206216 2020-01-14 00:08:59 -0800 PointerEvent has 0 as pressure when the input device is not pressure-sensitive 2026-01-12 09:04:41 -0800 1 1 1 Unclassified WebKit UI Events Safari 13 Mac macOS 10.15 NEW InRadar P2 Normal --- 1 nekomanma webkit-unassigned cdumez darin dbates esprehn+autocc ews-watchlist graouts kangil.han thorton webkit-bug-importer wenson_hsieh oldest_to_newest 1606287 0 nekomanma 2020-01-14 00:08:59 -0800 W3C's pointer events level 2 specification states: > For hardware and platforms that do not support pressure, the value MUST be 0.5 when in the active buttons state and 0 otherwise. https://www.w3.org/TR/2019/REC-pointerevents2-20190404/ However, WebKit sets 0 as pressure in such case. The behavior is confirmed with: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_2) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/13.0.4 Safari/605.1.15 Apparently, WebKit refers to pressure property of NSEvent. The documentation states: > For input devices that aren’t pressure-sensitive, the value is either 0.0 or 1.0. https://developer.apple.com/documentation/appkit/nsevent/1534543-pressure 1606813 1 webkit-bug-importer 2020-01-14 22:52:39 -0800 <rdar://problem/58596321> 1608644 2 388210 nekomanma 2020-01-20 00:26:50 -0800 Created attachment 388210 Patch 1608795 3 388210 darin 2020-01-20 10:14:25 -0800 Comment on attachment 388210 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=388210&action=review > Source/WebCore/dom/PointerEvent.cpp:115 > + , m_pressure(mouseEvent.buttons() ? 0.5 : 0) I am not 100% sure this is right. The specification calls for all pointerup events having pressure of 0. Are we guaranteed that buttons() will be 0 for *all* pointerup events? The test case covers only a simple case of a mouse with a single button. > LayoutTests/ChangeLog:11 > + * pointerevents/mouse/pointer-button-and-buttons-expected.txt: Removed. > + * pointerevents/mouse/pointer-button-and-buttons.html: Removed. > + * pointerevents/mouse/pointer-button-buttons-and-pressure-expected.txt: Added. > + * pointerevents/mouse/pointer-button-buttons-and-pressure.html: Added. Why are we renaming this test case? 1608796 4 darin 2020-01-20 10:17:20 -0800 (In reply to Akihiko Odaki from comment #0) > Apparently, WebKit refers to pressure property of NSEvent. The documentation > states: > > For input devices that aren’t pressure-sensitive, the value is either 0.0 or 1.0. > https://developer.apple.com/documentation/appkit/nsevent/1534543-pressure Aside from the bug fixed here, this implies that on *Cocoa* platforms, for non-pressure-sensitive devices, WebKit is setting 1.0 as the pressure, where the standard instead requests/requires that we set it to 0.5. May be challenging to fix if NSEvent does not tell us whether the device is pressure-sensitive. It would be good if someone investigated this and created tests. 