searching for reviewer(jnicol)
83488f147740483a4e59be8a0c4934303032d131: Bug 1485441 - Add diagnostics to GL compositor shader failures and avoid crashing release builds. r=jnicol
Nicolas Silva <nsilva@mozilla.com> - Thu, 23 Aug 2018 14:46:50 +0000 - rev 831024
Push 118868 by bmo:zjz@zjz.name at Fri, 24 Aug 2018 07:04:39 +0000
Bug 1485441 - Add diagnostics to GL compositor shader failures and avoid crashing release builds. r=jnicol Differential Revision: https://phabricator.services.mozilla.com/D4079
90c347a61bd56af575c2c812bac026e7ead65de2: Bug 1474466 - Convert rounded rect clips into paint geometry if the display item type support it. r=bas,jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Sun, 15 Jul 2018 20:26:46 +1200 - rev 818594
Push 116287 by cpeterson@mozilla.com at Sun, 15 Jul 2018 21:59:08 +0000
Bug 1474466 - Convert rounded rect clips into paint geometry if the display item type support it. r=bas,jnicol MozReview-Commit-ID: EkqgSSLufnp
6a228e47a3bda33d0e8c3a3d45a633960c6fd9bf: Bug 1473137 - Convert mOpacityIndices into a stack-based vector with preallocated storage to avoid allocations. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 03 Jul 2018 18:58:35 -0400 - rev 814051
Push 115070 by plingurar@mozilla.com at Wed, 04 Jul 2018 10:16:17 +0000
Bug 1473137 - Convert mOpacityIndices into a stack-based vector with preallocated storage to avoid allocations. r=jnicol MozReview-Commit-ID: Gi5z3YaH7kY
4ac82a817c402935031e762848f6617347ded9e0: Bug 1473137 - Convert mOpacityIndices into a stack-based vector with preallocated storage to avoid allocations. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 03 Jul 2018 18:58:35 -0400 - rev 813846
Push 115019 by mwoodrow@mozilla.com at Tue, 03 Jul 2018 22:59:07 +0000
Bug 1473137 - Convert mOpacityIndices into a stack-based vector with preallocated storage to avoid allocations. r?jnicol MozReview-Commit-ID: Gi5z3YaH7kY
4d1dccdf5a35f8ee1fdf121b98fb0630839cd5ce: Bug 1473137 - Convert mOpacityIndices into a stack-based vector with preallocated storage to avoid allocations. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 03 Jul 2018 18:42:42 -0400 - rev 813843
Push 115016 by mwoodrow@mozilla.com at Tue, 03 Jul 2018 22:43:02 +0000
Bug 1473137 - Convert mOpacityIndices into a stack-based vector with preallocated storage to avoid allocations. r?jnicol MozReview-Commit-ID: IS6ppVdit0M
0571c2da7c437eb965ebf0cb79a4d991e6d7b8c2: Bug 1460491 - Part 2: Only recompute visibility for items if they are newly added to this layer, or intersect one that changed. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Wed, 16 May 2018 16:02:37 +1200 - rev 795648
Push 110038 by bmo:alessio.placitelli@gmail.com at Wed, 16 May 2018 11:10:16 +0000
Bug 1460491 - Part 2: Only recompute visibility for items if they are newly added to this layer, or intersect one that changed. r=jnicol MozReview-Commit-ID: EAgvVQGsSPE
920d49a96e6ec78991466cb5e1850c81d39937f2: Bug 1460491 - Part 1: Split nsDisplayItem::mVisibleRect into two members, one for each stated purpose. Gets rid of the save/restore since we no longer modify the building rect. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Wed, 16 May 2018 15:56:32 +1200 - rev 795647
Push 110038 by bmo:alessio.placitelli@gmail.com at Wed, 16 May 2018 11:10:16 +0000
Bug 1460491 - Part 1: Split nsDisplayItem::mVisibleRect into two members, one for each stated purpose. Gets rid of the save/restore since we no longer modify the building rect. r=jnicol MozReview-Commit-ID: 5wcsSoZRN44
ab9f7030233e00f1b6ddec4064ed8712bd18d853: Bug 1460491 - Part 2: Only recompute visibility for items if they are newly added to this layer, or intersect one that changed. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Wed, 16 May 2018 16:02:37 +1200 - rev 795564
Push 110015 by mwoodrow@mozilla.com at Wed, 16 May 2018 04:15:10 +0000
Bug 1460491 - Part 2: Only recompute visibility for items if they are newly added to this layer, or intersect one that changed. r?jnicol MozReview-Commit-ID: EAgvVQGsSPE
