expanded date addition test coverage and added optimized snapshot test

[SATVIKsynopsis]

added updates to #7215

hi i have updated the date addition test suite with a more comprehensive set of cases based on earlier discussion.

the main date_addition_snapshot.rs test now covers a much wider range of dates (iso 2000–2004) so leap years, leap days, and month boundaries are properly executed.

this produces fairly large snapshot files but I believe it gets us closer to the complete coverage we were aiming for.

i also added date_optimized_snapshot.rs, which doesn’t produce snapshots but gives a compact human readable summary of the chosen date and duration ranges.

what i thought was the idea to make it easier to inspect the test configuration without scrolling through thousands of snapshot lines.
if this helper isn’t needed or could be structured better i am totally open to adjusting or removing it. and all the tests passed locally.

i am not fully sure that these changes do exactly what you had in mind, so please feel free to suggest changes. i will update things as said.

happy to revise anything here to better match expectations.

[SATVIKsynopsis]

CI is currently failing due to the size and runtime of the snapshot tests across calendars (macOS and test-gigo appear to fail first, with other jobs cancelled).

I validated the exhaustive date * duration * calendar matrix locally to confirm correctness (especially around leap days and leap months), but I understand committing exhaustive snapshots may not be desirable.

I am happy to reduce this to:

1. a small, human-readable snapshot targeting leap boundaries, or

2. a reduced per-calendar snapshot with a narrow date / duration range

Let me know which direction you would prefer, and I will adjust accordingly.

[SATVIKsynopsis]

i have now kept the exhaustive date × duration × calendar coverage as an ignored test, so it doesn’t affect ci runtime. i validated the full matrix locally to ensure leap days and leap months behave correctly across calendars.

for the pr itself, i am happy to proceed in one of the following ways:

1. commit a small human readable snapshot focused on leap boundaries

2. add a reduced per calendar snapshot with a narrow date / duration range

3. include the full snapshots if you feel that’s useful despite the size as you asked.

let me know which direction you would prefer.

[SATVIKsynopsis]

rebased onto latest main after the snapshot refactor and resolved conflicts.

the pr now keeps only the optimized snapshot test and a snapshot output file with no large snapshot files committed. all CI checks pass too.

happy to add reduced or full snapshots if you would like, but I kept this minimal for now.