CI run and only passed on the rerun, which skips ScreenTests.
serhiy-storchaka
left a comment
serhiy-storchaka
left a comment
Member
There was a problem hiding this comment.
Good catch, and thank you for the trace.
Please drop the NEWS entry. This is a fix to unreleased test. It does not need a NEWS entry.
Please simplify the comment, it is too verbose. One addition line would be enough.
Contributor
Author
|
Thanks for making the requested changes! : please review the changes made to this pull request. |
maurycy
added a commit
to maurycy/cpython
that referenced
this pull request
Sep 10, 2026
* main: (158 commits) pythongh-156924: Try reifying lazy imports in `ForwarRef.evaluate()` (python#156940) pythongh-156233: Fix typos and code snippet errors in the docs (python#156475) pythongh-156837: Refer to yield expressions in generator function definitions (pythonGH-156863) pythongh-155292: Skip updating unicodedata with mismatched interpreter (pythonGH-157066) pythongh-157170: Document the scope of global curses settings (pythonGH-157207) pythongh-155966: Correct handling of `math.tanpi` poles (python#155980) pythongh-157170: Restore use_env() after test_use_prescr_screen in test_curses (pythonGH-157171) pythongh-156910: fix deadlock in type_set_abstractmethods under free-threading (python#156948) pythongh-155648: Write the empty and placeholder IDLE tests (python#156260) pythongh-153569: centralize formatted-string state and source spans (python#156484) pythongh-157137: Mark the PEP 820 soft deprecations as 3.15, not `next` (python#157138) pythongh-152433: Use regular LoadLibrary in UWP for Windows system libs (pythonGH-156972) pythongh-121617: Fix Py_CLEAR() memcpy in C++: replace NULL with _Py_NULL (python#157188) pythongh-157135: Fix documentation errors in the `math.atan{2}pi` functions (python#157136) pythongh-121617: Fix Py_CLEAR() in C++: replace NULL with _Py_NULL (python#157067) pythongh-156774: Speed up pdb startup with asyncio guard (python#156775) pythongh-156109: Allow static, non-framework iOS builds (python#156110) pythongh-156780: Emscripten: add missing EM_JS_DEPS (python#156798) pythongh-152936: Make privileged functions available on Android (python#152977) pythongh-123018: Keep the libedit history file header when truncating (pythonGH-157165) ...
clin1234
pushed a commit
to clin1234/cpython
that referenced
this pull request
Sep 12, 2026
…st_curses (pythonGH-157171) curses.use_env() sets a process-wide ncurses setting, not a property of the screen it is called on.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
test_use_prescr_screenturnsuse_envoff to show that anew_prescr()screen can be used to affect state beforeinitscr(), and never turns it back on.curses.use_env()calls ncurses'use_env(), which sets the process-wide default (_nc_prescreen.use_env) that every later screen copies, not a property of the screen that is current at the time. With it off, ncurses sizes a screen from the terminfolines/colscapabilities alone, and thelinuxentry used by the CI runners has none, so everynewterm()for the rest of the process returns NULL. That is the 138 errorstest_cursesreports in the parallel phase of every Ubuntu CI run (all 31 Ubuntu job runs in the last 8 successful runs on main); the sequential rerun passes only because it re-runs the failed tests by name and so never runsScreenTestsfirst. Details, including an ncurses trace of the failingnewterm(), are in the issue.This restores the setting with
addCleanup(). Verified on Ubuntu 25.10 with ncurses 6.5 (the newterm-basedsetUp()needs ncurses 6.5 or newer):TERM=linux ./python -m test test_curses -u curses: 138 errors before, passes after (three consecutive runs).TERM=linux ./python -m test --fast-ci test_curses: no rerun needed after.TERM=xtermstill passes.On macOS the
ScreenTestsare skipped (ncurses 6.0), which is why this was invisible on developer machines.