24dea044d82b6d216c2c079d7c157a7877f626f1: Bug 1460491 - Part 1: Split nsDisplayItem::mVisibleRect into two members, one for each stated purpose. Gets rid of the save/restore since we no longer modify the building rect. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Wed, 16 May 2018 15:56:32 +1200 - rev 795563
Push 110015 by mwoodrow@mozilla.com at Wed, 16 May 2018 04:15:10 +0000
Bug 1460491 - Part 1: Split nsDisplayItem::mVisibleRect into two members, one for each stated purpose. Gets rid of the save/restore since we no longer modify the building rect. r?jnicol MozReview-Commit-ID: 5wcsSoZRN44
222772aa366b2dfaa4f1d60cb6e93a93cab53fb0: Bug 1460491 - Part 2: Only recompute visibility for items if they are newly added to this layer, or intersect one that changed. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Thu, 10 May 2018 16:22:41 +1200 - rev 793470
Push 109394 by mwoodrow@mozilla.com at Thu, 10 May 2018 04:23:18 +0000
Bug 1460491 - Part 2: Only recompute visibility for items if they are newly added to this layer, or intersect one that changed. r?jnicol MozReview-Commit-ID: FGFZqHcjHOZ
a0f84357e223a506d26d024c58a3a2056a53b81e: Bug 1460491 - Part 1: Split nsDisplayItem::mVisibleRect into two members, one for each stated purpose. Gets rid of the save/restore since we no longer modify the building rect. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Thu, 10 May 2018 16:22:39 +1200 - rev 793469
Push 109394 by mwoodrow@mozilla.com at Thu, 10 May 2018 04:23:18 +0000
Bug 1460491 - Part 1: Split nsDisplayItem::mVisibleRect into two members, one for each stated purpose. Gets rid of the save/restore since we no longer modify the building rect. r?jnicol MozReview-Commit-ID: 4ZmzOvDRp1i
fe859b4ec63cc8737ad26a1ca2bc2e7bfaa6dcdc: Bug 1440966 - Store optimized Layer in DisplayItemData as part of AddPaintedLayerFor. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Wed, 07 Mar 2018 15:42:43 +1300 - rev 764887
Push 101881 by bmo:snorp@snorp.net at Thu, 08 Mar 2018 15:11:59 +0000
Bug 1440966 - Store optimized Layer in DisplayItemData as part of AddPaintedLayerFor. r=jnicol
1a8e77cb2c21d4f42dd3e90fdcbece7a633b473b: Bug 1440966 - Store optimized Layer in DisplayItemData as part of AddPaintedLayerFor. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Wed, 07 Mar 2018 15:42:43 +1300 - rev 764106
Push 101664 by bmo:hsivonen@hsivonen.fi at Wed, 07 Mar 2018 07:45:15 +0000
Bug 1440966 - Store optimized Layer in DisplayItemData as part of AddPaintedLayerFor. r=jnicol
b3cf1b5c14cc373f90fda4f109e35bef5f69f84f: Bug 1440966 - Store optimized Layer in DisplayItemData as part of AddPaintedLayerFor. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Wed, 07 Mar 2018 15:42:43 +1300 - rev 764077
Push 101657 by nnethercote@mozilla.com at Wed, 07 Mar 2018 04:58:48 +0000
Bug 1440966 - Store optimized Layer in DisplayItemData as part of AddPaintedLayerFor. r=jnicol
58025bcb77cbc6505af35f4ac4e0cd8a0878841b: Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Fri, 02 Mar 2018 11:19:17 +1300 - rev 762419
Push 101169 by luca.greco@alcacoop.it at Fri, 02 Mar 2018 12:32:00 +0000
Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r=jnicol The display items are almost certainly gone from the cache at this point, so dereferencing them can take a while. This moves the DisplayItemData lookup, and the IsReused/HasMergedFrames checks into the ProcessDisplayItems pass (where we already use the items), and then just uses the AssignedDisplayItem entry for everything. MozReview-Commit-ID: 8udcE0bmyF3
4c62cca5f3ecb2ee93b4d50a4c9427b3fa8f8ccd: Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Mon, 19 Feb 2018 12:37:01 +1300 - rev 762418
Push 101169 by luca.greco@alcacoop.it at Fri, 02 Mar 2018 12:32:00 +0000
Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r=jnicol MozReview-Commit-ID: CfHL0Kqto6X
