issue_comments: 1475016834

This data as json

html_url	issue_url	id	node_id	user	created_at	updated_at	author_association	body	reactions	issue	performed_via_github_app
https://github.com/simonw/datasette/pull/1999#issuecomment-1475016834	https://api.github.com/repos/simonw/datasette/issues/1999	1475016834	IC_kwDOBm6k_c5X6viC	9599	2023-03-18T22:30:31Z	2023-03-18T22:30:31Z	OWNER	`test_paginate_using_link_header` will be tricky. The reason the tests are failing is that `json_renderer()` attempts to populate the `link` header using `data["next_url"]` - but that's not present unless `?_extra=next_url` has been passed: https://github.com/simonw/datasette/blob/2f38479dcc81f11a4362f4e28511fa88afc34e61/datasette/renderer.py#L101-L102 But I can only rely on `next` being present, not `next_url`. I thought I could maybe assemble the `link` header using `next`, by turning that into a `next_url` link - but there's some custom logic which kicks in for pagination against SQL views (offset/limit pagination) to calculate the `next_url` which isn't easily replicable at the `json_renderer()` layer, in the `_next_value_and_url()` utility function: https://github.com/simonw/datasette/blob/2f38479dcc81f11a4362f4e28511fa88afc34e61/datasette/views/table.py#L2275-L2282	{ "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }	1551694938