I'm having trouble with the /page/summary/{title} endpoint of the REST API at /api/rest_v1
Long story short, I'm trying to request specific language variants from wikipedias that could serve different dialects to the same url (ex: sr.wikipedia.org can render in raw Serbian, or in latinized Serbian sr_el, and zh.wikipedia.org can render Simplified Chinese zh_CN or Traditional Chinese zh_TW)
However, even though there is ostensibly an Accept-Language header param in the API that should be the way to set it, (based on past discussion in https://phabricator.wikimedia.org/T122942), using it doesn't seem to actually do anything.
For example, even with the example provided in docs at https://sr.wikipedia.org/api/rest_v1/#!/Page_content/get_page_summary_title, running Accept-Language sr-el still gives back the raw Serbian: {F27689857}
Same story if I call the API myself instead of through the UI in the docs: https://codepen.io/anon/pen/zyNVbq
(basically, I only ever get the extract as seen in https://sr.wikipedia.org/sr/%D0%88%D1%83%D1%82%D1%98%D1%83%D0%B1, not https://sr.wikipedia.org/sr-el/%D0%88%D1%83%D1%82%D1%98%D1%83%D0%B1 )
I've also tried this with "zh-tw", "tw", "zh-TW" and "zh_TW" for zh.wikipedia.org, but I only ever get Simplified Chinese in the response.
(compare https://zh.wikipedia.org/zh-cn/%E7%B4%90%E7%B4%84%E5%AE%A2 and https://zh.wikipedia.org/zh-tw/%E7%B4%90%E7%B4%84%E5%AE%A2)
What's going on here? Why isn't the accept-language param working?