7b7026149e56e8af45061fb7769f83b06c3c7dd8: Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Thu, 15 Feb 2018 14:45:35 +1300 - rev 762416
Push 101169 by luca.greco@alcacoop.it at Fri, 02 Mar 2018 12:32:00 +0000
Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r=jnicol MozReview-Commit-ID: 4RvU35pBz3z
ef972081de6964e22c38b6bbd1455d7bcb282723: Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 13 Feb 2018 15:56:43 +1300 - rev 762415
Push 101169 by luca.greco@alcacoop.it at Fri, 02 Mar 2018 12:32:00 +0000
Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r=jnicol These two structs store very similar state (including duplicating the mask layer common clip count), and the former uses an expensive hashtable for lookups. This patch combines the two, and uses a vector of entries instead of the hashtable so we can do the cleanup pass. * * * [mq]: fix MozReview-Commit-ID: KamhbGAIqpD
8d4b4e8e9d048da096f52f5b319290b079a808b3: Bug 1440177 - Part 1: Don't call GetLayerState from BuildContainerLayerFor as it recurses into child display items to find the answer. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 13 Feb 2018 14:25:55 +1300 - rev 762414
Push 101169 by luca.greco@alcacoop.it at Fri, 02 Mar 2018 12:32:00 +0000
Bug 1440177 - Part 1: Don't call GetLayerState from BuildContainerLayerFor as it recurses into child display items to find the answer. r=jnicol MozReview-Commit-ID: 3A4OBC5L6Ps
a77d06b2cf03b3b8824130a1e32c0b09d3ec138c: Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Fri, 02 Mar 2018 10:06:13 +1300 - rev 762386
Push 101169 by luca.greco@alcacoop.it at Fri, 02 Mar 2018 12:32:00 +0000
Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r=jnicol The display items are almost certainly gone from the cache at this point, so dereferencing them can take a while. This moves the DisplayItemData lookup, and the IsReused/HasMergedFrames checks into the ProcessDisplayItems pass (where we already use the items), and then just uses the AssignedDisplayItem entry for everything. MozReview-Commit-ID: JrRshEyZncb
928770efc9a3dab8eedf0e97357bd475d67c724a: Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Mon, 19 Feb 2018 12:37:01 +1300 - rev 762385
Push 101169 by luca.greco@alcacoop.it at Fri, 02 Mar 2018 12:32:00 +0000
Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r=jnicol MozReview-Commit-ID: IHrVPutPSC5
717a696b17e07a91edf9b6cbf9be0a611aea1f40: Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Thu, 15 Feb 2018 14:45:35 +1300 - rev 762383
Push 101169 by luca.greco@alcacoop.it at Fri, 02 Mar 2018 12:32:00 +0000
Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r=jnicol MozReview-Commit-ID: 1LIxNSVHHkp
0840994846bf50d1278920a8a677603f29e26582: Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 13 Feb 2018 15:56:43 +1300 - rev 762382
Push 101169 by luca.greco@alcacoop.it at Fri, 02 Mar 2018 12:32:00 +0000
Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r=jnicol These two structs store very similar state (including duplicating the mask layer common clip count), and the former uses an expensive hashtable for lookups. This patch combines the two, and uses a vector of entries instead of the hashtable so we can do the cleanup pass. * * * [mq]: fix MozReview-Commit-ID: KamhbGAIqpD
b1ba6c67264bce7b38de085b41183a5029023127: Bug 1440177 - Part 1: Don't call GetLayerState from BuildContainerLayerFor as it recurses into child display items to find the answer. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 13 Feb 2018 14:25:55 +1300 - rev 762381
Push 101169 by luca.greco@alcacoop.it at Fri, 02 Mar 2018 12:32:00 +0000
Bug 1440177 - Part 1: Don't call GetLayerState from BuildContainerLayerFor as it recurses into child display items to find the answer. r=jnicol MozReview-Commit-ID: 3A4OBC5L6Ps
f119175e79ca494d12ec5ba49e20b12c9e03ada4: Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Fri, 02 Mar 2018 11:19:17 +1300 - rev 762225
Push 101106 by mwoodrow@mozilla.com at Thu, 01 Mar 2018 22:19:48 +0000
Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r?jnicol The display items are almost certainly gone from the cache at this point, so dereferencing them can take a while. This moves the DisplayItemData lookup, and the IsReused/HasMergedFrames checks into the ProcessDisplayItems pass (where we already use the items), and then just uses the AssignedDisplayItem entry for everything. MozReview-Commit-ID: 8udcE0bmyF3
5190523dd6cf20ccae2d6621204e23e8d23c3d4c: Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Mon, 19 Feb 2018 12:37:01 +1300 - rev 762224
Push 101106 by mwoodrow@mozilla.com at Thu, 01 Mar 2018 22:19:48 +0000
Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r?jnicol MozReview-Commit-ID: CfHL0Kqto6X
a136e473da15901a9683d5ccf145ade9dc987eba: Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Thu, 15 Feb 2018 14:45:35 +1300 - rev 762222
Push 101106 by mwoodrow@mozilla.com at Thu, 01 Mar 2018 22:19:48 +0000
Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r?jnicol MozReview-Commit-ID: 4RvU35pBz3z
72b4a74dc82abc29eb94bdc464498f22d99707bc: Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 13 Feb 2018 15:56:43 +1300 - rev 762221
Push 101106 by mwoodrow@mozilla.com at Thu, 01 Mar 2018 22:19:48 +0000
Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r?jnicol These two structs store very similar state (including duplicating the mask layer common clip count), and the former uses an expensive hashtable for lookups. This patch combines the two, and uses a vector of entries instead of the hashtable so we can do the cleanup pass. * * * [mq]: fix MozReview-Commit-ID: KamhbGAIqpD
9dea9d905d9c493fcd3c4c4f9ca1e4a5616144f1: Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Fri, 02 Mar 2018 10:06:13 +1300 - rev 762118
Push 101090 by mwoodrow@mozilla.com at Thu, 01 Mar 2018 21:06:37 +0000
Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r?jnicol The display items are almost certainly gone from the cache at this point, so dereferencing them can take a while. This moves the DisplayItemData lookup, and the IsReused/HasMergedFrames checks into the ProcessDisplayItems pass (where we already use the items), and then just uses the AssignedDisplayItem entry for everything. MozReview-Commit-ID: JrRshEyZncb
4dd730074a96a91598ea7f349e3bf76fc5b69a2c: Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Mon, 19 Feb 2018 12:37:01 +1300 - rev 762117
Push 101090 by mwoodrow@mozilla.com at Thu, 01 Mar 2018 21:06:37 +0000
Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r?jnicol MozReview-Commit-ID: IHrVPutPSC5
8cc609ffe68636c40c71951d7bbaf6a9a59da886: Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Thu, 15 Feb 2018 14:45:35 +1300 - rev 762115
Push 101090 by mwoodrow@mozilla.com at Thu, 01 Mar 2018 21:06:37 +0000
Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r?jnicol MozReview-Commit-ID: 1LIxNSVHHkp
aa5a1c6043e5a1682bf77e9a2ddabefbc809c5b9: Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 13 Feb 2018 15:56:43 +1300 - rev 762114
Push 101090 by mwoodrow@mozilla.com at Thu, 01 Mar 2018 21:06:37 +0000
Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r?jnicol These two structs store very similar state (including duplicating the mask layer common clip count), and the former uses an expensive hashtable for lookups. This patch combines the two, and uses a vector of entries instead of the hashtable so we can do the cleanup pass. * * * [mq]: fix MozReview-Commit-ID: KamhbGAIqpD
d63c5a6c13ae9cb97b108b27f880ad190ea76af6: Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Thu, 01 Mar 2018 11:42:27 +1300 - rev 762010
Push 101062 by bmo:eoger@fastmail.com at Thu, 01 Mar 2018 18:27:01 +0000
Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r=jnicol The display items are almost certainly gone from the cache at this point, so dereferencing them can take a while. This moves the DisplayItemData lookup, and the IsReused/HasMergedFrames checks into the ProcessDisplayItems pass (where we already use the items), and then just uses the AssignedDisplayItem entry for everything. MozReview-Commit-ID: 3NibNGSVsax
