Repository navigation
Ask a modulus's argument for its whole limit where its parts say nothing - #2372
Merged
Rafael-SOWNet merged 1 commit intoNov 4, 2026
Merged
Conversation
abs(x * ln(x)) at 0 was NaN, the claim that the limit does not exist, and unevaluated from each side, where it is 0 from both: it is real on both sides, and from the left |x| |ln|x| + i pi| tends to 0 too. The modulus took its argument's limit as the descent reads it, and the descent reads x * ln(x) as 0 * -oo, which is NaN. The same on 2.6.0. Where that reading has no value, the modulus now asks its argument's whole limit, one level deep and outside a limit budget as a quotient does, and takes the modulus of a value. x * ln(x) itself keeps the two-sided convention, and so does x^x; abs(x^x), real on both sides since |x^x| is |x|^x to the left of 0, is 1 there. limitcheck at DEPTH=3, 21 seeds 14-51, master against this: 20 more compared over 8 seeds, 10 NaNs where the values settle answered, each taking the modulus of something that holds a power like x^x at 0 -- abs(x^x), abs(sinh(x^x)), abs(x + 1 - x^x) -- and nothing moved to another value. Suite 18994 passed, AOT and the eight harnesses clean. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #2372 +/- ##
==========================================
- Coverage 86.77% 86.76% -0.01%
==========================================
Files 312 312
Lines 68908 68911 +3
Branches 20081 20082 +1
==========================================
- Hits 59798 59794 -4
- Misses 4912 4917 +5
- Partials 4198 4200 +2
Flags with carried forward coverage won't be shown. Click here to find out more.
🚀 New features to boost your workflow:
|
Rafael-SOWNet
deleted the
fix/a-modulus-asks-its-arguments-whole-limit
branch
November 4, 2026 21:44
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
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.
abs(x * ln(x))at 0 wasNaN, the claim that the limit does not exist, and it was unevaluated from each side. The limit is 0 from both sides. The expression is real on both sides, and from the left |x|·|ln|x| + iπ| tends to 0 too.The modulus took its argument's limit as the descent reads it, and the descent reads
x * ln(x)as 0 · (-oo), which is NaN. 2.6.0 gives the same results.Now, where that reading has no value, the modulus asks its argument's whole limit and takes the modulus of a value. It does this through
AsTheWholeLimit, as a quotient already does: one level deep, and not under a limit budget.On the two-sided convention (#2365):
x * ln(x)andx^xstayNaNat 0, because neither has real values to the left. The tests pin both. A modulus of them does have real values on both sides, since |x^x| is |x|^x to the left of 0. So under the same conventionabs(x^x)has a limit there, and it is now1.abs(x * ln(x))NaN; unevaluated from each side0abs(sin(x) * ln(x))NaN0abs(x^2 * ln(x))NaN0abs(x^x)NaN1Measured
DEPTH=3on 21 seeds (14 to 51), master against this branch: 20 more rows compared, across 8 seeds. 10 "NaN where the values settle" findings are answered. Each takes the modulus of something that holds a power like x^x at 0, for exampleabs(x^x),abs(sinh(x^x))andabs(x + 1 - x^x). No row moved to another value.Found while checking #2371. Recorded in BREAKING-CHANGES.md.
🤖 Generated with Claude Code