1608871 5 graouts 2020-01-20 14:57:08 -0800 (In reply to Darin Adler from comment #3) > Comment on attachment 388210 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=388210&action=review > > > Source/WebCore/dom/PointerEvent.cpp:115 > > + , m_pressure(mouseEvent.buttons() ? 0.5 : 0) > > I am not 100% sure this is right. The specification calls for all pointerup > events having pressure of 0. Are we guaranteed that buttons() will be 0 for > *all* pointerup events? The test case covers only a simple case of a mouse > with a single button. Maybe it would be better to use the provided event type to set the pressure value. 1609450 6 nekomanma 2020-01-21 21:25:38 -0800 (In reply to Darin Adler from comment #3) > Comment on attachment 388210 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=388210&action=review > > > Source/WebCore/dom/PointerEvent.cpp:115 > > + , m_pressure(mouseEvent.buttons() ? 0.5 : 0) > > I am not 100% sure this is right. The specification calls for all pointerup > events having pressure of 0. Are we guaranteed that buttons() will be 0 for > *all* pointerup events? The test case covers only a simple case of a mouse > with a single button. Yes, the specification states: > A user agent MUST fire a pointer event named pointerup when a pointer leaves the active buttons state. > > > LayoutTests/ChangeLog:11 > > + * pointerevents/mouse/pointer-button-and-buttons-expected.txt: Removed. > > + * pointerevents/mouse/pointer-button-and-buttons.html: Removed. > > + * pointerevents/mouse/pointer-button-buttons-and-pressure-expected.txt: Added. > > + * pointerevents/mouse/pointer-button-buttons-and-pressure.html: Added. > > Why are we renaming this test case? Because it is now testing pressure property too. The name simply represents tested properties. (In reply to Darin Adler from comment #4) > (In reply to Akihiko Odaki from comment #0) > > Apparently, WebKit refers to pressure property of NSEvent. The documentation > > states: > > > For input devices that aren’t pressure-sensitive, the value is either 0.0 or 1.0. > > https://developer.apple.com/documentation/appkit/nsevent/1534543-pressure > > Aside from the bug fixed here, this implies that on *Cocoa* platforms, for > non-pressure-sensitive devices, WebKit is setting 1.0 as the pressure, where > the standard instead requests/requires that we set it to 0.5. May be > challenging to fix if NSEvent does not tell us whether the device is > pressure-sensitive. It would be good if someone investigated this and > created tests. I wrote that statement before carefully looking at the code and it is misleading. In reality, pressure property of NSEvent is completely ignored because MouseEvent provided to PointerEvent to create a new instance does not have pressure information. (In reply to Antoine Quint from comment #5) > (In reply to Darin Adler from comment #3) > > Comment on attachment 388210 [details] > > Patch > > > > View in context: > > https://bugs.webkit.org/attachment.cgi?id=388210&action=review > > > > > Source/WebCore/dom/PointerEvent.cpp:115 > > > + , m_pressure(mouseEvent.buttons() ? 