A logarithm of x is matched to a logarithm of a multiple of x below the bar

[Rafael-SOWNet]

By parts on li(b x)/x takes the antiderivative of 1/x. With ln(x), the remainder was b ln(x)/ln(b x), which nothing read. ln(b x) is as much an antiderivative of 1/x, and taken so, the logarithm the derivative of li(b x) divides by cancels: li(b x)/x is li(b x) ln(b x) - b x, which is Rubi's answer. The constant of integration is now matched to a logarithm of a multiple of x below the bar, as it is matched to a polynomial divisor.

Measured

Input	Was (2.5.0)	Now
"li(b*x)/x".Integrate("x")	li * b * x + C, with li a variable	li(b * x) * ln(b * x) - b * x + C

• Rubi, master at 16203aa6 and this change on it, run side by side, 0 wrong everywhere:
  • Family 8: 380 → 381 of 420, the one being li(b x)/x.
  • Family 3 at twenty a file: 172 and 173 of 180.
  • The independent suites: 1756 of 1814 on both.
  • Families 1, 2 and 4–7 at five a file: 771 and 770 of 867.
• Run alone, the three that moved are:
  • li(b x)/x, answered here and not on master;
  • 3.2.3 Rename EntitySet -> EntityList #30 and 6.7.1 Let's steal some issues from https://github.com/davidedc/Algebrite! >:3 #180, answered on both in 19 to 24 s. Their side-by-side difference was load.
• The unit tests pass: 14,262, none failed (net10.0). On master with A square of a special function of a shifted argument is asked term by term #1643, the classes this touches pass (106). The native AOT publish of the C++ wrapper builds with no trim or AOT warnings.
• The performance gate passes on c6f7a34c, which is this change on 16203aa6. Allocation matches the baseline on all 19 gated benchmarks.

SpecialFunctionsByPartsTest checks li(b x)/x at positive points, where li(b x) is real.

Part of #1501.

🤖 Generated with Claude Code