Closed tollmanz closed 8 years ago
This was done in https://github.com/CondeNast/purgely/commit/10038d3982a0bb392973c9db03e8a84a534758bc. I do still like the idea of related URL purging for situations when you have a URL, but don't know what the related URLs might be. Anyway, nothing more to do here.
The more I think about the idea of finding related URLs and issuing individual purge requests for those URLs, I think that we are not leveraging surrogate keys to their full potential. Currently, when a URL purge is requested with related URLs, related URLs are found (e.g., term pages, feed URLs, author pages) and each is purged separately.
Theoretically, a
post-{id}
surrogate key should be available on each of those "related URLs". Thus, surrogate purging should be sufficient.I don't think anything is lost with this transition. In fact, the cache purging is likely to be more precise and comprehensive with this method.