0.5 : 0) > > > > I am not 100% sure this is right. The specification calls for all pointerup > > events having pressure of 0. Are we guaranteed that buttons() will be 0 for > > *all* pointerup events? The test case covers only a simple case of a mouse > > with a single button. > > Maybe it would be better to use the provided event type to set the pressure > value. The specification states it is guaranteed, so I do not think it makes difference. We can still add an assertion, but such an assertion should cover all (including one generated by pressure-sensitive device) PointerEvents and I think it is out of scope of this bug. 388210 2020-01-20 00:26:50 -0800 2020-01-20 10:14:25 -0800 Patch bug-206216-20200120172649.patch text/plain 5697 nekomanma SW5kZXg6IFNvdXJjZS9XZWJDb3JlL0NoYW5nZUxvZwo9PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBTb3VyY2UvV2Vi Q29yZS9DaGFuZ2VMb2cJKHJldmlzaW9uIDI1NDgxMSkKKysrIFNvdXJjZS9XZWJDb3JlL0NoYW5n ZUxvZwkod29ya2luZyBjb3B5KQpAQCAtMSwzICsxLDE0IEBACisyMDIwLTAxLTIwICBBa2loaWtv IE9kYWtpICA8bmVrb21hbm1hQHBpeGl2LmNvbT4KKworICAgICAgICBQb2ludGVyRXZlbnQgaGFz IDAgYXMgcHJlc3N1cmUgd2hlbiB0aGUgaW5wdXQgZGV2aWNlIGlzIG5vdCBwcmVzc3VyZS1zZW5z aXRpdmUKKyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTIw NjIxNgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgIFRl c3Q6IHBvaW50ZXJldmVudHMvbW91c2UvcG9pbnRlci1idXR0b24tYnV0dG9ucy1hbmQtcHJlc3N1 cmUuaHRtbAorCisgICAgICAgICogZG9tL1BvaW50ZXJFdmVudC5jcHA6CisKIDIwMjAtMDEtMjAg IFRha2FzaGkgS29tb3JpICA8VGFrYXNoaS5Lb21vcmlAc29ueS5jb20+CiAKICAgICAgICAgS2V5 ZWREZWNvZGVyR2VuZXJpYyBjcmFzaGVzIHdoZW4gaXQgYWNjZXNzZXMgZGF0YSB3aXRoIG5vbi1l eGlzdGluZyBrZXkKSW5kZXg6IFNvdXJjZS9XZWJDb3JlL2RvbS9Qb2ludGVyRXZlbnQuY3BwCj09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT0KLS0tIFNvdXJjZS9XZWJDb3JlL2RvbS9Qb2ludGVyRXZlbnQuY3BwCShyZXZpc2lv biAyNTQ4MDIpCisrKyBTb3VyY2UvV2ViQ29yZS9kb20vUG9pbnRlckV2ZW50LmNwcAkod29ya2lu ZyBjb3B5KQpAQCAtMTEyLDYgKzExMiw3IEBAIFBvaW50ZXJFdmVudDo6UG9pbnRlckV2ZW50KGNv bnN0IEF0b21TdHIKIAogUG9pbnRlckV2ZW50OjpQb2ludGVyRXZlbnQoY29uc3QgQXRvbVN0cmlu ZyYgdHlwZSwgc2hvcnQgYnV0dG9uLCBjb25zdCBNb3VzZUV2ZW50JiBtb3VzZUV2ZW50KQogICAg IDogTW91c2VFdmVudCh0eXBlLCB0eXBlQ2FuQnViYmxlKHR5cGUpLCB0eXBlSXNDYW5jZWxhYmxl KHR5cGUpLCB0eXBlSXNDb21wb3NlZCh0eXBlKSwgbW91c2VFdmVudC52aWV3KCksIG1vdXNlRXZl bnQuZGV0YWlsKCksIG1vdXNlRXZlbnQuc2NyZWVuTG9jYXRpb24oKSwgeyBtb3VzZUV2ZW50LmNs aWVudFgoKSwgbW91c2VFdmVudC5jbGllbnRZKCkgfSwgbW91c2VFdmVudC5tb2RpZmllcktleXMo KSwgYnV0dG9uLCBtb3VzZUV2ZW50LmJ1dHRvbnMoKSwgbW91c2VFdmVudC5zeW50aGV0aWNDbGlj