Bug 42186 – Failing 2d.path.arcTo.ensuresubpath.* philip canvas tests

• Component: Layout and Rendering
• Status: RESOLVED FIXED

[Matthew Delaney]

2d.path.arcTo.ensuresubpath.2

[Andreas Kling]

2d.path.arcTo.ensuresubpath.1 is also skipped on Mac.

[Andreas Kling]

Created attachment 61872 [details]
Proposed patch

[Andreas Kling]

Created attachment 61873 [details]
Proposed patch v2

Missed the fact that Path::isEmpty() and !Path::hasCurrentPoint() mean different things.

[Sam Weinig]

(In reply to comment #3)
> Created an attachment (id=61873) [details]
> Proposed patch v2
> 
> Missed the fact that Path::isEmpty() and !Path::hasCurrentPoint() mean different things.

In the CG code path at least, Path::isEmpty() == !Path::hasCurrentPoint() I think.  On what platforms do these differ and how?

[Andreas Kling]

From Path.h:

    // Gets the current point of the current path, which is conceptually the final point reached by the path so far.
    // Note the Path can be empty (isEmpty() == true) and still have a current point.

Most Path implementations have Path::isEmpty() == !Path::hasCurrentPoint(), only Skia and OpenVG differ.

PathSkia.cpp:
=============

bool Path::isEmpty() const
{
    return m_path->isEmpty();
}

bool Path::hasCurrentPoint() const
{
    return m_path->getPoints(NULL, 0) != 0;
}

PathOpenVG.cpp:
===============

bool Path::isEmpty() const
{
    m_path->makeCompatibleContextCurrent();
    return !vgGetParameteri(m_path->vgPath(), VG_PATH_NUM_SEGMENTS);
}

bool Path::hasCurrentPoint() const
{
    m_path->makeCompatibleContextCurrent();
    return vgGetParameteri(m_path->vgPath(), VG_PATH_NUM_SEGMENTS) > 0;
}

[WebKit Commit Bot]

Comment on attachment 61873 [details]
Proposed patch v2

Clearing flags on attachment: 61873

Committed r63607: <http://trac.webkit.org/changeset/63607>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.