alt Hacker Newsit's not an insult. yes. cache heavily. shitty php stacks serving trillions of dollars of ecommerce sales have managed to do it for a long time.
Replies
Those ecommerce stacks serve a large fixed number of pages. cgit does not.
Imagine if WooCommerce had a route "/product/<sku1>/compare/<sku2>" which displayed an auto-generated comparison between any two product pages.
Now imagine running a million-SKU WooCommerce site, where each product page was 100kb of text.
Now imagine scrapers are permuting those URLs.
How would you cache that?
That's what cgit/kernel.org and many other Git forges are dealing with. These aren't static websites, even if the underlying Git repo is largely static; they're rendering arbitrary diffs and other generated-on-the-fly views into Git history. The ability to do that is a large part of the value of a Git UI.
Can you explain how a cache lets you avoid serving a request for the first time?