Fix DEC and longitude conversions in Meade commands - #296

Heads-up about a merge hazard rather than a review comment, since #304 touches the same thing.

This and #304 shouldn't both land. #296 negates the longitude in MeadeCommandProcessor::onSetSiteLongitude, #304 does it in the parser. Different files, so git merges them happily and the result negates twice: the reader gives -121°53', then the handler flips it back to +121°53' East. No conflict, and no test that would catch it.

Three other things for whoever picks a direction:

• This reintroduces the sign-of-zero problem on both sides (getTotalHours() > 0, and Longitude(-degrees, ...) where sign(0) is 1), which is what fix(meade): keep the sign of coordinates whose degrees component is zero #305 is removing.
• It doesn't touch readLongitude, so the unsigned :Sg121*53# that INDI actually sends still won't parse.
• The DEC half predates fix(meade): restore hemisphere conversion for DEC coordinates #300 and fix(meade): build declination from wire components without losing the sign #302, and it conflicts with fix(meade): keep the sign of coordinates whose degrees component is zero #305's struct change.

The getter-negation idea here is the right one and is what #304 does now, so credit where it's due if this ends up superseded.

Closing — both fixes in this PR have now landed via the post-#291 repair series, and the diagnosis here was correct on both counts, including the longitude convention (which turned out to be the contested one).

• DEC hemisphere conversion (decToInternal / decFrom): restored in fix(meade): restore hemisphere conversion for DEC coordinates #300, then hardened in fix(meade): build declination from wire components without losing the sign #302 and fix(meade): keep the sign of coordinates whose degrees component is zero #305. fix(meade): keep the sign of coordinates whose degrees component is zero #305 carries the sign as a separate negative channel, so wire forms like -00*30:00 survive where a sign-in-degrees struct cannot express them; the round trips are pinned by tests in unit_tests/test_core/meade/.
• Longitude — east-negative on the wire vs. east-positive internal: restored in fix(meade): accept the unsigned site longitude INDI sends #304. readLongitude/writeLongitude fold both the signed form and the legacy unsigned (0..360 westward) form through one mapping, matching pre-refactor: Enhance Meade command parsing and response handling #291 Longitude::ParseFromMeade, the LX200 spec wording, and what INDI actually sends. When the convention was weighed before merging fix(meade): accept the unsigned site longitude INDI sends #304, this PR's position is the one that won.

The merged versions also distinguish the sub-degree cases (000*30 west vs 359*30 east) that a negated-degrees approach cannot. Thanks for the report and the correct read on the convention.