d22462ac7ee164e320ec72c71f7b732eaf52b993: Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Mon, 19 Feb 2018 12:37:01 +1300 - rev 762009
Push 101062 by bmo:eoger@fastmail.com at Thu, 01 Mar 2018 18:27:01 +0000
Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r=jnicol MozReview-Commit-ID: IJrH0MLLPdQ
f929871841727b9823667b031cd1fe11c7c3a5f3: Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Thu, 15 Feb 2018 14:45:35 +1300 - rev 762007
Push 101062 by bmo:eoger@fastmail.com at Thu, 01 Mar 2018 18:27:01 +0000
Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r=jnicol MozReview-Commit-ID: JnxY10oiffD
e6808a7dda3e733fb8b321916c81fb3ad1c3de88: Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 13 Feb 2018 15:56:43 +1300 - rev 762006
Push 101062 by bmo:eoger@fastmail.com at Thu, 01 Mar 2018 18:27:01 +0000
Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r=jnicol These two structs store very similar state (including duplicating the mask layer common clip count), and the former uses an expensive hashtable for lookups. This patch combines the two, and uses a vector of entries instead of the hashtable so we can do the cleanup pass. MozReview-Commit-ID: 3gKbB3OBIDd
ce353b632a3108600e45cd1c21786ea9fb21be7c: Bug 1440177 - Part 1: Don't call GetLayerState from BuildContainerLayerFor as it recurses into child display items to find the answer. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 13 Feb 2018 14:25:55 +1300 - rev 762005
Push 101062 by bmo:eoger@fastmail.com at Thu, 01 Mar 2018 18:27:01 +0000
Bug 1440177 - Part 1: Don't call GetLayerState from BuildContainerLayerFor as it recurses into child display items to find the answer. r=jnicol MozReview-Commit-ID: 3A4OBC5L6Ps
bf9c377398ac25bd058f4cde060bd55141de6f31: Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Thu, 01 Mar 2018 11:42:27 +1300 - rev 761378
Push 100947 by mwoodrow@mozilla.com at Wed, 28 Feb 2018 23:56:10 +0000
Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r?jnicol The display items are almost certainly gone from the cache at this point, so dereferencing them can take a while. This moves the DisplayItemData lookup, and the IsReused/HasMergedFrames checks into the ProcessDisplayItems pass (where we already use the items), and then just uses the AssignedDisplayItem entry for everything. MozReview-Commit-ID: 3NibNGSVsax
cc6a0f2123a35e90740fe7a9e9d883a17d6dac5e: Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Mon, 19 Feb 2018 12:37:01 +1300 - rev 761377
Push 100947 by mwoodrow@mozilla.com at Wed, 28 Feb 2018 23:56:10 +0000
Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r?jnicol MozReview-Commit-ID: IJrH0MLLPdQ
97422dfe0632c9b7c7b4bc4c88acca9bda8875db: Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Thu, 15 Feb 2018 14:45:35 +1300 - rev 761375
Push 100947 by mwoodrow@mozilla.com at Wed, 28 Feb 2018 23:56:10 +0000
Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r?jnicol MozReview-Commit-ID: JnxY10oiffD
ade40e075474d95a42f8859cb3d20a160ae25bc8: Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 13 Feb 2018 15:56:43 +1300 - rev 761374
Push 100947 by mwoodrow@mozilla.com at Wed, 28 Feb 2018 23:56:10 +0000
Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r?jnicol These two structs store very similar state (including duplicating the mask layer common clip count), and the former uses an expensive hashtable for lookups. This patch combines the two, and uses a vector of entries instead of the hashtable so we can do the cleanup pass. MozReview-Commit-ID: 3gKbB3OBIDd
0e1ca79b33cf7de2562308d7dab90f481f8247c1: Bug 1440177 - Part 1: Don't call GetLayerState from BuildContainerLayerFor as it recurses into child display items to find the answer. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 13 Feb 2018 14:25:55 +1300 - rev 761373
Push 100947 by mwoodrow@mozilla.com at Wed, 28 Feb 2018 23:56:10 +0000
