Expose X-GM-THRID via Fetch::gmail_thread_id()

[benchub]

Summary

imap-proto already parses the Gmail thread id (X-GM-THRID) into AttributeValue::GmailThrId, but Fetch provided no getter for it — unlike its siblings gmail_msg_id() (X-GM-MSGID) and gmail_labels() (X-GM-LABELS). This adds the missing accessor:

/// Get the Gmail thread id (X-GM-THRID). Only present if the server
/// returned it (Gmail's X-GM-EXT-1).
pub fn gmail_thread_id(&self) -> Option<&u64>

It mirrors gmail_msg_id() exactly, just matching AttributeValue::GmailThrId instead of AttributeValue::GmailMsgId.

Motivation: X-GM-THRID lets clients group messages into Gmail conversations/threads without re-implementing Gmail's threading heuristics. Since gmail_msg_id/gmail_labels are already exposed, this fills the gap for the one remaining Gmail extension attribute imap-proto parses.

Test plan

• Added parse_fetches_gmail_thread_id in src/parse.rs, following the existing parse_fetches_test convention: feeds a * 24 FETCH (... X-GM-THRID 1278455344230334865) response through parse_fetches and asserts gmail_thread_id() returns the parsed value, and None for a FETCH response without it.
• cargo fmt --check and cargo test --features runtime-async-std pass (70 unit tests, including the new one).

🤖 Generated with Claude Code

[benchub]

One thing worth flagging for the record: AttributeValue::GmailThrId was introduced in imap-proto 0.16.6 (0.16.4/0.16.5 only have GmailMsgId/GmailLabels). That's newer than the imap-proto = "0.16.4" floor this crate currently declares in Cargo.toml.

In practice this isn't a problem under normal dependency resolution — Cargo will happily pick 0.16.6+ within that ^0.16.4 range, and that's exactly what cargo build/cargo test did in my testing. But a strict minimal-versions build (cargo +nightly build -Z minimal-versions) pinned to exactly imap-proto 0.16.4 would fail to compile against this new accessor.

I left the Cargo.toml floor alone to keep this diff minimal and not second-guess your dependency-floor policy, but happy to bump imap-proto to "0.16.6" here (or in a follow-up) if you'd prefer that be explicit. Just let me know!

[benchub]

Oh hey, this duplicates #133 and #129.... but this time with tests. I guess there's a desire.