a1R5cGUoKSwgbW91c2VFdmVudC5yZWxhdGVkVGFyZ2V0KCkpCisgICAgLCBtX3ByZXNzdXJlKG1v dXNlRXZlbnQuYnV0dG9ucygpID8gMC41IDogMCkKICAgICAsIG1faXNQcmltYXJ5KHRydWUpCiB7 CiB9CkluZGV4OiBMYXlvdXRUZXN0cy9DaGFuZ2VMb2cKPT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gTGF5b3V0VGVz dHMvQ2hhbmdlTG9nCShyZXZpc2lvbiAyNTQ4MDIpCisrKyBMYXlvdXRUZXN0cy9DaGFuZ2VMb2cJ KHdvcmtpbmcgY29weSkKQEAgLTEsMyArMSwxNSBAQAorMjAyMC0wMS0yMCAgQWtpaGlrbyBPZGFr aSAgPG5la29tYW5tYUBwaXhpdi5jb20+CisKKyAgICAgICAgUG9pbnRlckV2ZW50IGhhcyAwIGFz IHByZXNzdXJlIHdoZW4gdGhlIGlucHV0IGRldmljZSBpcyBub3QgcHJlc3N1cmUtc2Vuc2l0aXZl CisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD0yMDYyMTYK KworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICAqIHBvaW50 ZXJldmVudHMvbW91c2UvcG9pbnRlci1idXR0b24tYW5kLWJ1dHRvbnMtZXhwZWN0ZWQudHh0OiBS ZW1vdmVkLgorICAgICAgICAqIHBvaW50ZXJldmVudHMvbW91c2UvcG9pbnRlci1idXR0b24tYW5k LWJ1dHRvbnMuaHRtbDogUmVtb3ZlZC4KKyAgICAgICAgKiBwb2ludGVyZXZlbnRzL21vdXNlL3Bv aW50ZXItYnV0dG9uLWJ1dHRvbnMtYW5kLXByZXNzdXJlLWV4cGVjdGVkLnR4dDogQWRkZWQuCisg ICAgICAgICogcG9pbnRlcmV2ZW50cy9tb3VzZS9wb2ludGVyLWJ1dHRvbi1idXR0b25zLWFuZC1w cmVzc3VyZS5odG1sOiBBZGRlZC4KKwogMjAyMC0wMS0xOCAgRGllZ28gUGlubyBHYXJjaWEgIDxk cGlub0BpZ2FsaWEuY29tPgogCiAgICAgICAgIFtXUEVdIEdhcmRlbmluZywgc3luY2hyb25pemUg d2l0aCBHVEsgVGVzdEV4cGVjdGF0aW9ucwpJbmRleDogTGF5b3V0VGVzdHMvcG9pbnRlcmV2ZW50 cy9tb3VzZS9wb2ludGVyLWJ1dHRvbi1hbmQtYnV0dG9ucy1leHBlY3RlZC50eHQKPT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PQotLS0gTGF5b3V0VGVzdHMvcG9pbnRlcmV2ZW50cy9tb3VzZS9wb2ludGVyLWJ1dHRvbi1hbmQt YnV0dG9ucy1leHBlY3RlZC50eHQJKHJldmlzaW9uIDI1NDgwMikKKysrIExheW91dFRlc3RzL3Bv aW50ZXJldmVudHMvbW91c2UvcG9pbnRlci1idXR0b24tYW5kLWJ1dHRvbnMtZXhwZWN0ZWQudHh0 CShub25leGlzdGVudCkKQEAgLTEsMyArMCwwIEBACi0KLVBBU1MgVGVzdGluZyB0aGF0IHRoZSAi YnV0dG9uIiBhbmQgImJ1dHRvbnMiIHByb3BlcnRpZXMgcmVwb3J0IHRoZSBleHBlY3RlZCB2YWx1 ZXMuIAotCkluZGV4OiBMYXlvdXRUZXN0cy9wb2ludGVyZXZlbnRzL21vdXNlL3BvaW50ZXItYnV0 dG9uLWFuZC1idXR0b25zLmh0bWwKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gTGF5b3V0VGVzdHMvcG9pbnRlcmV2 ZW50cy9tb3VzZS9wb2ludGVyLWJ1dHRvbi1hbmQtYnV0dG9ucy5odG1sCShyZXZpc2lvbiAyNTQ4 MDIpCisrKyBMYXlvdXRUZXN0cy9wb2ludGVyZXZlbnRzL21vdXNlL3BvaW50ZXItYnV0dG9uLWFu ZC1idXR0b25zLmh0bWwJKG5vbmV4aXN0ZW50KQpAQCAtMSwzMSArMCwwIEBACi08IURPQ1RZUEUg aHRtbD4KLTxodG1sPgotPGhlYWQ+Ci08bWV0YSBjaGFyc2V0PXV0Zi04PgotPC9oZWFkPgotPGJv ZHk+Ci08c2NyaXB0IHNyYz0iLi4vLi4vcmVzb3VyY2VzL3Rlc3RoYXJuZXNzLmpzIj48L3Njcmlw dD4KLTxzY3JpcHQgc3JjPSIuLi8uLi9yZXNvdXJjZXMvdGVzdGhhcm5lc3NyZXBvcnQuanMiPjwv c2NyaXB0PgotPHNjcmlwdCBzcmM9Ii4uL3V0aWxzLmpzIj48L3NjcmlwdD4KLTxzY3JpcHQ+Ci0K LSd1c2Ugc3RyaWN0JzsKLQotdGFyZ2V0X3Rlc3QoKHRhcmdldCwgdGVzdCkgPT4gewotICAgIGNv bnN0IGV2ZW50VHJhY2tlciA9IG5ldyBFdmVudFRyYWNrZXIodGFyZ2V0LCBbInBvaW50ZXJkb3du IiwgInBvaW50ZXJ1cCJdKTsKLQotICAgIC8vIENsaWNrIHdpdGhvdXQgYSBtb3ZlLgotICAgIGV2 