[Rafael-SOWNet]

From one side, a product whose factors tend to 0 and to an infinity, one of them a quotient, is now asked as that quotient with the other factor in its dividend, its whole limit once.

Read as a product nothing settled the form, and the rewriting solvers took x * (ln(x) / ln(2 + x)) at 0+ through l'Hôpital's rule for longer than any caller waits. Three samples of the hung probe put the time in ApplylHopitalRuleImpl's Simplify, reached through SolveByRewritingTowardsInfinity and SolveAsDifferenceOfInfinities. x ln(x) / ln(2 + x), the same thing over one divisor, is 0 at once.

Not while the sides of a two-sided limit are compared, as powers already are not (MayReadAPowerThroughItsExponent). Two-sided, x * ln(x) / ln(2 + x) at 0 is NaN, because its left side is complex. Regrouped during the comparison, the same product written with brackets came back 0. Two-sided answers do not move.

Measured on master → this branch, one probe each:

	master	branch
x * log(2 + x, x) at 0+	did not finish in 30 s	0, 0.24 s
x * (ln(x) / ln(2 + x)) at 0+	did not finish in 25 s	0, 0.26 s
x * log(2 + x, x) at 0-	0, 11.8 s	0, 0.26 s
x * (ln(-x) / ln(2 + x)) at 0-	0, 11.7 s	0, 0.25 s
x * log(2 + x, x^3) at 0+	0, 1.15 s	0, 0.26 s
x * log(2 + x, x) at 0, two-sided	NaN	NaN

limitcheck A/B against master (DEPTH=3, the usual seeds): five answers gained, none lost, none wrong.

	master	branch
sin(ln(x) * x) at 0+	unevaluated	0
floor(x * log(pi, x)) at 0+	unevaluated	-1
sgn(sin(x) * log(pi, x)) at 0+	unevaluated	-1
arcsin(-x * log(-1, x)) at 0- and 0+	unevaluated	0

One row moved from a decline to a timeout at the harness's 5 s cap: (cos(x) * sqrt(x))^arctan(tanh(x)) at 0+. Alone it declines on both builds in 7.1–7.7 s, so the move is jitter. Each new answer is in BREAKING-CHANGES.md, measured on 2.6.0, which leaves all of them unevaluated.

The full suite (19,093), the wrapper suites, AOT and the ten harnesses pass.

Found by limitcheck at DEPTH=3, SEED=59, under the logarithm-of-a-power change (next PR). That change writes ln(x^3) / ln(2 + x) as 3 ln(x) / ln(2 + x) and so ran into this. No issue was filed.

🤖 Generated with Claude Code

[codecov]

Codecov Report

❌ Patch coverage is 94.73684% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 86.78%. Comparing base (0f2f41d) to head (5a25201).

Files with missing lines	Patch %	Lines
...nctions/Continuous/Limits/Solvers/Limit.Classes.cs	94.44%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #2384   +/-   ##
=======================================
  Coverage   86.78%   86.78%           
=======================================
  Files         312      312           
  Lines       69110    69129   +19     
  Branches    20156    20159    +3     
=======================================
+ Hits        59978    59995   +17     
- Misses       4916     4917    +1     
- Partials     4216     4217    +1     

Flag	Coverage Δ
cpp	37.77% <ø> (ø)
csharp	87.32% <94.73%> (+<0.01%)	⬆️
fsharp	32.97% <ø> (ø)
interactive	40.72% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
...ns/Continuous/Limits/Solvers/Solvers.Definition.cs	90.37% <100.00%> (+0.04%)	⬆️
...nctions/Continuous/Limits/Solvers/Limit.Classes.cs	74.93% <94.44%> (+0.89%)	⬆️

... and 4 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.