This does several things:
1. Teaches sky about asynchronous script execution. Previously once all imports
were loaded and the script text was available, we executed a script and assumed
it completed synchronously. We left the parser loop to do so, but that was fine
as the next chunk from the background thread would resume the parser. In this
change scripts now load and execute separately. The "load" step may trigger
further dart import loads which may cause the execution to happen asynchronously
which required teaching both the DartController and the HTMLScriptRunner to
take callbacks to allow HTMLDocumentParser to know to continue parsing after
the Dart script has resolved its imports and executed.
This required re-working some of how the parser executes scripts and I
re-purposed isWaitingForScripts to include "is the parser blocked" where
as before it was limited only to "does the treebuilder have a script", even
though the imports system may have had pending scripts as well.
I made HTMLScriptRunner live only as long as the script it was executing
since it only contained per-script state at this point.
2. Fixed an error reporting bug whereby we would not show errors when "init"
failed to execute, only "main". This required using the dart_mirrors_api.h
which required adding an include path to the core build. :(
3. Made it possible for a single sky file to contain multiple dart <script>
tags. Each <script> is a separate library and executes as
soon as </script> is seen. main or init is called for each. This required
mangling "urls" for these script blocks since Dart unique's libraries by urls.
Before this change it may have been possible to do <import 'foo.sky'> and then
<script>import 'foo.sky'</script> and have it work!?
R=abarth@chromium.org
BUG=
Review URL: https://codereview.chromium.org/938623005
We shouldn't need pendingScripts in ScriptRunner since
we should never be trying to run scripts when we're
not ready to run them.
However this wasn't completely true in the case of imports
there was code to have us break before any start tag after
an <import> was seen, but it was subtly wrong in that it
it would include the start-tag it was trying to break before
in the chunk it sent to the main thread.
This didn't run out to be the problem I was facing, but I fixed
it anyway. The problem which was actually preventing me from
removing pendingScripts was adding a check inside
didRecieveParsedChunk... to check if imports were pending and
add the chunk to the list of pending chunks.
I also renamed m_speculations to m_pendingChunks since these
chunks are never speculative anymore.
We can't test the off-by-one import-breaking code with our
current system, but it would be trivial to test with a
self-closing custom element if/when we ever add custom
elements back to the system.
R=abarth@chromium.org
BUG=
Review URL: https://codereview.chromium.org/934083002
Normal trace events already go directly to base. This CL removes the remaining
trace events that were still indirected through blink::Platform. These were
just the half-finished inspector timeline trace events, which currently aren't
actually hooked up to anything.
This CL also removes the redunant "convertable to trace format" wrappers and
moves their one remaining use over to just using the version in base directly.
R=eseidel@google.com, ojan@chromium.org
Review URL: https://codereview.chromium.org/889823002
This CL makes the parser yield before non-import start tags if there are
outstanding imports. This will let use remove the upgrade path from custom
elements because we can require the registration to happen before the tree
building.
R=eseidel@chromium.org
Review URL: https://codereview.chromium.org/814173005
This caused us to lose our gn check certification. :(
Turns out gn check was just ignoring all the header
paths it didn't understand and so gn check passing
for sky wasn't meaning much. I tried to straighten
out some of the mess in this CL, but its going to take
several more rounds of massaging before gn check
passes again. On the bright side (almost) all of
our headers are absolute now. Turns out my script
(attached to the bug) didn't notice ../ includes
but I'll fix that in the next patch.
R=abarth@chromium.org
BUG=435361
Review URL: https://codereview.chromium.org/746023002
Fix (most) generated includes to have gen/ in their path.
This makes it easier to tell where files exist on disk.
Unfortunately I had to leave the old include path
in engine/BUILD.gn to support all the v8 includes
which were too many to deal with in this patch.
It's a little nasty to have the raw build directory
in our include path, but it produces nicer paths.
R=abarth@chromium.org
This CL adds the beginning of a new module loading system that matches the
conceptual module in the spec. I've writed it up to the import API on
AbstractModule. In a future CL, I'll replace the old module loading system with
this one.
R=eseidel@chromium.org
Review URL: https://codereview.chromium.org/720903002
This CL cleans up the DocumentParser interface to match what's actually needed
by clients. As part of this cleanup, I've removed the asHTMLDocumentParser
dynamic cast and just exposed virtual functions for the state folks were
accessing via the dynamic cast.
R=ojan@chromium.org, eseidel@chromium.org
Review URL: https://codereview.chromium.org/722743002
I used do-webcore-rename from Blink/WebKit
which is very good at doing this kind
of search-replace.
Also removed toRefPrtNativeArray after
conversion since it previously had two
separate flavors. Both versions are no longer
used so I've removed the code until we
need one again.
https://www.irccloud.com/pastebin/5C16p5cE
is the diff I used to do-webcore-rename
TBR=abarth@chromium.org
This CL plumbs the data pipe from the network stack all the way to the parser
thread. We now drain the data pipe on the parser thread, which means network
bytes don't need to transit the main thread to reach the parser.
R=eseidel@chromium.org
Review URL: https://codereview.chromium.org/664573004
