Extend the still-running-dont-update-again delay

This reduces unnecessary bandwidth for everybody. Shouldn't happen much,
but good to avoid regardless. We do want to keep some kind of checking,
in case one release quickly follows another, which could happen if
there's an unexpectedly broken release or an urgent hotfix required.

Author: pimterry

src/index.ts

@@ -236,12 +236,12 @@ export async function runHTK(options: {
                 if (isErrorLike(error)) {
                     // Did we receive a successful update, that wants to restart the server:
                     if (error.code === 'EEXIT') {
-                        // Block future update checks for one hour.
+                        // Block future update checks for 6 hours.
 
                         // If we don't, we'll redownload the same update again every check.
                         // We don't want to block it completely though, in case this server
                         // stays open for a very long time.
-                        return delay(1000 * 60 * 60);
+                        return delay(1000 * 60 * 60 * 6);
                     }
 
                     // Report any HTTP response errors cleanly & explicitly: