111532 2013-03-06 00:03:47 -0800 Clarify the size of a document view that causes FrameView to emit the resize event. 2013-07-09 15:52:59 -0700 1 1 1 Unclassified WebKit UI Events 528+ (Nightly build) Unspecified Unspecified NEW P2 Normal --- 80242 107424 81209 111408 1 dongseong.hwang dongseong.hwang aelias allan.jensen bdakin jturcotte kenneth simon.fraser oldest_to_newest 848533 0 dongseong.hwang 2013-03-06 00:03:47 -0800 The spec DOM Level 2 Events states that the resize event occurs when a document view is resized. After r141053, FrameView::unscaledVisibleContentSize stands for a document view size. When turning on fixed layout, fixed layout size stands for a document view size. This patch applies the term of the spec to FrameView exactly. In addition, this removes delegatesScrolling check because delegatesScrolling is not related to the size of a document view. 848536 1 191665 dongseong.hwang 2013-03-06 00:07:03 -0800 Created attachment 191665 Patch 848539 2 dongseong.hwang 2013-03-06 00:09:01 -0800 This patch is the follow-up patch of Bug 80242. r141053 that the changelog refers to is Bug 107424. I tested only Qt and EFL ports, so I need various feedback. 848564 3 kenneth 2013-03-06 00:41:05 -0800 Why isnt this tested? 848570 4 191665 kenneth 2013-03-06 00:43:17 -0800 Comment on attachment 191665 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=191665&action=review > Source/WebCore/ChangeLog:13 > + After r141053, FrameView::unscaledVisibleContentSize stands for a document view > + size. When turning on fixed layout, fixed layout size stands for a document view > + size. This patch applies the term of the spec to FrameView exactly. view? Not viewport? I don't get this. With fixed layout, the fixed layout size is the size of the document, not the viewport > Source/WebCore/ChangeLog:19 > + > + Covered by existing tests: fast/events/resize-events.html > + Sorry, it must be early yet. I missed this one! :-) 848651 5 allan.jensen 2013-03-06 02:10:54 -0800 This is great, but can the subtle difference be tested. I guess it makes a difference on with fixedLayoutSize but disabled delegatedScrolling? 849134 6 aelias 2013-03-06 11:34:28 -0800 I think there's a bit of an open question about whether, in the presence of page scale, we should report to Javascript (1) the true user-visible scaled viewport, or (2) the fixed layout size (and a scroll offset properly clamped to go with it). The specs are ambiguous about this since page scale didn't exist back then. I'm inclined to favor (2) since that's more backwards compatible, and it looks like this patch is moving in that direction, so it's okay with me. Though in that case, to be more consistent, it seems like the viewport size sent to Javascript should also changed be unscaledVisibleContentSize. 