Alignment: deprecate old Transform* shims, migrate all in-tree callers to *JD

[ckemper67]

Reviewer note

This PR is stacked on #2387 and targets that branch as its base. After #2387 merges to master, this branch will be rebased onto master and the
PR base retargeted accordingly.

Summary

Follow-up to #2387. That PR added TransformCelestialToTelescopeJD /
TransformTelescopeToCelestialJD to the math plugin stack and converted the old
TransformCelestialToTelescope / TransformTelescopeToCelestial methods in
MathPluginManagement into compat shims. This PR completes Phase 2 and Phase 3 of
the migration plan described in that PR's follow-up section:

• Phase 2 -- Mark the old shim methods [[deprecated]] in MathPluginManagement.h
  so any caller that has not yet migrated gets a compiler warning.
• Phase 3 -- Thread an explicit JD parameter through
  AlignmentSubsystemForDrivers and update every in-tree driver and example call
  site to use the *JD variants directly.

Changes

libs/alignment/MathPluginManagement.h

Added [[deprecated("Use TransformCelestialToTelescopeJD with an explicit Julian Date")]]
to both old public shim methods. Their implementations are unchanged.

Note: the old pure-virtual methods in MathPlugin.h are intentionally left
un-deprecated. External math plugins must still implement them; their deprecation is
Phase 4 work after the broader ecosystem migrates.

libs/alignment/AlignmentSubsystemForDrivers.h/.cpp

Added double JD = ln_get_julian_from_sys() as a trailing default parameter to all
six public helper methods:

• AddAlignmentEntryEquatorial / AddAlignmentEntryAltAz -- use JD for
  ObservationJulianDate (previously always captured ln_get_julian_from_sys())
• SkyToTelescopeEquatorial / SkyToTelescopeAltAz -- call
  TransformCelestialToTelescopeJD(ra, dec, JD, TDV) instead of the old shim
• TelescopeEquatorialToSky / TelescopeAltAzToSky -- call
  TransformTelescopeToCelestialJD(TDV, ra, dec, JD) instead of the old shim

Existing callers that omit the JD argument see no source change and identical
behavior (the default evaluates ln_get_julian_from_sys() at the call site).
Callers that want simulated time support can now pass an explicit JD.

Driver and example call sites

All remaining in-tree uses of the old MathPluginManagement shims are updated to
call *JD directly with ln_get_julian_from_sys():

File	Sites
drivers/telescope/skywatcherAPIMount.cpp	9
drivers/telescope/astrotrac.cpp	2
drivers/telescope/telescope_simulator.cpp	2
drivers/telescope/temmadriver.cpp	3
drivers/telescope/dsc.cpp	2
examples/tutorial_seven/simple_telescope_simulator.cpp	3

skywatcherAPIMount non-zero offset sites (JulianOffset axis property and bracket
tracking) now compute the absolute JD explicitly
(ln_get_julian_from_sys() + JDOffset) instead of passing a relative offset through
the shim. The JDnow variable used by the bracket tracking else-branch is hoisted
before the if/else so both branches sample the clock at the same instant.

Backward compatibility

No behavior change for any existing caller. All drivers that go through
AlignmentSubsystemForDrivers pick up the default JD argument and continue to
call the system clock as before. Third-party drivers (indi-celestronaux, indi-eqmod,
etc.) that call MathPluginManagement directly will see deprecation warnings on
their next rebuild but continue to function correctly through the shim.

Test results

1/2 Test  #9: test-alignment ...................   Passed    0.50 sec
2/2 Test #10: test-alignment-plugins ...........   Passed    0.36 sec

100% tests passed, 0 tests failed out of 2

Follow-up (not in this PR)

indi-3rdparty driver migration: Third-party drivers that call
TransformCelestialToTelescope / TransformTelescopeToCelestial directly
(notably indi-celestronaux and indi-eqmod) will see deprecation warnings on
their next rebuild. A follow-up PR to indi-3rdparty should update those call
sites to use *JD with explicit ln_get_julian_from_sys(), mirroring the
pattern applied here.

Phase 4 -- Remove: Once external plugins have migrated, the compat layer can be
fully removed: make *JD pure virtual in MathPlugin.h, remove the old pure
virtuals and their shim overrides in built-in plugins, remove the deprecated
MathPluginManagement shim methods. The full sequence is described in the
follow-up section of #2387.

[ckemper67]

How would we fix the drivers in indi-3rdparty? i.e. what is the sequence of events to make a clean transition? If we merge this, then a build with -Werror would fail if it encounters the deprecated methods, right? I think some of our CI pipelines do that.

Merge #2387, then fix 3rdparty, then merge #2388?
Or remove the deprecation from #2388, merge both PRs, then update 3rd-party, then add the deprecation after everything is clean?

[knro]

Thanks. Now the 3rd party drivers needs to be fixed next