bpo-32888: Improve exception message in ast.literal_eval by Rosuav · Pull Request #340 · python/cpython · GitHub

When a non-literal is given to literal_eval, attempt to be more
helpful with the message, rather than calling it 'malformed'.

https://bugs.python.org/issue32888

…#340, python#344, python#348, python#349, python#350)

…#340, python#344, python#348, python#349, python#350) (3.5->3.6)

…#340, python#344, python#348, python#349, python#350) (3.6->3.7)

This looks good, do you mind opening a bugs.python.org ticket and adding a news entry?

When a non-literal is given to literal_eval, attempt to be more helpful with the message, rather than calling it 'malformed'.

Rebased onto current master, which includes shifting where the actual change happens. NEWS entry added, issue number added.

Please add a special test in test_ast.

In what cases this exception is raised? Is it covered by tests?

The easiest way to trigger it is to use something that isn't permitted in literal_eval, but is valid syntax. There's already one such test - proving that f-strings aren't literal_evallable - and I guess there could be more, but since this is a whitelist, the error is basically "anything else, raise". The only difference in this patch is that the message is different for "syntactically valid but not acceptable by literal_eval" and "syntactically invalid".

I meant what input triggers the "malformed node or string" error now?

Calling literal_eval with an AST node that has non-AST children in it. I don't think it's possible (post-patch) to trigger this error by calling literal_eval with a string, but I can't say for sure, so I don't want to say just "malformed node".

Please add "Patch by..."

Should that be done by amending the current commit, or adding an additional commit?

It is better for review to add a new commit. In any case all commits are squashed before merging into the repository.

👍

Overall this looks good, just one suggestion on the test implementation.

A helpful thing to use in between lines 594 and 595 would be with self.subTest(test=test, message=message): and then indent the parts below it. That way each of these end up being tested independently rather than potentially having the items loop have a case fail and cause the rest of the cases to fail to execute.

Check out https://docs.python.org/3/library/unittest.html#distinguishing-test-iterations-using-subtests

A Python core developer has requested some changes be made to your pull request before we can consider merging it. If you could please address their requests along with any other requests in other reviews from core developers that would be appreciated.

Once you have made the requested changes, please leave a comment on this pull request containing the phrase I have made the requested changes; please review again. I will then notify any core developers who have left a review that you're ready for them to take another look at this pull request.

@Rosuav, please address the last change request by @briancurtin and also rebase to fix the conflict. Thanks!

BPO cites another couple of issues, one of which is still open, as dependencies. And it's not really a high priority - it's nothing more than a nice-to-have. There are a lot of barriers to getting patches accepted and it's often not worth the effort unless you REALLY care about the patch.

(Note that you can rebase or do a regular merge, which is often easier! All PRs are squash merged in the end @csabella)

@serhiy-storchaka Should this wait for bpo-32893 per your last message on the ticket or could it be merged soon? (when conflicts are resolved + suggestion from Brian)

A different approach using ast.dump(): PR #16620.

What's the status of this PR?

@Rosuav: Are you still working on this?

A different approach using ast.dump(): PR #16620.

Oh, I just closed it since it could produce very long error message: https://bugs.python.org/issue38396#msg354132

No, I'm not. It was a nice-to-have but between the pushback (no core dev seemed to want it to happen) and the constant need to fix merge conflicts meant that I abandoned this proposal ages ago. If someone else wants to champion this fix, then sure, but I don't have the energy to push everything, and this one wasn't worth it.

@isidentical: Are you interested to take over this change? I'm not the good candidate to review it. Maybe @pablogsal or @serhiy-storchaka.

Sure.

Closing this pull request as it is now GH-17662.

Bumps [celery](https://github.com/celery/celery) from 4.4.3 to 4.4.4. - [Release notes](https://github.com/celery/celery/releases) - [Changelog](https://github.com/celery/celery/blob/master/Changelog.rst) - [Commits](celery/celery@4.4.3...v4.4.4) Signed-off-by: dependabot-preview[bot] <support@dependabot.com> Co-authored-by: dependabot-preview[bot] <27856297+dependabot-preview[bot]@users.noreply.github.com>

v2.1.1 single-rule extension to A10 per supervisor 01:05:42Z 2026-05-12 + pythia python#341 independent-push-window observation. Authored under structural escalation from pythia python#335(c) (POST-A10 empirical class- repeat trigger): 7th feedback_fabrication_detector_verify instance (supervisor self at 22:36Z) was the 1st POST-A10-codification instance. A10.1: when issuing CONFIRMED-FABRICATION verdict (per A10 trigger), verdict post MUST inline (a) verbatim git command output snippet supporting the accusation, (b) explicit HEAD-at-claim disambiguation distinguishing claim-time HEAD from current HEAD when they differ. Closes the python#340(2) grep-scope-mismatch gap that produced both POST-A10 incidents (librarian 21:26Z + supervisor 22:36Z self-violation). Explicit non-coverage: routine verification claims (gate posts, state reports, ABBA results, build-status, push-verify) remain governed by feedback_no_tool_execution_citation (Alex 2026-04-29 directive). A10.1 verbatim-snippet requirement is FABRICATION-VERDICT-CLASS-ONLY, not generic. The two rules coexist via claim-class distinction. Adoption record updated to note v2.1.1 amendment on top of v2.1 352cc1f.

