Documentation
¶
Overview ¶
Package fetch provides a thin adapter between the crawl orchestrator and the Jira Cloud HTTP client. Its single responsibility is: given an issue key, return the raw JSON bytes for that issue.
fetch knows nothing about parsing, link extraction, Markdown rendering, or crawl queue management. It delegates all HTTP concerns (auth, retries, rate-limit backoff, typed error sentinels) to the client package.
Fetcher interface ¶
The Fetcher interface is the contract the crawl package depends on. Callers that need to inject a fake in tests can implement Fetcher without spinning up an httptest.Server.
Error propagation ¶
Typed sentinel errors from the client package (ErrUnauthorized, ErrForbidden, ErrNotFound, ErrRateLimited) are propagated without additional wrapping so that callers can use errors.Is directly:
raw, err := f.Fetch(ctx, "PROJ-1")
if errors.Is(err, client.ErrUnauthorized) { /* abort crawl */ }
if errors.Is(err, client.ErrForbidden) { /* render stub */ }
if errors.Is(err, client.ErrNotFound) { /* render stub */ }
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type ClientFetcher ¶
type ClientFetcher struct {
// contains filtered or unexported fields
}
ClientFetcher is the production implementation of Fetcher. It wraps a *client.Client and delegates all HTTP concerns to it. Construct via New or NewFromConfig; the zero value is not valid.
func New ¶
func New(c *client.Client) *ClientFetcher
New constructs a ClientFetcher from an already-constructed *client.Client. Use this when the caller already holds a client (e.g. the gojira facade that shares one client across fetch and other operations).
func NewFromConfig ¶
NewFromConfig is a convenience constructor that builds both a *client.Client and a *ClientFetcher from a validated config.Config. Useful for callers that do not need to share the underlying client with other components.
Any client.Option values (e.g. WithHTTPClient for tests) are forwarded to client.New unchanged. In addition, client.WithLogger and client.WithRoundTripper are commonly forwarded by the crawl observability instrument to install the httplog RoundTripper around the underlying client's transport; both flow through this variadic unchanged.
func (*ClientFetcher) Fetch ¶
Fetch retrieves the raw JSON bytes for the Jira issue identified by key.
On success it returns the bytes exactly as received from the Jira API. On failure it returns one of the client sentinel errors (client.ErrUnauthorized, client.ErrForbidden, client.ErrNotFound, client.ErrRateLimited) or a wrapped transport error. All sentinels remain reachable via errors.Is.
Fetch does not add retry logic (that is the client's responsibility), does not emit events (that is the crawl's responsibility), and does not parse the returned bytes.
The underlying GET request is issued with `expand=names` so the response includes a top-level "names" object mapping every field ID (including customfield_NNNNN) to its human-readable label. `internal/parse` consumes that object and surfaces it as `parse.Issue.Names`, which `internal/render` uses to render custom fields with the label the Jira UI shows rather than the opaque numeric ID. The expansion is per-issue and adds no extra HTTP requests; the cost is a slightly larger response body.
Source Files