mirror of
https://github.com/alam00000/bentopdf.git
synced 2026-11-05 20:02:47 +01:00
David Cumps
2b56c45177
Review follow-up: the origin regex also matched a port (https://www.bentopdf.com:8443) and userinfo (https://[email protected]), which are different origins; the lookahead now excludes ':' and '@'. The replacement is a callback, so `$&`/`$1` in SITE_URL stay literal. Tests cover both.
24 lines
1.1 KiB
JavaScript
24 lines
1.1 KiB
JavaScript
/**
|
|
* The public origin the build writes into canonical / og:url / JSON-LD URLs.
|
|
* Self-hosters set SITE_URL at build time; the default is the official site.
|
|
*/
|
|
export const DEFAULT_SITE_URL = 'https://www.bentopdf.com';
|
|
|
|
// The exact authority only: not a longer host (`.com.example`), a port
|
|
// (`:8443`, a different origin) or userinfo (`@other.example`, another host).
|
|
const DEFAULT_ORIGIN_RE = /https:\/\/www\.bentopdf\.com(?![\w.:@-])/g;
|
|
|
|
/**
|
|
* Rewrites absolute URLs on the default origin to `siteUrl` (no trailing
|
|
* slash). Used for hand-written pages that carry the official origin
|
|
* literally, so a build with a custom SITE_URL emits canonicals on its own
|
|
* host. Other origins are left untouched: look-alike hosts such as
|
|
* `https://www.bentopdf.com.example`, other ports and userinfo URLs.
|
|
*/
|
|
export function rebaseDefaultOrigin(html, siteUrl) {
|
|
if (siteUrl === DEFAULT_SITE_URL) return html;
|
|
// A callback keeps `siteUrl` literal: a string replacement would expand
|
|
// `$&`, `$1`, … inside it.
|
|
return html.replace(DEFAULT_ORIGIN_RE, () => siteUrl);
|
|
}
|