Bug 1440177 - Part 1: Don't call GetLayerState from BuildContainerLayerFor as it recurses into child display items to find the answer. r?jnicol MozReview-Commit-ID: 3A4OBC5L6Ps
35b0d12a64297e50f791c6144ee23739dbb0cd6a: Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Thu, 22 Feb 2018 14:32:35 +1300 - rev 758222
Push 99985 by mwoodrow@mozilla.com at Thu, 22 Feb 2018 01:35:14 +0000
Bug 1440177 - Part 6: Don't dereference display items during AddPaintedDisplayItem for the LAYER_NONE case. r?jnicol The display items are almost certainly gone from the cache at this point, so dereferencing them can take a while. This moves the DisplayItemData lookup, and the IsReused/HasMergedFrames checks into the ProcessDisplayItems pass (where we already use the items), and then just uses the AssignedDisplayItem entry for everything. MozReview-Commit-ID: 5t7CZ3Eu4m3
df08878dcf245adb5dcd107defa5618233b1cc54: Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Mon, 19 Feb 2018 12:37:01 +1300 - rev 758221
Push 99985 by mwoodrow@mozilla.com at Thu, 22 Feb 2018 01:35:14 +0000
Bug 1440177 - Part 5: Don't call GetDisplayItemDataForManager in AddPaintedDisplayItem since we already have it passed in as a parameter. r?jnicol MozReview-Commit-ID: IkKmPAyXK9E
cf69789b9887238f15eee613c5b0f3c467267126: Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Thu, 15 Feb 2018 14:45:35 +1300 - rev 758219
Push 99985 by mwoodrow@mozilla.com at Thu, 22 Feb 2018 01:35:14 +0000
Bug 1440177 - Part 3: Preallocate a small number of PaintedLayerData objects and only resize the mPaintedLayerDataStack once. r?jnicol MozReview-Commit-ID: 9yTod0AVJKm
ff7bcf7304390dfc4af342b0ff22f57adb4dd8a6: Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 13 Feb 2018 15:56:43 +1300 - rev 758218
Push 99985 by mwoodrow@mozilla.com at Thu, 22 Feb 2018 01:35:14 +0000
Bug 1440177 - Part 2: Combine PaintedLayerItemsEntry and PaintedDisplayItemLayerUserData into a single struct. r?jnicol These two structs store very similar state (including duplicating the mask layer common clip count), and the former uses an expensive hashtable for lookups. This patch combines the two, and uses a vector of entries instead of the hashtable so we can do the cleanup pass. MozReview-Commit-ID: IYdhFlKC9s6
938d227fd70b687df441e808c7b8260801b6b5c0: Bug 1440177 - Part 1: Don't call GetLayerState from BuildContainerLayerFor as it recurses into child display items to find the answer. r?jnicol draft
Matt Woodrow <mwoodrow@mozilla.com> - Tue, 13 Feb 2018 14:25:55 +1300 - rev 758217
Push 99985 by mwoodrow@mozilla.com at Thu, 22 Feb 2018 01:35:14 +0000
Bug 1440177 - Part 1: Don't call GetLayerState from BuildContainerLayerFor as it recurses into child display items to find the answer. r?jnicol MozReview-Commit-ID: LvNPrk4Xrxb
4aa9b74724b8326d5f13cad627ea1fddc9bcaf47: Bug 1435649 - Don't deference the display item when attempting the early return in ComputeGeometryChangeForItem to avoid a cache miss. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Mon, 12 Feb 2018 19:03:09 +1300 - rev 753803
Push 98685 by mak77@bonardo.net at Mon, 12 Feb 2018 13:47:25 +0000
Bug 1435649 - Don't deference the display item when attempting the early return in ComputeGeometryChangeForItem to avoid a cache miss. r=jnicol
f840f8d4b7b13206a25e2e69702b966f2674a612: Bug 1435650 - Cache the offset from AGR to reference frame in ProcessDisplayItems. r=jnicol
Matt Woodrow <mwoodrow@mozilla.com> - Mon, 12 Feb 2018 19:03:09 +1300 - rev 753802
Push 98685 by mak77@bonardo.net at Mon, 12 Feb 2018 13:47:25 +0000
Bug 1435650 - Cache the offset from AGR to reference frame in ProcessDisplayItems. r=jnicol
efc5b1e03ead10625a44444f24885e10cc9e229c: Bug 1416015 - Ensure SurfaceTexture desctruction happens correctly r=jnicol
James Willcox <snorp@snorp.net> - Wed, 15 Nov 2017 15:20:26 -0600 - rev 701628
Push 90225 by bmo:kinetik@flim.org at Wed, 22 Nov 2017 02:25:32 +0000
Bug 1416015 - Ensure SurfaceTexture desctruction happens correctly r=jnicol MozReview-Commit-ID: I4X1jQQC7ry