NEWS

Rtumblr 0.2.0

The HTTP backend has been migrated from httr to httr2. OAuth 1.0a signing and the interactive browser authentication flow are now implemented in-package; the user-facing setup is unchanged (set RTUMBLR_TOKEN="consumer_key;consumer_secret").
rtumblr_auth() is exported to (re-)run the OAuth flow explicitly. The access token is now cached under tools::R_user_dir("Rtumblr", "cache") instead of .httr-oauth, so you will be asked to authorize once after upgrading. Power users can supply a pre-issued access token via a four-field RTUMBLR_TOKEN ("consumer_key;consumer_secret;oauth_token;oauth_token_secret") to skip the browser.
get_blog_posts() now paginates automatically. Pass n to control how many posts to retrieve across pages (n = Inf fetches the whole blog). The same n argument was added to get_blog_likes(), get_blog_following(), get_blog_followers() and get_blog_blocks().
The default limit for get_blog_posts() (and the following/followers helpers) is now 20, matching the Tumblr API maximum (the previous default of 50 was silently capped).
Fixed a bug where get_blog_blocks(), get_blog_following() and get_blog_followers() ignored the offset argument.
New read-only endpoints: get_user_info(), get_user_limits(), get_dashboard(), get_user_likes(), get_user_following(), get_post() and get_post_notes().
Tests now use httptest2 instead of vcr.