the-knights-who-say-ni added the CLA signed label Feb 27, 2017

akruis pushed a commit to stackless-dev/stackless that referenced this pull request Nov 1, 2017

Issue python#28556 : merge 5 more typing changes from upstream (python…

83ec302

…#340, python#344, python#348, python#349, python#350)

akruis pushed a commit to stackless-dev/stackless that referenced this pull request Nov 1, 2017

Issue python#28556 : merge 5 more typing changes from upstream (python…

043a8bc

…#340, python#344, python#348, python#349, python#350) (3.5->3.6)

akruis pushed a commit to stackless-dev/stackless that referenced this pull request Nov 1, 2017

Issue python#28556 : merge 5 more typing changes from upstream (python…

52f2959

…#340, python#344, python#348, python#349, python#350) (3.6->3.7)

Rosuav changed the title Improve exception message in ast.literal_eval bpo-32888: Improve exception message in ast.literal_eval Feb 20, 2018

Improve exception message in ast.literal_eval

26ba98e

When a non-literal is given to literal_eval, attempt to be more helpful with the message, rather than calling it 'malformed'.

Rosuav force-pushed the literal_eval-exception branch from b9b6e9d to 26ba98e Compare February 20, 2018 17:41

serhiy-storchaka reviewed Feb 20, 2018

View reviewed changes

zware added the awaiting review label Feb 20, 2018

Rosuav added 3 commits February 21, 2018 05:10

Include patch author name in NEWS

08ffc03

Distinguish different causes of error

fb83598

Add test of different forms of exception message

cce0c6c

briancurtin requested changes Jul 3, 2018

View reviewed changes

bedevere-bot added awaiting changes and removed awaiting review labels Jul 3, 2018

isidentical mentioned this pull request Dec 14, 2019

bpo-38396: dump(node) instead of repr(node) on ast.literal_eval #16620

Closed

isidentical mentioned this pull request Dec 19, 2019

gh-77069: add context info to ast.literal_eval error messages #17662

Closed

csabella closed this Jan 12, 2020

                           return node.n
+                      elif isinstance(node, AST):
+                          raise ValueError('%s not allowed in literal' % type(node).__name__)
                       raise ValueError('malformed node or string: ' + repr(node))

		@@ -0,0 +1,2 @@
		Improve wording of error messages from ast.literal_eval, distinguishing
		valid-but-unacceptable nodes from those that are actually malformed.

+                          ast.BinOp(ast.Num(1), ast.Add(), 'oops'): "malformed node or string: 'oops'",
+                      }
+                      for test, message in tests.items():
+                          with self.assertRaises(ValueError) as cm:

bpo-32888: Improve exception message in ast.literal_eval by Rosuav · Pull Request #340 · python/cpython · GitHub

Repository navigation

bpo-32888: Improve exception message in ast.literal_eval - #340

bpo-32888: Improve exception message in ast.literal_eval#340Rosuav wants to merge 4 commits intopython:masterpython/cpython:masterfrom Rosuav:literal_eval-exceptionRosuav/cpython:literal_eval-exceptionCopy head branch name to clipboard

Conversation

Rosuav commented Feb 27, 2017 • edited by bedevere-bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

merwok commented Feb 18, 2018

Uh oh!

Rosuav commented Feb 20, 2018

Uh oh!

serhiy-storchaka left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

briancurtin left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bedevere-bot commented Jul 3, 2018

Uh oh!

csabella commented May 13, 2019

Uh oh!

Rosuav commented May 13, 2019

Uh oh!

merwok commented May 13, 2019

Uh oh!

merwok commented May 13, 2019 • edited by bedevere-bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

vstinner commented Dec 14, 2019

Uh oh!

vstinner commented Dec 14, 2019

Uh oh!

Rosuav commented Dec 14, 2019

Uh oh!

vstinner commented Dec 15, 2019

Uh oh!

isidentical commented Dec 19, 2019

Uh oh!

csabella commented Jan 12, 2020

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

10 participants

bpo-32888: Improve exception message in ast.literal_eval#340
Rosuav wants to merge 4 commits into
python:masterfrom
Rosuav:literal_eval-exception

Rosuav commented Feb 27, 2017 •
edited by bedevere-bot
Loading

merwok commented May 13, 2019 •
edited by bedevere-bot
Loading