Documentation
¶
Overview ¶
Package webfeed is the contract behind txco://web/feed: the request a rule may make, the answer it gets back, and the reading of a feed document (RSS, Atom or JSON Feed) into that answer. It does no networking: the handler in chassis/server fetches through chassis/fetch and hands the body here.
A feed is what a site publishes about itself: its newest items, each a title, an address, a date and a few lines. Reading it answers "what has this site published" without fetching the site's pages, which many refuse to a program. Items are external text: the stack that reads them decides what they may do.
Index ¶
Constants ¶
const ( MaxCount = 50 DefaultCount = 20 // MaxScanItems is how many items are read from one document; any past // them are left unread and the answer says it was truncated. A feed // holds its newest few dozen. MaxScanItems = 500 MaxTitleBytes = 1 << 10 MaxSummaryBytes = 4 << 10 MaxURLBytes = 2 << 10 MaxIDBytes = 1 << 10 MaxAuthorBytes = 256 MaxAnswerBytes = 256 << 10 )
The contract's bounds.
const ( CodeInvalidRequest = "invalid_request" CodeDisabled = "disabled" CodeRateLimited = "rate_limited" CodeBusy = "busy" CodeNotAFeed = "not_a_feed" CodeInvalidFeed = "invalid_feed" CodeNoTenant = "no_tenant" CodeInternal = "internal_error" )
Error codes. The handler prefixes them txco_feed_; the fetcher's own codes (timeout, destination_denied, http_error, …) pass through under the same prefix.
const DefaultInto = "_feed"
DefaultInto is where the answer lands when a rule names no `into`.
Variables ¶
This section is empty.
Functions ¶
func Discover ¶
Discover lists the feeds an HTML page advertises with <link rel="alternate" type="application/rss+xml" href="…"> and its kin, resolved against base, in page order, without repeats.
func Parse ¶
Parse reads and checks the op's WITH object. The URL itself is checked by the handler with the fetcher's policy. An *Error names the first field at fault.
Types ¶
type Answer ¶
type Answer struct {
OK bool `json:"ok"`
URL string `json:"url"`
FinalURL string `json:"final_url,omitempty"`
Status int `json:"status,omitempty"`
// Format is rss (RSS 2.0, 0.9x and RSS 1.0), atom or json.
Format string `json:"format"`
Feed Info `json:"feed"`
Items []Item `json:"items"`
Count int `json:"count"`
// Total is how many items the document holds (up to MaxScanItems),
// before `since` and `count`.
Total int `json:"total"`
Truncated bool `json:"truncated"`
}
Answer is the op's result object on success.
type Document ¶
type Document struct {
Format string
Feed Info
Items []Item
// Total is how many items the document holds, up to MaxScanItems.
Total int
// Truncated says the document was not read whole: more items than
// MaxScanItems, more elements than the reader holds, or a syntax
// error after some items were read.
Truncated bool
}
Document is a feed as read: its format, what it says about itself, and its items in document order, each already plain text with its date in RFC 3339 and its URL resolved and checked. Nothing is cut yet: Select applies the request and the caps.
type Error ¶
type Error struct {
Code string `json:"code"`
Message string `json:"message"`
Status int `json:"status,omitempty"`
Feeds []string `json:"feeds,omitempty"`
}
Error is the in-band failure. Feeds is set on not_a_feed when the address was an HTML page that advertises feeds: the addresses to try.
type Failure ¶
type Failure struct {
OK bool `json:"ok"`
URL string `json:"url"`
FinalURL string `json:"final_url,omitempty"`
ContentType string `json:"content_type,omitempty"`
Error *Error `json:"error"`
}
Failure is the op's result object when the feed was not read. The body is never in it.
type Info ¶
type Info struct {
Title string `json:"title"`
Link string `json:"link,omitempty"`
Description string `json:"description,omitempty"`
Updated *string `json:"updated"`
}
Info is what the feed says about itself.
type Item ¶
type Item struct {
ID string `json:"id"`
Title string `json:"title"`
URL string `json:"url"`
// Date is the item's own date (published, else updated), RFC 3339 in
// UTC, or null.
Date *string `json:"date"`
Summary string `json:"summary"`
Author string `json:"author,omitempty"`
// Source is the item URL's host, lower case.
Source string `json:"source,omitempty"`
}
Item is one entry of the feed.
func Select ¶
Select applies the request to a document's items: those before `since` go; when every remaining item is dated they are put newest first, otherwise the document's order stands; the first `count` are kept; and each is cut to the contract's caps (a title, summary or author at its cap on a character boundary; an id or URL over its cap dropped from the item, since a cut one points somewhere else). The answer as a whole stays under MaxAnswerBytes by dropping items from the end. Truncated reports any cut or drop for size; `count` is a choice, not a cut.
Source Files