855707 7 dongseong.hwang 2013-03-14 22:36:16 -0700 (In reply to comment #5) > This is great, but can the subtle difference be tested. I guess it makes a difference on with fixedLayoutSize but disabled delegatedScrolling? Thank you for review! Yes, you're right: the port with fixedLayoutSize but disabled delegatedScrolling now keeps fixedLayoutSize() as a m_lastViewportSize. the port can be tested using fast/events/resize-events.html also. I think this change goes to right direction. I think a document view is layout viewport that http://www.quirksmode.org/mobile/viewports2.html said, and when enabling fixed layout, layout viewport would be fixed layout size. (In reply to comment #6) > I think there's a bit of an open question about whether, in the presence of page scale, we should report to Javascript (1) the true user-visible scaled viewport, or (2) the fixed layout size (and a scroll offset properly clamped to go with it). The specs are ambiguous about this since page scale didn't exist back then. Thank you for good open questions! In my understanding based on http://www.quirksmode.org/mobile/viewports2.html 1. If a user uses window.innerWidth/Height, WebKit should return user-visible scaled viewport, which is visibleContentRect(IncludeScrollbars).size(). 2. If a user uses document.documentElement.clientWidth/Height, WebKit should return unscaledVisibleContentSize. And if enabling fixed layout, WebKit should return fixed layout size. > I'm inclined to favor (2) since that's more backwards compatible, and it looks like this patch is moving in that direction, so it's okay with me. Though in that case, to be more consistent, it seems like the viewport size sent to Javascript should also changed be unscaledVisibleContentSize. I manually test http://www.quirksmode.org/m/tests/widthtest.html on Qt WK2, both window.innerWidth and document.documentElement.clientWidth work well. If do you think we should change anything, could you explain more? 191665 2013-03-06 00:07:03 -0800 2013-07-09 07:00:59 -0700 Patch bug-111532-20130306170233.patch text/plain 3085 dongseong.hwang U3VidmVyc2lvbiBSZXZpc2lvbjogMTQ0ODQxCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9D aGFuZ2VMb2cgYi9Tb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXggODlhN2MwODliYWNlZTRm ZDFlZDg4NWFlNjdmZGQ3MTM3ZDI3OTkwNS4uZjEzOGMwYjAyNTczNDhhMWEyNzBmZTE3OTZmODVl NDRmZjg3YzBmMyAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCisrKyBiL1Nv dXJjZS9XZWJDb3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDI2IEBACisyMDEzLTAzLTA2ICBIdWFu ZyBEb25nc3VuZyAgPGx1eHRlbGxhQGNvbXBhbnkxMDAubmV0PgorCisgICAgICAgIENsYXJpZnkg dGhlIHNpemUgb2YgYSBkb2N1bWVudCB2aWV3IHRoYXQgY2F1c2VzIEZyYW1lVmlldyB0byBlbWl0 IHRoZSByZXNpemUgZXZlbnQuCisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3df YnVnLmNnaT9pZD0xMTE1MzIKKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4K KworICAgICAgICBUaGUgc3BlYyBET00gTGV2ZWwgMiBFdmVudHMgc3RhdGVzIHRoYXQgdGhlIHJl c2l6ZSBldmVudCBvY2N1cnMgd2hlbiBhIGRvY3VtZW50CisgICAgICAgIHZpZXcgaXMgcmVzaXpl ZC4KKworICAgICAgICBBZnRlciByMTQxMDUzLCBGcmFtZVZpZXc6OnVuc2NhbGVkVmlzaWJsZUNv bnRlbnRTaXplIHN0YW5kcyBmb3IgYSBkb2N1bWVudCB2aWV3CisgICAgICAgIHNpemUuIFdoZW4g dHVybmluZyBvbiBmaXhlZCBsYXlvdXQsIGZpeGVkIGxheW91dCBzaXplIHN0YW5kcyBmb3IgYSBk b2N1bWVudCB2aWV3CisgICAgICAgIHNpemUuIFRoaXMgcGF0Y2ggYXBwbGllcyB0aGUgdGVybSBv