[3]: Same as [2], but for mobile.

SELECT
  COUNT(body), SUM(p1), SUM(p2), SUM(p3), SUM(p4), SUM(p5), SUM(p6), SUM(p7), SUM(p8), SUM(p9), SUM(p0), SUM(p1 + p2 + p3 + p4 + p5 + p6 + p7 + p8 + p9 + p0 > 0)
FROM (
SELECT
  body,
  REGEXP_MATCH(body, "![a-zA-Z]+\\.DeviceOrientationEvent") as p1,
  REGEXP_MATCH(body, "[a-zA-Z]+\\.DeviceOrientationEvent ?\\&\\&|\\&\\& ?[a-zA-Z]+\\.DeviceOrientationEvent") as p2,
  REGEXP_MATCH(body, "[a-zA-Z]+\\.DeviceOrientationEvent ?\\|\\||\\|\\| ?[a-zA-Z]+\\.DeviceOrientationEvent") as p3,
  REGEXP_MATCH(body, "== ?[a-zA-Z]+\\.DeviceOrientationEvent|[a-zA-Z]+\\.DeviceOrientationEvent ?==|[a-zA-Z]+\\.DeviceOrientationEvent ?!=") as p4,
  REGEXP_MATCH(body, "[a-zA-Z]+\\.DeviceOrientationEvent ?\\?") as p5,
  REGEXP_MATCH(body, "typeof ([a-zA-Z]+\\.)?DeviceOrientationEvent") as p6,
  REGEXP_MATCH(body, "[a-zA-Z]+\\[[\'\"]DeviceOrientationEvent[\'\"]\\]") as p7,
  REGEXP_MATCH(body, "[\'\"]DeviceOrientationEvent[\'\"] ?in ?[a-zA-Z]+") as p8,
  REGEXP_MATCH(body, "if ?\\( ?[a-zA-Z]+\\.DeviceOrientationEvent ?\\)") as p9,
  REGEXP_MATCH(body, "\\|DeviceOrientationEvent\\||DataView DeviceOrientationEvent Document") as p0, // False positives.
FROM [httparchive:response_bodies.2019_01_01_mobile]
WHERE url LIKE '%.js'
  AND body CONTAINS "DeviceOrientationEvent"
  AND REGEXP_MATCH(body, "DeviceOrientationEvent[^:]")
GROUP BY body, p1, p2, p3, p4, p5, p6, p7, p8, p9, p0)

Row	f0_	f1_	f2_	f3_	f4_	f5_	f6_	f7_	f8_	f9_	f10_	f11_	 
1	10010	1945	1474	337	955	187	280	4	1557	506	2963	9501	 

--> (f11-f10)/(f0-f10) = 92.7%


[4]: Same as [3], but without `url LIKE '%.js'`

Row	f0_	f1_	f2_	f3_	f4_	f5_	f6_	f7_	f8_	f9_	f10_	f11_	 
1	13428	2652	2529	448	1133	323	415	6	2249	813	3184	12733	 

--> (f11-f10)/(f0-f10) =  93.2%


[5]: Occurrences of code pattern `if(DeviceOrientationEvent)`.

SELECT page, url
  FROM [httparchive:response_bodies.2019_01_01_mobile]
  WHERE url LIKE '%.js'
    AND body CONTAINS "DeviceOrientationEvent" AND REGEXP_MATCH(body, "if ?\\( ?DeviceOrientationEvent ?\\)")
  LIMIT 100

Row	page	url	 
1	https://prod.gq.com/	https://d160accw6snlyf.cloudfront.net/public/js/thetake-injector.js	 
2	https://intothegloss.com/	https://d160accw6snlyf.cloudfront.net/public/js/thetake-injector.js	 
3	http://beautyisboring.com/	https://d160accw6snlyf.cloudfront.net/public/js/thetake-injector.js	 
4	https://www.purewow.com/	https://d160accw6snlyf.cloudfront.net/public/js/thetake-injector.js	 
5	https://mstag.gq.com/	https://d160accw6snlyf.cloudfront.net/public/js/thetake-injector.js	 
6	https://www.gq.com/	https://d160accw6snlyf.cloudfront.net/public/js/thetake-injector.js	 

The only site not served over HTTPS has an Alexa rank of 1944803.


