I don't think anyone is planning to review the SH4-specific changes in
detail,
as AFAIK nobody on the team knows the SH4 architecture well enough to
spot any
mistakes.
The sounds logical.
The team has discussed this topic, and before we land SH4 support, it
would be
nice to understand the motivation behind the port in order for us to be
able
to
decide to what extent we should support it.
We ported v8 on sh4 because that's something that some of our clients want.
This
is mainly for a faster WebKit or just to run node.js.
I don't think the first support level is the right one for sh4 as I will
stay as
maintainer of the port. So "basic support" look appropriate.
[...]. For the latter to work, the platform port must be in a shape where
it fits into our infrastructure and workflow: we must at least be able to
compile it and run the test suite, which means it must have a simulator
and
the
test expectations must be up to date so that the test suite passes. We're
not
willing to introduce dependencies on third-party tools for this, e.g.
QEmu.
I guess that providing a tar file with everythind that you need in order to
build and test v8 for sh4 is not enought to change this?
Once you have all of the code in place (simulator, anything missing to
make
the
majority of the tests work, expectation file for the others), we will
review
the
style and general structure of the SH4 specific code. This may take a few
weeks.
To be clear, once the patch is approved, we will land it into the V8
tree, but
it will continue to be your responsibility to make sure it keeps working
and
stays up to date, and we reserve the right to remove the port at a future
date
if you are no longer able to maintain it.
Sound ok too.
I will come back to you when I have the simulator for sh4.
Thanks for the review.
--
Rémi Duraffort
https://codereview.chromium.org/11275184/
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
