When you host something on Drive, it takes the url https://googledrive.com/host/fileid , so it actually still continues to work from the trash bin. Yesterday I thought it was time to clean out my trash so I permanently deleted everything, including the hosted gadget – so every page on this site started showing an error where the missing gadget used to be.
Luckily I had the source for the gadget on GitHub, so I could easily recreate it, but the problem with using Drive as a host is that the fileid is a oneoff thing – meaning that I cannot just create another file with the same id.
So I was faced with either manually editing each of 700+ pages on this site and reinserting a new gadget, or patching up the existing gadget with the new fileid.
Here’s a snippet to do it.
DANGER and IMPORTANT – make sure you do lots of checking on what will be changed before actually doing the setHtmlContent() . This could mess up your site if you replace the wrong thing with the wrong thing. NOTE also that some deprecated gadgets will stop working when rewritten. For example , those old Adsense gadgets that Google deprecated on Sites. If your page gets rewritten and you still have some of them on your site – they will disappear forever. If you have gadgets on your site, test thoroughly that they will not be screwed up by rewriting the html before running.
All I had to do was set the old fileId
var searchRx = /0B92ExLh4POiZT081dWl6WjdwbFk/g;
And the new one
var replaceWith = "0B92ExLh4POiZb19Qc3hidFgzS2M";
Always worth giving it a trial run first by setting this to false
var actuallyDoIt = true;
function patch() {
// set this to true to actually do the update as opposed to just reporing it
var actuallyDoIt = false;
// this is the site i'm working with
var site = SitesApp.getSite("mcpher.com", "share");
// get all the pages on the site
var pages = getPages(site);
// get all the pages that contain the search term in their html
var searchRx = /0B92ExLh4POiZT081dWl6WjdwbFk/g;
var workingPages = pages.reduce (function (p,c) {
var html = c.getHtmlContent();
// if its a target page then select it and avoid getting the html again later
if (html.search (searchRx , html) >=0) {
p.push ({html:html, page:c});
}
return p;
},[]);
//now we'll update these pages with the replacement value
var replaceWith = "0B92ExLh4POiZb19Qc3hidFgzS2M";
workingPages.forEach (function (d,i) {
Logger.log (""+i+" "+(actuallyDoIt ? "updating " : "would have updated ") + d.page.getUrl());
// REMEMBER TO DO LOTS OF LOGGING AND CHECKING HERE BEFORE SETTING ACTUALLYDOIT TO TRUE
if (actuallyDoIt) {
d.page.setHtmlContent ( d.html.replace (searchRx, replaceWith) );
}
});
}
/**
* get all the pages on my site - into a heirach object
* @param {Site} site the site
* @param {number} optMax max number of pages to go for
* @param {string} optQuery a search query (unfortunately this doesnt search html)
* @param {number} optStart you can start later - useful for restarting if you run out of execution time
* @return {Array.Page} all the pages on the site
**/
function getPages(site,optMax,optQuery,optStart) {
var chunk,pages = [], options = {};
do {
options.start= pages.length + (optStart || 0);
if (optMax) options.max = optMax - pages.length ;
if (optQuery) options.search = optQuery;
chunk = cUseful.rateLimitExpBackoff( function (){
return site.getAllDescendants(options);
});
cUseful.arrayAppend(pages, chunk);
}
while (chunk.length && (!optMax || pages.length
For this snippet you'll need my cUseful library. You can find the details below or get it from github.
Why not join our forum,follow the blog or follow me on Twitter to ensure you get updates when they are available.