Fixed links to old, 404'ing volvo.com pages

Post by **matthew1** » 12 Oct 2024, 12:47

packetfire wrote: ↑12 Oct 2024, 12:34 Researching this, I stumbled across this document:

https://volvo.custhelp.com/app/answers/ ... 9-warranty

Which I attach a pdf of here in case the website is reorganized.
Partial Zero Emissions Vehicle (PZEV) Warranty.pdf

They DO reorganize their customer-facing pages often. Sooo many links here 404 from stuff I've posted and many other users have posted over the years. It's frustrating, and I understand somewhat because I worked in this industry (web dev) for 15 years, but man alive, redirects are a thing, Volvo North America.

Thanks for uploading the PDF, packetfire.

Post by **packetfire** » 12 Oct 2024, 13:17

> Sooo many links here 404 from stuff I've posted and many other users have posted over the years.

Archive.org?
Keyword search with volvo's domain as the "site" limitation?

Post by **matthew1** » 16 Oct 2024, 14:21

packetfire wrote: ↑12 Oct 2024, 13:17 > Sooo many links here 404 from stuff I've posted and many other users have posted over the years.

Archive.org?
Keyword search with volvo's domain as the "site" limitation?

Good suggestions. I need to compile a list of URLs in posts here, put it in a spreadsheet, see how big the scope of this is.

Just now I found another (soft) 404 from a link in a reputable source (DuckDuckGo):

which goes to

I'm not even looking for these, I'm just doing basic research. I've never set out to go looking for these, it's always just using the internet to do normal person stuff.

If I was even peripherally involved in Volvo's website, I'd speak up about this epidemic of soft 404s from old links. Redirects have been around for decades. If I was actually in charge of the Volvo website, I'd call an all-hands meeting yesterday and ask for a project timeline on my desk in the morning. You can't control the URLs pointing to you on other sites, but you can control how your site reacts to them. (Which is strangely analogous to human behavior.)

It's a little odd for a large corporation to leave this low-hanging fruit hanging low.

Post by **packetfire** » 19 Oct 2024, 13:56

@matthew1

OK, let's go all hyper-techno-geeky on the wrench-turners here...

So, you write a spider that crawls Matthews' as a whole, and for every URL that gets back a bad code or a "301", which Volvo seems to use a lot to toss you into a pit, You replace:

http://www.volvopage.volvosite.com/

With

http://web.archive.org//www.volvopage.volvosite.com/

In the post, and make the think open up archive.org, which gives one a darn good chance of seeing the old content linked-to by the author.

Spider and rewrite on board a night in the wee hours of each night, and you are done fairly quickly

Post by **matthew1** » 19 Oct 2024, 14:20

Yes, exactly, except no spiders. Just a db query for *volvo.com* inside a [URL] tag in all posts. Look at the result and make sure it's only stuff I want to do this to. Then execute a tricky find/replace on that result. (it's tricky because the database will hold two strings for any single URL instance in a post you and I see... one for the display URL, and one for the actual URL)

I've done things like this before here a handful of times, for instance when I switched over to https from http... and another time when I did a sweep to catch non-www in the URLs. Both of those were internal links... MVS links pointing to MVS pages.

Some voice in my head is telling me Volvo used a different domain... and I just now checked and yes there is more than just volvo.com:

https://www.volvocars.com/us/

... and that voice says there was another, long ago...

Now I need to shut up about this and Just Do It.

(you don't have to shut up about it, I'm just saying I need to get this thing done)

Thanks for bringing this up and your discourse on it, PF.

Post by **matthew1** » 19 Oct 2024, 15:30

I'm going to use this thread as a notepad.

SQL:

Code: Select all

SELECT post_id, post_text FROM `phpbb_posts` where post_text LIKE "%volvocars.com%";

839 results

Code: Select all

<URL url="http://new.volvocars.com/ownersdocs/2000/2000_SV70/00sv70_06.htm#pg97"><LINK_TEXT text="http://new.volvocars.com/ownersdocs/200 ... 6.htm#pg97">http://new.volvocars.com/ownersdocs/2000/2000_SV70/00sv70_06.htm#pg97</LINK_TEXT></URL></r>

There's a typical old URL (no https), and that's how the db stores them (and the actual display that we humans read in these posts). If anyone is interested, it's post_id 21989, so viewtopic.php?p=21989#p21989

The Wayback Machine strips off the #pg97, but the page is there:

https://web.archive.org/web/20080225214 ... 6.htm#pg97

But it puts that weird 20080225214415 in there, so what if I find and replace and don't put anything in that slot?

http://new.volvocars.com/ownersdocs/199 ... 850_1.html

will get prepended with

Code: Select all

https://web.archive.org/web/

and that URL when viewed becomes

https://web.archive.org/web/20171017041 ... 850_1.html

... and that works, even though it gets rewritten to add back that strange identifier after "/web". So a quick, limited proof of concept proves workable.

Post by **matthew1** » 19 Oct 2024, 15:46

Post by **packetfire** » 24 Oct 2024, 20:18

Well it is tricky to write a string match for EXACTLY the domain name you want, unless you put the "www." in front to make sure, and the ".com" or whatever on the end.

I think each variant of the volvo domain will need its own search string, and its own replace strategy.

But I have no clue how good the string searching is in whatever backend SQL is behind this beastie.

Post by **matthew1** » 28 Oct 2024, 14:10

Code: Select all

UPDATE phpbb_posts SET post_text = REPLACE(post_text, '<URL url="http://new.volvocars.com', '<URL url="https://web.archive.org/web/http://new.volvocars.com') WHERE post_text LIKE '%<URL url=\"http://new.volvocars.com%';

I just ran that, 357 changes made. Let's see if the MVS is a smoking ruin now. If not, I'll do more iterations of this find/replace.

Post by **matthew1** » 10 Nov 2024, 15:58

I split all the above posts off from Packetfire's thread. This project is complete. Results, number of changed links of each in parentheses:

Code: Select all

http://www.volvocars.com (117)
http://new.volvocars.com (0)
http://www.customers.volvocars.com (0)
http://accessories.volvocars.com (23)
https://www.volvocars.com (35)
https://new.volvocars.com (0)
https://www.customers.volvocars.com (85)
https://accessories.volvocars.com (40)