ZiB0aGUgc3BlYyB0byBGcmFtZVZpZXcgZXhhY3RseS4KKworICAgICAgICBJbiBhZGRpdGlvbiwg dGhpcyByZW1vdmVzIGRlbGVnYXRlc1Njcm9sbGluZyBjaGVjayBiZWNhdXNlIGRlbGVnYXRlc1Nj cm9sbGluZyBpcworICAgICAgICBub3QgcmVsYXRlZCB0byB0aGUgc2l6ZSBvZiBhIGRvY3VtZW50 IHZpZXcuCisKKyAgICAgICAgQ292ZXJlZCBieSBleGlzdGluZyB0ZXN0czogZmFzdC9ldmVudHMv cmVzaXplLWV2ZW50cy5odG1sCisKKyAgICAgICAgKiBwYWdlL0ZyYW1lVmlldy5jcHA6CisgICAg ICAgIChXZWJDb3JlOjpGcmFtZVZpZXc6OmxheW91dCk6CisgICAgICAgIChXZWJDb3JlOjpGcmFt ZVZpZXc6OnBlcmZvcm1Qb3N0TGF5b3V0VGFza3MpOgorCiAyMDEzLTAzLTA1ICBTaGF3biBTaW5n aCAgPHNoYXduc2luZ2hAY2hyb21pdW0ub3JnPgogCiAgICAgICAgIEZpeCBkZWZhdWx0IGJhY2tn cm91bmQgb2YgYSBkcmFnZ2VkIGltYWdlLgpkaWZmIC0tZ2l0IGEvU291cmNlL1dlYkNvcmUvcGFn ZS9GcmFtZVZpZXcuY3BwIGIvU291cmNlL1dlYkNvcmUvcGFnZS9GcmFtZVZpZXcuY3BwCmluZGV4 IGE4NGU1NmU4MTQ5N2Y0OGZmNjFiMDc4YjUxYzM5ODVkMDQ1MTA1M2YuLmQxNGMxOGIzMzliZTlh OGQ0ZTM2YWQ1ZDliMjg5MjdiZWI3NWQxYWUgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9XZWJDb3JlL3Bh Z2UvRnJhbWVWaWV3LmNwcAorKysgYi9Tb3VyY2UvV2ViQ29yZS9wYWdlL0ZyYW1lVmlldy5jcHAK QEAgLTExOTIsMTAgKzExOTIsMTAgQEAgdm9pZCBGcmFtZVZpZXc6OmxheW91dChib29sIGFsbG93 U3VidHJlZSkKIAogICAgICAgICAgICAgICAgICAgICBtX2ZpcnN0TGF5b3V0ID0gZmFsc2U7CiAg ICAgICAgICAgICAgICAgICAgIG1fZmlyc3RMYXlvdXRDYWxsYmFja1BlbmRpbmcgPSB0cnVlOwot ICAgICAgICAgICAgICAgICAgICBpZiAodXNlRml4ZWRMYXlvdXQoKSAmJiAhZml4ZWRMYXlvdXRT aXplKCkuaXNFbXB0eSgpICYmIGRlbGVnYXRlc1Njcm9sbGluZygpKQorICAgICAgICAgICAgICAg ICAgICBpZiAodXNlRml4ZWRMYXlvdXQoKSAmJiAhZml4ZWRMYXlvdXRTaXplKCkuaXNFbXB0eSgp KQogICAgICAgICAgICAgICAgICAgICAgICAgbV9sYXN0Vmlld3BvcnRTaXplID0gZml4ZWRMYXlv dXRTaXplKCk7CiAgICAgICAgICAgICAgICAgICAgIGVsc2UKLSAgICAgICAgICAgICAgICAgICAg ICAgIG1fbGFzdFZpZXdwb3J0U2l6ZSA9IHZpc2libGVDb250ZW50UmVjdChJbmNsdWRlU2Nyb2xs YmFycykuc2l6ZSgpOworICAgICAgICAgICAgICAgICAgICAgICAgbV9sYXN0Vmlld3BvcnRTaXpl ID0gdW5zY2FsZWRWaXNpYmxlQ29udGVudFNpemUoSW5jbHVkZVNjcm9sbGJhcnMpOwogICAgICAg ICAgICAgICAgICAgICBtX2xhc3Rab29tRmFjdG9yID0gcm9vdC0+c3R5bGUoKS0+em9vbSgpOwog CiAgICAgICAgICAgICAgICAgICAgIC8vIFNldCB0aGUgaW5pdGlhbCB2TW9kZSB0byBBbHdheXNP biBpZiB3ZSdyZSBhdXRvLgpAQCAtMjY0NSwxMCArMjY0NSwxMCBAQCB2b2lkIEZyYW1lVmlldzo6 cGVyZm9ybVBvc3RMYXlvdXRUYXNrcygpCiAKICAgICBpZiAocmVuZGVyVmlldyAmJiAhcmVuZGVy Vmlldy0+cHJpbnRpbmcoKSkgewogICAgICAgICBJbnRTaXplIGN1cnJlbnRTaXplOwotICAgICAg ICBpZiAodXNlRml4ZWRMYXlvdXQoKSAmJiAhZml4ZWRMYXlvdXRTaXplKCkuaXNFbXB0eSgpICYm IGRlbGVnYXRlc1Njcm9sbGluZygpKQorICAgICAgICBpZiAodXNlRml4ZWRMYXlvdXQoKSAmJiAh Zml4ZWRMYXlvdXRTaXplKCkuaXNFbXB0eSgpKQogICAgICAgICAgICAgY3VycmVudFNpemUgPSBm aXhlZExheW91dFNpemUoKTsKICAgICAgICAgZWxzZQotICAgICAgICAgICAgY3VycmVudFNpemUg PSB2aXNpYmxlQ29udGVudFJlY3QoSW5jbHVkZVNjcm9sbGJhcnMpLnNpemUoKTsKKyAgICAgICAg ICAgIGN1cnJlbnRTaXplID0gdW5zY2FsZWRWaXNpYmxlQ29udGVudFNpemUoSW5jbHVkZVNjcm9s bGJhcnMpOwogICAgICAgICBmbG9hdCBjdXJyZW50Wm9vbUZhY3RvciA9IHJlbmRlclZpZXctPnN0 eWxlKCktPnpvb20oKTsKICAgICAgICAgYm9vbCByZXNpemVkID0gIW1fZmlyc3RMYXlvdXQgJiYg KGN1cnJlbnRTaXplICE9IG1fbGFzdFZpZXdwb3J0U2l6ZSB8fCBjdXJyZW50Wm9vbUZhY3RvciAh PSBtX2xhc3Rab29tRmFjdG9yKTsKICAgICAgICAgbV9sYXN0Vmlld3BvcnRTaXplID0gY3VycmVu dFNpemU7Cg==