serve Atom feeds for the site, a section and a tag

Membership is a publication date, not a type declaration (ADR-0043). The parked
feed shape said "every type declared primary", which would have made feeds wait on
declared types a third time — but the thing that distinguishes a feed item is
already on disk. Pages and section landings drop out because they have no date,
which is the right reason. The parked idea stays parked with a sharper trigger:
someone wanting a *dated* bundle kept out.

Built with encoding/xml from typed structs, never a template: XML in html/template
is escaping for the wrong grammar, and that is a correctness trap rather than a
matter of taste.

A bug the evidence found, older than feeds: content.URL("", lang) built "//", so a
whole-site feed's id and alternate link were https://khosra.example// — every
entry identity wrong in every reader. The root is "/" now, with a test, and the
hand-built "/" the resolver carried for the same reason can follow later.

Two counters re-scoped rather than incremented, the same way transforms was:

Views now counts *per-bundle selection* — the thing architecture.md means by the
View layer, still at zero consumers. Output formats are not it: HTML, sitemap XML
and Atom are three functions with nothing to share, so an interface over them
would have one member and no leverage.

Effects stays at 1. A feed is generated per request like the sitemap, so it is not
a second Effect and the runner is not yet due — the next thing that writes files
off the request path is.

internal/web/resolve.go

@@ -20,6 +20,8 @@ type resolution struct {
 	// redirect is the canonical path when the request named a non-canonical one. Non-empty means answer
 	// with a permanent redirect and nothing else.
 	redirect string
+	// feed means the request named a feed of whatever scope key and tag describe (ADR-0043).
+	feed bool
 }
 
 // resolve maps a request path to a bundle key and language.
@@ -54,6 +56,15 @@ func resolve(path string, site *content.Site) (resolution, bool) {
 		return resolution{redirect: "/"}, true
 	}
 
+	// A trailing feed.xml names a feed of the scope before it. It is a file rather than a page, so none of
+	// the trailing-slash canonicalisation below applies to it (ADR-0043).
+	if rest, isFeed := cutFeed(key); isFeed {
+		if tag, section, isTag := cutTag(rest); isTag {
+			return resolution{key: section, tag: tag, lang: lang, feed: true}, true
+		}
+		return resolution{key: rest, lang: lang, feed: true}, true
+	}
+
 	page := 1
 	// A trailing /page/N/ is pagination, not part of the key (ADR-0028). Page one is the bare listing
 	// URL, so /page/1/ is a second spelling and redirects.
@@ -80,6 +91,18 @@ func resolve(path string, site *content.Site) (resolution, bool) {
 	return resolution{key: key, lang: lang, page: page}, true
 }
 
+// cutFeed strips a trailing feed.xml, reporting whether one was there. What remains is the scope: empty for
+// the whole site, a section, or a tag path.
+func cutFeed(key string) (rest string, ok bool) {
+	if key == feedFile {
+		return "", true
+	}
+	if trimmed, found := strings.CutSuffix(key, "/"+feedFile); found {
+		return trimmed, true
+	}
+	return key, false
+}
+
 // cutTag splits a tag listing key into its term and the section it is narrowed to.
 func cutTag(key string) (tag, section string, ok bool) {
 	if rest, found := strings.CutPrefix(key, content.TagsSegment+"/"); found {