fix: drop broken using meosType = MeosType; alias in tydef.hpp

[estebanzimanyi]

High-leverage 1-file fix unblocking every in-flight PR's Linux arm64 build.

The bug

src/include/tydef.hpp:18 reads:

using meosType = MeosType;

This was added in c8cad6d ("Pre-existing build issue surfaced when the headers were touched — vcpkg's MEOS now exposes MeosType instead of meosType…") on the assumption that the pinned MEOS exposed the upper-case spelling. That assumption is no longer true.

The pinned MEOS (vcpkg_ports/meos/portfile.cmake REF f11b7443ee98…) exposes meosType (lower-case) via meos/include/temporal/meos_catalog.h:

extern uint32 datum_hash(Datum d, meosType basetype);
extern Set *set_in(const char *str, meosType basetype);
// …

So the alias resolves using meosType = MeosType; where MeosType doesn't exist → every PR's Linux arm64 build dies with:

/duckdb_build_dir/src/include/tydef.hpp:18:18:
  error: ‘MeosType’ does not name a type; did you mean ‘meosType’?

Confirmation that MobilityDuck doesn't actually need MeosType

$ grep -rn '\bMeosType\b' src/ --include='*.cpp' --include='*.hpp'
src/include/tydef.hpp:14:// Forward-compat alias for the meosType → MeosType rename…
src/include/tydef.hpp:15:// pr785-sync-script).  Vcpkg's MEOS exposes \`MeosType\`; existing
src/include/tydef.hpp:18:using meosType = MeosType;     # ← the broken line

MeosType is referenced only by the alias line and its comment. Every other call-site uses the lower-case meosType (which MEOS already provides):

$ grep -rn '\bmeosType\b' src/ --include='*.cpp' --include='*.hpp' | wc -l
30+   # ≥ 30 active call-sites, all in src/temporal/spanset_functions.cpp and similar

The fix

Delete the alias line and replace its comment with one that explains the current state (so future MEOS-pin bumps don't re-introduce the same bug). Zero behavioural change to MobilityDuck.

Impact: unblocks the whole open-PR backlog

Every Linux arm64 build currently fails identically — this includes:

PR	Title
#126	parity batch — bearing + eCovers/tCovers + stbox dim + seqSetGaps
#130	feat(parity): close final 14-name gap to reach 100% addressable coverage
#149	Longjmp out of the MEOS error handler
#158	edge-to-cloud quickstart + temporalFooter + SRID/geodetic fix
#159	doc: PR Reviewer Guide (rebased #117)
#160	docs+scripts: PR coordination policy (rebased #120)
#148 / #150 / #151 / #153 / #155 / #156	the feat/*_port_core extended-type stack

…all of which would build green once this merges.

Refs

Unblocks #126, #130, #149, #158, #159, #160 + the *_port_core stack.

[estebanzimanyi]

Reviewer's quickstart — ~2 minutes

What this PR does in one sentence: removes one line (using meosType = MeosType;) from src/include/tydef.hpp that was added on the assumption that the vcpkg MEOS pin had been bumped past the consolidation rename — but the current pin still exposes the pre-consolidation meosType, so the alias references a non-existent type and breaks every PR's Linux arm64 build.

Files to read (1 file, 1 line removed):

• src/include/tydef.hpp — drop line 18 + update the comment above it to document the pre/post-consolidation distinction so the next pin bump knows what to do.

Naming context (for clarity, not action):

• meosType (lowercase) = pre-consolidation spelling, exposed by the current pin f11b7443ee98…
• MeosType (uppercase) = post-consolidation target (the rename is upstream-MobilityDB work in flight; not yet reflected in the pin)
• MobilityDuck's source consistently uses meosType (matches the pin) — grep -rn '\bMeosType\b' src/ confirms MeosType is referenced ONLY on the alias line itself.

Smoke verification (Linux arm64 CI): already SUCCESS on this PR's most recent run. Linux amd64 FAILURE was a transient curl network error in vcpkg, not a code issue — re-trigger to confirm.

Why it's safe to merge: no-op for MobilityDuck's own code (which uses meosType directly); unblocks ~10 in-flight PRs (#126, #130, #149, #158, #159, #160 + the feat/*_port_core stack) that all fail the same 'MeosType' does not name a type build step on Linux arm64.

When the MEOS pin is bumped past consolidation (future PR): either restore the bridge (using meosType = MeosType; will then be valid because MeosType will exist) or sweep MobilityDuck source meosType → MeosType in one PR — the comment in the modified file documents both options.