1675128 - Convert ReflowInput'::mComputedOffsets to logical coordinates

Status: RESOLVED FIXED

(Core :: Layout, task)

Description

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

The approach of this bug should be similar to bug 1674450 comment 0.

Comment 1

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Bug 1675128 Part 1 - Add WritingMode argument to ReflowInput::SetComputedLogicalOffsets().

Comment 2

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Bug 1675128 Part 2 - Add WritingMode argument to ReflowInput::ComputedLogicalOffsets().

Depends on D96125

Comment 3

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Bug 1675128 Part 3 - Make ReflowInput::ComputeRelativeOffsets() return the offset in LogicalMargin.

Also, convert its implementation to use more logical APIs.

Depends on D96126

Comment 4

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Bug 1675128 Part 4 - Remove ReflowInput::ComputedPhysicalOffsets() that returns a writable-reference.

While I'm at this, I also convert the computation in RecomputePosition()
into logical coordinates.

Depends on D96127

Comment 5

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Bug 1675128 Part 5 - Store ReflowInput::mComputedOffsets in LogicalMargin.

Depends on D96128

Comment 6

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Bug 1675128 Part 6 - Remove nsStylePosition::OffsetHasPercent().

Remove it to encourage people to use existing API to get logical side.

Depends on D96129

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/1b44b1d83458
Part 1 - Add WritingMode argument to ReflowInput::SetComputedLogicalOffsets(). r=emilio
https://hg.mozilla.org/integration/autoland/rev/78261df44adc
Part 2 - Add WritingMode argument to ReflowInput::ComputedLogicalOffsets(). r=emilio
https://hg.mozilla.org/integration/autoland/rev/276db1fd9cf2
Part 3 - Make ReflowInput::ComputeRelativeOffsets() return the offset in LogicalMargin. r=emilio
https://hg.mozilla.org/integration/autoland/rev/576b43399318
Part 4 - Remove ReflowInput::ComputedPhysicalOffsets() that returns a writable-reference. r=emilio
https://hg.mozilla.org/integration/autoland/rev/09693271e582
Part 5 - Store ReflowInput::mComputedOffsets in LogicalMargin. r=emilio
https://hg.mozilla.org/integration/autoland/rev/0b430b271213
Part 6 - Remove nsStylePosition::OffsetHasPercent(). r=emilio

Comment 8

[Narcis Beleuzu [:NarcisB]]

https://hg.mozilla.org/mozilla-central/rev/1b44b1d83458
https://hg.mozilla.org/mozilla-central/rev/78261df44adc
https://hg.mozilla.org/mozilla-central/rev/276db1fd9cf2
https://hg.mozilla.org/mozilla-central/rev/576b43399318
https://hg.mozilla.org/mozilla-central/rev/09693271e582
https://hg.mozilla.org/mozilla-central/rev/0b430b271213

Comment 9

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Bug 1675128 Part 1 - Change available block and inline size to store in LogicalSize.

I put the private member section at the end of ReflowInput's
declarations. In later patches, other members are going to live there
when they're converted to logical coordinates.

Remove unused AvailableWidth() and AvailableHeight() that return
writable-references.

I didn't port the comment section verbatim, but revise them according to
the best of my knowledge.

Comment 10

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Bug 1675128 Part 2 - Reimplement SetComputedISize() and SetComputedBSize().

We can use them to implement SetComputedWidth() and SetComputedHeight()
rather than the other way around.

The logic involving whether to call InitResizeFlags() for viewport frame
can be simplified. That is, we don't want to call InitResizeFlags() for
viewport frame in SetComputedISize().

Remove unused ComputedWidth() and ComputedHeight() that return
writable-references.

Depends on D99499

Comment 11

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Bug 1675128 Part 3 - Change the computed inline-size and block-size to store in LogicalSize.

Delete the comment describing that replaced frames should use their
intrinsic size if they have NS_UNCONSTRAINEDSIZE in mComputedWidth and
mComputedHeight, because we always have a definite size computed by
ComputeSizeWithIntrinsicDimensions().

Depends on D99500

Comment 12

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Bug 1675128 Part 4 - Change computed min-size and max-size to store in LogicalSize.

Remove unused ComputedMinWidth(), ComputedMaxWidth(),
ComputedMinHeight(), and ComputedMaxHeight() that return
writable-references.

Depends on D99501

Comment 13

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Bug 1675128 Part 5 - Revise comments in ReflowInput.h.

s/aContainingBlockSize/aCBSize/ for ComputeMinMaxValues() because that's
the parameter name used in definition.

Depends on D99502

Comment 14

[Phabricator Automation]

Comment on attachment 9192691 [details]
Bug 1675128 Part 1 - Change available block and inline size to store in LogicalSize.

Revision D99499 was moved to bug 1681849. Setting attachment 9192691 [details] to obsolete.

Comment 15

[Phabricator Automation]

Comment on attachment 9192692 [details]
Bug 1675128 Part 2 - Reimplement SetComputedISize() and SetComputedBSize().

Revision D99500 was moved to bug 1681849. Setting attachment 9192692 [details] to obsolete.

Comment 16

[Phabricator Automation]

Comment on attachment 9192693 [details]
Bug 1675128 Part 3 - Change the computed inline-size and block-size to store in LogicalSize.

Revision D99501 was moved to bug 1681849. Setting attachment 9192693 [details] to obsolete.

Comment 17

[Phabricator Automation]

Comment on attachment 9192694 [details]
Bug 1675128 Part 4 - Change computed min-size and max-size to store in LogicalSize.

Revision D99502 was moved to bug 1681849. Setting attachment 9192694 [details] to obsolete.

Comment 18

[Phabricator Automation]

Comment on attachment 9192695 [details]
Bug 1675128 Part 5 - Revise comments in ReflowInput.h.

Revision D99503 was moved to bug 1681849. Setting attachment 9192695 [details] to obsolete.