[6]: Same as [5], but for `new DeviceOrientationEvent`.

Row	page	url
1	https://www.catprint.com/	https://www.catprint.com/assets/application-1dcbc2614efc594a9eee5fdee54bb91ec6735a7cb06952722e4838d2e04e80b0.js	

Alexa rank 71716.


[7]: Same as [5], but for `(evt: DeviceOrientationEvent)`, and without restricting to *.js.

Row	page	url	 
1	https://playground.babylonjs.com/	https://playground.babylonjs.com/babylon.d.txt	 
2	https://www.babylonjs-playground.com/	https://www.babylonjs-playground.com/babylon.d.txt	

This is not JS.


[8]: Same as [5], but for `(DeviceOrientationEvent)` more generically:

SELECT page, url
  FROM [httparchive:response_bodies.2019_01_01_mobile]
  WHERE url LIKE '%.js'
    AND body CONTAINS "DeviceOrientationEvent" AND REGEXP_MATCH(body, "[a-zA-Z0-9]\\( ?DeviceOrientationEvent?\\)")

Result is 3349 sites using `http://static.adsafeprotected.com/sca.17.4.95.js`, which has:

function(){return"function"==typeof DeviceOrientationEvent&&a239.a341.a66(DeviceOrientationEvent)}

This seems like an extremely popular script, but it correctly implementing feature-probing.


[9]:

SELECT page, url
  FROM [httparchive:response_bodies.2019_01_01_mobile]
  WHERE url LIKE '%.js'
    AND body CONTAINS "ondeviceorientation" AND REGEXP_MATCH(body, "if ?\\( ?ondeviceorientation?\\)")

No results.


[10]: Search code snippets for the `ondeviceorientation` event handler attribute.

SELECT
   REGEXP_EXTRACT(reduced_body, "(.{0,99}[^a-zA-Z]ondeviceorientation[^a-zA-Z].{0,99})") AS snippet
FROM (
  SELECT
    body,
    REGEXP_REPLACE(body, "[a-zA-Z]+\\.ondeviceorientation[^.]|typeof ondeviceorientation|\'ondeviceorientation\'|\"ondeviceorientation\"|\\|ondeviceorientation\\|", "") AS reduced_body
  FROM [httparchive:response_bodies.2019_01_01_mobile]
  WHERE url LIKE '%.js'
    AND body CONTAINS "ondeviceorientation"   
  GROUP BY body, reduced_body)
WHERE REGEXP_MATCH(reduced_body, "[^a-zA-Z]ondeviceorientation[^a-zA-Z]")
GROUP BY snippet

Nothing interesting, i.e. no usages that would immediately throw if we removed the attribute.


[11]: Same as [3], but counts feature probing for DeviceMotion.

