@ai-ecoverse/slicc-man 1.1.0 → 1.1.1

package/README.md

@@ -134,7 +134,7 @@ Pages, the sitemap and the topic index are cached:
 
 Each site has its own directory in the cache, named by the base URL's host and port (`www.sliccy.com`, `127.0.0.1-8080`), with one JSON file per page, `sitemap.json` and `index.json`. A cached copy is fresh for the response's `Cache-Control` `max-age` (the site sends 4 hours), and 4 hours without one. A stale copy is revalidated with `If-Modified-Since` (the response's `Last-Modified`): `304` keeps it and makes it fresh again, `200` replaces it, and `404` drops it, so the topic has no page. When the site can't be reached, or answers 5xx, a cached copy is shown however old, after one line on stderr, `<cmd>: <url>: <reason>; showing the copy cached <date>`. With nothing cached, the command prints `<cmd>: <url>: <reason>` and exits 2. The cache is written atomically (a temporary file renamed into place), so two commands at once never see half a page; a cache that can't be written is skipped without a word.
 
-`apropos` and `whatis` read the index: every `/man/<topic>` entry of the sitemap, with its description, the first paragraph of its page, rendered as plain text on one line. The index is built from the cached sitemap and pages, fetching what is missing or stale (pages through the same cache, several at a time), and cached itself under the same rules. A page that fails to load is listed without a description.
+`apropos` and `whatis` read the index: every `/man/<topic>` entry of the sitemap, with its description, the first paragraph of its page, rendered as plain text on one line. The index is built from the cached sitemap and pages, fetching what is missing or stale (pages through the same cache, several at a time), and cached itself under the same rules. A page that fails to load is listed without a description; the other pages are still fetched, and an index with a failed page isn't cached, so the next command tries that page again.
 
 `SLICC_MAN_URL` overrides the base URL `https://www.sliccy.com`: pages are `$SLICC_MAN_URL/man/<topic>.md` and the sitemap `$SLICC_MAN_URL/sitemap.xml`, whose entries count by their path, whatever their host. The tests point it at a local server.
 

package/lib/site.mjs

@@ -154,12 +154,12 @@ export function site({ url, store, fetch, now = Date.now, name }) {
   async function build(list, offline) {
     const entries = new Array(list.length);
     let next = 0;
-    let down = offline;
+    let down = false;
     async function worker() {
       while (next < list.length) {
         const at = next++;
         const topic = list[at];
-        const found = await page(topic, { offline: down });
+        const found = await page(topic, { offline });
         if (found.stale || found.offline) down = true;
         entries[at] = { topic, description: found.status === 'ok' ? describe(found.body) : '' };
       }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ai-ecoverse/slicc-man",
-  "version": "1.1.0",
+  "version": "1.1.1",
   "description": "man, apropos and whatis for SLICC, and scman for SLICC, macOS, Linux and Windows, reading pages from www.sliccy.com/man",
   "type": "module",
   "license": "Apache-2.0",