ZW50U2VuZGVyLm1vdXNlTW92ZVRvKDUwLCA1MCk7Ci0gICAgZXZlbnRTZW5kZXIubW91c2VEb3du KCk7Ci0gICAgZXZlbnRTZW5kZXIubW91c2VVcCgpOwotCi0gICAgZXZlbnRUcmFja2VyLmFzc2Vy dE1hdGNoZXNFdmVudHMoWwotICAgICAgICB7IHR5cGUgOiAicG9pbnRlcmRvd24iLCB4OiA1MCwg eTogNTAsIGJ1dHRvbjogMCwgYnV0dG9uczogMSB9LAotICAgICAgICB7IHR5cGUgOiAicG9pbnRl cnVwIiwgeDogNTAsIHk6IDUwLCBidXR0b246IDAsIGJ1dHRvbnM6IDAgfSwKLSAgICBdKTsKLSAg ICB0ZXN0LmRvbmUoKTsKLX0sIGBUZXN0aW5nIHRoYXQgdGhlICJidXR0b24iIGFuZCAiYnV0dG9u cyIgcHJvcGVydGllcyByZXBvcnQgdGhlIGV4cGVjdGVkIHZhbHVlcy5gKTsKLQotPC9zY3JpcHQ+ Ci08L2JvZHk+Ci08L2h0bWw+ClwgTm8gbmV3bGluZSBhdCBlbmQgb2YgZmlsZQpJbmRleDogTGF5 b3V0VGVzdHMvcG9pbnRlcmV2ZW50cy9tb3VzZS9wb2ludGVyLWJ1dHRvbi1idXR0b25zLWFuZC1w cmVzc3VyZS1leHBlY3RlZC50eHQKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gTGF5b3V0VGVzdHMvcG9pbnRlcmV2 ZW50cy9tb3VzZS9wb2ludGVyLWJ1dHRvbi1idXR0b25zLWFuZC1wcmVzc3VyZS1leHBlY3RlZC50 eHQJKG5vbmV4aXN0ZW50KQorKysgTGF5b3V0VGVzdHMvcG9pbnRlcmV2ZW50cy9tb3VzZS9wb2lu dGVyLWJ1dHRvbi1idXR0b25zLWFuZC1wcmVzc3VyZS1leHBlY3RlZC50eHQJKHdvcmtpbmcgY29w eSkKQEAgLTAsMCArMSwzIEBACisKK1BBU1MgVGVzdGluZyB0aGF0IHRoZSAiYnV0dG9uIiwgImJ1 dHRvbnMiIGFuZCAicHJlc3N1cmUiIHByb3BlcnRpZXMgcmVwb3J0IHRoZSBleHBlY3RlZCB2YWx1 ZXMuIAorCkluZGV4OiBMYXlvdXRUZXN0cy9wb2ludGVyZXZlbnRzL21vdXNlL3BvaW50ZXItYnV0 dG9uLWJ1dHRvbnMtYW5kLXByZXNzdXJlLmh0bWwKPT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gTGF5b3V0VGVzdHMv cG9pbnRlcmV2ZW50cy9tb3VzZS9wb2ludGVyLWJ1dHRvbi1idXR0b25zLWFuZC1wcmVzc3VyZS5o dG1sCShub25leGlzdGVudCkKKysrIExheW91dFRlc3RzL3BvaW50ZXJldmVudHMvbW91c2UvcG9p bnRlci1idXR0b24tYnV0dG9ucy1hbmQtcHJlc3N1cmUuaHRtbAkod29ya2luZyBjb3B5KQpAQCAt MCwwICsxLDMxIEBACis8IURPQ1RZUEUgaHRtbD4KKzxodG1sPgorPGhlYWQ+Cis8bWV0YSBjaGFy c2V0PXV0Zi04PgorPC9oZWFkPgorPGJvZHk+Cis8c2NyaXB0IHNyYz0iLi4vLi4vcmVzb3VyY2Vz L3Rlc3RoYXJuZXNzLmpzIj48L3NjcmlwdD4KKzxzY3JpcHQgc3JjPSIuLi8uLi9yZXNvdXJjZXMv dGVzdGhhcm5lc3NyZXBvcnQuanMiPjwvc2NyaXB0PgorPHNjcmlwdCBzcmM9Ii4uL3V0aWxzLmpz Ij48L3NjcmlwdD4KKzxzY3JpcHQ+CisKKyd1c2Ugc3RyaWN0JzsKKwordGFyZ2V0X3Rlc3QoKHRh cmdldCwgdGVzdCkgPT4geworICAgIGNvbnN0IGV2ZW50VHJhY2tlciA9IG5ldyBFdmVudFRyYWNr ZXIodGFyZ2V0LCBbInBvaW50ZXJkb3duIiwgInBvaW50ZXJ1cCJdKTsKKworICAgIC8vIENsaWNr IHdpdGhvdXQgYSBtb3ZlLgorICAgIGV2ZW50U2VuZGVyLm1vdXNlTW92ZVRvKDUwLCA1MCk7Cisg ICAgZXZlbnRTZW5kZXIubW91c2VEb3duKCk7CisgICAgZXZlbnRTZW5kZXIubW91c2VVcCgpOwor CisgICAgZXZlbnRUcmFja2VyLmFzc2VydE1hdGNoZXNFdmVudHMoWworICAgICAgICB7IHR5cGUg OiAicG9pbnRlcmRvd24iLCB4OiA1MCwgeTogNTAsIGJ1dHRvbjogMCwgYnV0dG9uczogMSwgcHJl c3N1cmU6IDAuNSB9LAorICAgICAgICB7IHR5cGUgOiAicG9pbnRlcnVwIiwgeDogNTAsIHk6IDUw LCBidXR0b246IDAsIGJ1dHRvbnM6IDAsIHByZXNzdXJlOiAwIH0sCisgICAgXSk7CisgICAgdGVz dC5kb25lKCk7Cit9LCBgVGVzdGluZyB0aGF0IHRoZSAiYnV0dG9uIiwgImJ1dHRvbnMiIGFuZCAi cHJlc3N1cmUiIHByb3BlcnRpZXMgcmVwb3J0IHRoZSBleHBlY3RlZCB2YWx1ZXMuYCk7CisKKzwv c2NyaXB0PgorPC9ib2R5PgorPC9odG1sPgpcIE5vIG5ld2xpbmUgYXQgZW5kIG9mIGZpbGUK