SELECT
  COUNT(body), SUM(p1), SUM(p2), SUM(p3), SUM(p4), SUM(p5), SUM(p6), SUM(p7), SUM(p8), SUM(p9), SUM(p0), SUM(p1 + p2 + p3 + p4 + p5 + p6 + p7 + p8 + p9 + p0 > 0)
FROM (
SELECT
  body,
  REGEXP_MATCH(body, "![a-zA-Z]+\\.DeviceMotionEvent") as p1,
  REGEXP_MATCH(body, "[a-zA-Z]+\\.DeviceMotionEvent ?\\&\\&|\\&\\& ?[a-zA-Z]+\\.DeviceMotionEvent") as p2,
  REGEXP_MATCH(body, "[a-zA-Z]+\\.DeviceMotionEvent ?\\|\\||\\|\\| ?[a-zA-Z]+\\.DeviceMotionEvent") as p3,
  REGEXP_MATCH(body, "== ?[a-zA-Z]+\\.DeviceMotionEvent|[a-zA-Z]+\\.DeviceMotionEvent ?==|[a-zA-Z]+\\.DeviceMotionEvent ?!=") as p4,
  REGEXP_MATCH(body, "[a-zA-Z]+\\.DeviceMotionEvent ?\\?") as p5,
  REGEXP_MATCH(body, "typeof ([a-zA-Z]+\\.)?DeviceMotionEvent") as p6,
  REGEXP_MATCH(body, "[a-zA-Z]+\\[[\'\"]DeviceMotionEvent[\'\"]\\]") as p7,
  REGEXP_MATCH(body, "[\'\"]DeviceMotionEvent[\'\"] ?in ?[a-zA-Z]+") as p8,
  REGEXP_MATCH(body, "if ?\\( ?[a-zA-Z]+\\.DeviceMotionEvent ?\\)") as p9,
  REGEXP_MATCH(body, "\\|DeviceMotionEvent\\||DataView DeviceMotionEvent Document") as p0, // False positives.
FROM [httparchive:response_bodies.2019_01_01_mobile]
WHERE url LIKE '%.js'
  AND body CONTAINS "DeviceMotionEvent"
  AND REGEXP_MATCH(body, "DeviceMotionEvent[^:]")
GROUP BY body, p1, p2, p3, p4, p5, p6, p7, p8, p9, p0)

Row	f0_	f1_	f2_	f3_	f4_	f5_	f6_	f7_	f8_	f9_	f10_	f11_	 
1	2984	1597	423	34	79	228	11	31	747	35	25	2835	

--> (f11-f10)/(f0-f10) = 95%


[12]: Same as [1], but not restricted to *.js, and looks at mobile.

SELECT
   REGEXP_EXTRACT(reduced_body, "(.{0,99}[^a-zA-Z]DeviceOrientationEvent[^a-zA-Z].{0,99})") AS snippet
FROM (
  SELECT
    body,
    REGEXP_REPLACE(body, "[a-zA-Z]+\\.DeviceOrientationEvent[^.]|typeof DeviceOrientationEvent|\'DeviceOrientationEvent\'|\"DeviceOrientationEvent\"|\\|DeviceOrientationEvent\\|", "") AS reduced_body
  FROM [httparchive:response_bodies.2019_01_01_mobile]
  WHERE body CONTAINS "DeviceOrientationEvent"   
  GROUP BY body, reduced_body)
WHERE REGEXP_MATCH(reduced_body, "[^a-zA-Z]DeviceOrientationEvent[^a-zA-Z]")
GROUP BY snippet

Three interesting results:

onOrientationEvent(evt: DeviceOrientationEvent): void;
function(){return"function"==typeof DeviceOrientationEvent&&a239.a341.a66(DeviceOrientationEvent)}
return new DeviceOrientationEvent(e,t)


[13]: Same as [12], but for DeviceMotionEvent.

SELECT
   REGEXP_EXTRACT(reduced_body, "(.{0,99}[^a-zA-Z]DeviceMotionEvent[^a-zA-Z].{0,99})") AS snippet
FROM (
  SELECT
    body,
    REGEXP_REPLACE(body, "[a-zA-Z]+\\.DeviceMotionEvent[^.]|typeof DeviceMotionEvent|\'DeviceMotionEvent\'|\"DeviceMotionEvent\"|\\|DeviceMotionEvent\\|", "") AS reduced_body
  FROM [httparchive:response_bodies.2019_01_01_mobile]
  WHERE body CONTAINS "DeviceMotionEvent"   
  GROUP BY body, reduced_body)
WHERE REGEXP_MATCH(reduced_body, "[^a-zA-Z]DeviceMotionEvent[^a-zA-Z]")
GROUP BY snippet

Two interesting results:

a239.a341.a66(DeviceMotionEvent)
return new DeviceMotionEvent(e,t)