Pin seven open issues that no longer reproduce (#263, #347, #362, #424, #415, #550, #164)

[Rafael-SOWNet]

Seven issues that are open in the tracker but do not reproduce on current master.

This PR contains no fix — only tests. Each pins behaviour that already works, so
the issues can be closed without leaving whatever closed them unprotected. Every case
is the reporter's own, copied from the issue.

Issue	Reported as	Today
#263	InnerSimplify hangs on (x^15 - 3^15 - 15·3^14·(x-3))/(x-3)^2	returns, and so does the limit it was reached from
#347	stack overflow on "a and x".Limit("x", "+oo")	returns the limit unevaluated, no crash
#362	(-1/2d-3/2d)t → (-3/2*d - d/2)*t	(-2) * d * t
#424	Piecewise should be compileable	Compile<double, double> works
#415	simplify intervals	[1;5] /\ [3;8] → [3;5], \/ → [1;8]
#550	equation system fails on a simple case	the reporter's 3×3 system solves
#164	(x+1)^2*(x+2-1)^2 leaves sixteen uncollected terms	(1 + x)^4

A few notes on how they are written:

• The two that used to hang assert termination inside a bounded wait, not a
  duration. They cannot go flaky on a slow machine, and a regression fails the run
  instead of hanging CI.
• Equation system solver fails on simple case #550 is checked by substituting the solution back into all three equations, so it
  does not depend on the form the solver hands the roots back in. The 25×26 system from
  the same report is a question of scale and is left open at
  #608.
• Stackoverflow when computing a limit #347 still returns the limit unevaluated, which is the honest answer for a boolean;
  what the issue reported was the crash, and that is what is pinned.

I checked two neighbours that looked like they belonged here and do not, so they
are deliberately absent:

• no answer (null) to limit 1/x + ln(x) #209 — lim x→0+ (1/x + ln(x)) still comes back unevaluated, not +oo. The two
  terms diverge in opposite directions and deciding the sum needs the rates compared.
• Can't evaluate limit of arcsin #333 — "arcsin(x)".Limit("x", "+oo") still comes back unevaluated. The related
  arcsin(x)/x at 0 does give 1, but that is not what was reported.

Testing

UnitTests 3669 passed / 0 failed. The branch is master plus one new test file, so
the numbers are master's own plus these eleven tests.

This is the fourth of four independent PRs — the others are #663, #664 and #665, and
this one touches no file any of them touch.

🤖 Generated with Claude Code

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 79.43%. Comparing base (90c00a8) to head (ac93b26).
⚠️ Report is 40 commits behind head on master.
❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #666      +/-   ##
==========================================
- Coverage   80.99%   79.43%   -1.56%     
==========================================
  Files         155      155              
  Lines       13687    12294    -1393     
  Branches     1957     1981      +24     
==========================================
- Hits        11086     9766    -1320     
+ Misses       1990     1941      -49     
+ Partials      611      587      -24     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

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