From IndieWeb

In 2015 I moved my site from Nucleus to ProcessWire. In the process, I decided to update my permalink URL structure. Below I will document problems and the solutions I've found.

Existing URLs

  • article permalinks used URL path /YYYY/MM/[post id]/
  • notes permalinks used URL path /notes/YYYY/MM/DD/[ordinal]/

New URLs

  • article permalinks will use URL path /YYYY/MM/[slug]/
  • notes permalinks will use URL path /YYYY/MM/[slug]/


Accessing Old URLs

The existing URL paths are stored along with each page in ProcessWire. The ProcessWire templates detect the old URL path and return a 301 redirect to the new permalink.

De-duping Webmentions

  • To uniquely identify and de-dupe webmentions, my code uses md5(source + target). Since the permalinks will be changing in my notes h-feed, I suspect bridgy will try to re-send recent webmentions to the new URLs. Since the md5 hash will be different, they will not be properly de-duped.
    • My solution: When migrating webmentions to ProcessWire, I re-calculated the hash to be md5(source + target path sans scheme). For example, instead of using in the hash, I will use Leaving the scheme off is a bit of future-proofing, since I intend to deliver all my pages https-only eventually.
  • . . .