This CL teaches the Sky compositor to record an SkPicture instead of just
drawing in immediate mode. Using an SkPicture will let us experiment with
various optimizations in Skia that preprocess the SkPicture before translating
the Skia commands to GL commands.
This CL also refactors the "display delegate" system now that we can hand off
an SkPicture to a rasterization backend. The new design requires fewer
backpointers and removes the static display delegate factory.
R=eseidel@google.com
Review URL: https://codereview.chromium.org/854303002
Previously, we were creating a new texture for every frame. Using tracing on a
Nexus 5, that appears to cost about 6ms per frame on the GPU thread. After this
CL, we keep a cache of recently used textures and only allocate a new one when
either (1) we don't have one free or (2) our size requirements have changed.
If our size requirements change, we dump the whole texture cache. In the
future, we'll probably need something fancier if we ever need more than one
size texture per frame.
R=jamesr@chromium.org
Review URL: https://codereview.chromium.org/845963006
In preperation for removing the Client annotation from ServiceProvider
this passes a second parameter of type ServiceProvider in the shell and
application Connect calls. In this patch the type signatures are updated
but the second parameter is basically unused. The intention is that the
first parameter |services| will be used for the connecting application to
request services from the connected application (as it does currently)
and the second parameter |exported_services| be used for the connecting
application to provide services to the connected application. We have
very few services exported in the second direction today - I'll update
them to use the second parameter in a follow-up patch.
R=darin@chromium.org
Review URL: https://codereview.chromium.org/845593003
Surfaces identifiers have a local and a namespace component. The
namespace is particular to a connection. The local component is up to
the endpoint of the connection to manage. In contexts where a surface
that may be from anywhere is referenced, a fully qualified ID with both
the local and namespace component is needed in order to be unambiguous.
In contexts that can only apply to local surfaces only the local id is
needed.
This updates the mojo.Surface APIs that can only refer to local IDs to
only take the local component. In particular creating, destroying, or
submitting a frame can only refer to surfaces created on that connection.
References to surfaces within a frame may refer to local or foreign
surfaces so they use fully qualified IDs.
This also skips the SurfacesService indirection since many applications
can perform useful operations on a mojo.Surface interface such as create
surfaces and submit frames without knowing their ID namespace. The
namespace component is needed only to pass fully qualified IDs to other
actors that may wish to reference the produced frame.
R=sky@chromium.org
Review URL: https://codereview.chromium.org/826423008
This way we get some broad pixel test coverage.
Do this by dumping the png from the flights app,
uploading it to google storage and then having
the reference file load the uploaded file.
This isn't a good long-term solution. Long-term
we should dump paint commands (or some other
textual representation) and/or make more targetted
reftests that don't need pngs. As such, the process
for doing this is painful and manual. If we decide
we want to have pixel tests properly, we'll want
to do something more automated obviously.
Also, turn on the discard_transparency bool when
encoding the pngs. Otherwise, the png is lossy
and none of this works.
R=esprehn@chromium.org
Review URL: https://codereview.chromium.org/836363003
Surface IDs are composed of a namespace component and a local
identifier component. This splits up the two pieces in the mojom to make
it clearer.
This also simplifies the path for connecting to the surfaces service. In
order to perform any operations with surfaces, each client must know
what its id namespace value is. This was done by first connecting to a
SurfacesService interface and then calling CreateSurfaceConnection which
returned a Surface pointer and the namespace. Having to connect to this
extra interface and receive the Surface impl asynchronously complicated
startup code for applications by adding extra states in startup.
Instead of that, this allows connecting to the Surface interface
directly and promises that the ID namespace will be provided as the
first message in the pipe directed at the SurfaceClient. Callers can
then either block on startup or handle setting the ID namespace
asynchronously depending on what other things that thread could be doing
at the time.
In a follow-up, I plan to define the id namespace value 0 as meaning "the
namespace of this connection" which will allow creating surfaces and
submitting frames before learning the connection's namespace. The only
thing the namespace will be passing surface IDs to other clients for them
to embed.
This also removes the Size parameter from CreateSurface, which wasn't
used by anything. The size of submitted frames is part of the embedder /
embedee contract.
R=esprehn@chromium.org, sky@chromium.org
Review URL: https://codereview.chromium.org/807733002
-Add a --testing flag to sky_viewer and cause it to paint into an
SkBitmap instead of a ganesh surface so we can get the pixels out.
-Add GetPixelsForTesting to layer.cc to actually grab out the pixels.
-Add a reftest and a mismatch reftest. They need a setTimeout after
the load event. Unclear why or what the right fix is. Maybe we should
give internals some way to force the paint? If we don't have the
setTimeout, we paint a white page (so we do a paint, but with no
content).
-Add a DisplayDelegate to Layer so that Viewer can decide whether
to use the real ganesh backend or the SkBitmap one without littering
the whole code-base with is_testing bools and logic.
R=esprehn@chromium.org
Review URL: https://codereview.chromium.org/797063002
This CL goes from this:
//mojo/services/public/cpp/geometry
//mojo/services/public/interfaces/geometry
to this:
//mojo/services/geometry/public/cpp
//mojo/services/geometry/public/interfaces
This CL also makes the Mojo-side changes required to roll this change into
Chromium.
TBR=beng
Review URL: https://codereview.chromium.org/790213002
This CL goes from this:
//mojo/services/public/cpp/surfaces
//mojo/services/public/interfaces/surfaces
to this:
//mojo/services/surfaces/public/cpp
//mojo/services/surfaces/public/interfaces
This CL also makes the Mojo-side changes required to roll this change into
Chromium.
TBR=beng
Review URL: https://codereview.chromium.org/792813002
Since we don't support using the component build to produce mojo apps,
we can simplify the build targets in a few ways:
*) every mojo_native_application must depend on the c system thunks,
so just make that part of the template instead of requiring the dep
*) there's no such thing as depending on gles2 headers from a component,
so delete the forwarding group.
Most targets that want to use the gles2 headers in a mojo context
want to depend on an implementation through the thunks, so
//mojo/public/c/gles2 does just that. A smaller number of targets (such
as the implementation of the thunks) want to just depend on the headers
but not an impl, so they can depend on //mojo/public/c/gles2:headers.
The //mojo/public/gles target isn't that useful since the only thing we
expose is a set of C entry points.
We can probably also simplify the c system targets, but that's trickier
due to more extensive use from the chromium side.
BUG=438701
R=viettrungluu@chromium.org
Review URL: https://codereview.chromium.org/780733002
Previously, there was no way for clients of the surfaces service to
appropriately rate-limit their frame submissions. We tried using the "return
resources" signal to rate-limit, but that's really a measure of how quickly
you're submitting new resources rather than how quickly the system is putting
up frames.
Currently, only the Sky compositor listens to this signal. Using this signal,
we're able to run sky/examples/spinning-square.sky in sync with the surfaces
service (that is, submitting exactly one frames every 17ms).
R=jamesr@chromium.org
Review URL: https://codereview.chromium.org/756673004
This CL fixes two bugs:
1) We need to enter the Ganesh context in order to destroy the GrContext.
2) We had a race condition whereby we'd try to upload a frame to a surface that
didn't yet exist. This CL fixes that race by adding a state to LayerHost to
wait for the surface to be created before trying to upload to it.
R=esprehn@chromium.org
Review URL: https://codereview.chromium.org/753643002
Turns out we need to reset the context before switching back to drawing with
Ganesh (i.e., call resetContext on the GrContext) to clean out whatever state we've
changed in the GL context. Now the city-list and the flights-app demos work.
R=esprehn@chromium.org
Review URL: https://codereview.chromium.org/752653002
