release: 0.1.2

[singhpratech]

Version bump to 0.1.2, from scripts/bump_version.py --set 0.1.2; --check passes.

What 0.1.2 carries (both merged on main):

• Honour declared temporal precision and zones on PostgreSQL; db_schema option; schema catalog fill (validation P1-P4, P7, P8) #107 — on PostgreSQL through psqlodbc, columns declared with reduced precision now read at that precision (TIMESTAMP(0) → timestamp[s], TIMESTAMPTZ(3) → timestamp[ms, UTC], TIME(3) → time32[ms]); bulk ingest creates TIMESTAMP(n) / TIMESTAMP(n) WITH TIME ZONE / TIME(n) from the Arrow type; adbc.connection.db_schema; catalog names on schema-level GetObjects. Validation suite on PostgreSQL 293 / 327, the rest declared backend limits. Behaviour change to note in the release notes: code that assumed every PostgreSQL timestamp arrives as microseconds sees the declared unit instead; the values are unchanged.
• ingest: bind PostgreSQL array literals wide; LFConversion=0 for psqlodbc on Windows #109 — Windows: PostgreSQL array ingest bound its literals narrow, so non-ASCII text was stored double-encoded through the ANSI code page; now bound wide like the row path. psqlodbc gets LFConversion=0 on Windows so fetched text keeps its newlines as stored.