This is done. Took a while to figure out, lot of soft-404s. The IABot database is also updated, so those URLs will be repaired on other wikis also. --
GreenC15:43, 20 March 2023 (UTC)
I started making these changes, but there is too many for non-bot edits. In addition, in the citations, it might be worth removing the link to the Internet archive and removing the parameter with the status of a dead link.
Thanks! Search results have not updated yet, so there are still
almost 1200 matches. Some of these seem real, for example
[1] on
Zürich. I've not used JWT, but if it finds articles by regex search you might need to switch to a
broader search, because I'm not sure regex search is working right now.
This kind of work is more complicate then it might seem there are different file types (pdf, txt, wav etc) that should be handled differently, and different types of file structures at IA to deal with. it took me a while to find all the issues and develop code for it that gives reasonable results. It's not a simple search-replace thought sometimes it is. I have developed code for it and can process them, but some people get upset about it, so I have been hesitant to try and fix them all in one go, and just catch them incidentally when the bot happens to process an article with one there. Possibly I could modify it to only process when the link is dead so there is no question that something should be fixed. --
GreenC14:30, 15 March 2023 (UTC)
Handled differently how? Do you have an example? The links in the form https?://ia[0-9]+.us.archive.org/[0-9]+/items/(.+) always go to the mere file download just like the /download/ URL. The trouble begins if you start operating on the various other kinds of URLs which might include things like us.archive.org/view_archive.php or various views like /stream/ . I'm not proposing a complete normalisation of all IA URLs, only the /download/ ones which are definitely going to break.
Nemo07:14, 23 March 2023 (UTC)
It's in about 1,300 pages. I'm running it now, for links that are already dead. The others will rerun periodically after they die. Some users don't get it, I don't want to deal with them unless the link is already dead in which case they have no basis. --
GreenC16:13, 23 March 2023 (UTC)