--component as a first-class query filter with case-insensitive name expansion

[arcaven]

Proposal

--component as a first-class query filter — on issue list and every command that accepts --jql (e.g. the proposed issue comment add --jql #586, issue changelog --jql #587):

jr issue list --component Okta                          # component = resolved-ids
jr issue list --component Okta --component Teleport     # OR: component in (…)
jr issue list --component not:Vector                    # exclusion
jr issue list --component none                          # component is EMPTY (untagged-review queue)
jr issue list --component all:Vector --component all:Kubernetes   # issue has BOTH

Composes by AND with all other structured filters and with --jql (structured flags append AND (…)).

Resolution must be deterministic, not name-string trust

A primary purpose of components is search, and component names are load-bearing strings: cross-project queries match by name, and casing drift (CloudSmith vs Cloudsmith, Github vs GitHub, n8n vs N8N — real examples from our two projects) silently splits results. So resolve client-side by ID:

1. Fetch components for the target project(s) — one cached call per project.
2. NFKC-normalize + Unicode-casefold both sides; collect all matching component IDs across case variants and projects.
3. Emit component in (id1, id2, …) — deterministic regardless of server-side name semantics; multi-project correctness (same name, different per-project IDs) falls out for free.
4. Zero matches → edit-distance suggestion: did you mean: CloudSmith?. --exact opts out of case-folding.

Related: the jr component subcommand family and the generalized filter grammar are filed separately.