I've been trying to switch from using scons to gyp to build V8, but I still
have yet to build V8 using gyp. Every time I try, MSVC (same in 2008 and
2010) says that 'certain properties of the solution could not be read', or
something a lot like that. V8.base and V8.preparser seem to build with
Yes, I have both Cygwin and Python 2.6. Both are the google supplied
versions. I'm not really sure how I could have made scons work without at
least Python 2.6.
I am also not really sure how having those can change what Visual C++ does.
--
v8-users mailing list
Probably Cygwin, I have it installed to C:\Cygwin, which I know is the
normal place for it on its own, but I don't know if that's right for V8.
I'll go through the building instructions again, being sure Cygwin goes
into the right place.
--
v8-users mailing list
v8-users@googlegroups.com
Hurrah, I have successfully built V8 using Gyp, and successfully built my
program against the generated files!
It was indeed cygwin not being in the third party folder. That shows me for
thinking I know better than the developers.
Thank you.
--
v8-users mailing list
v8-users@googlegroups.com
I've been making a program that uses V8. It used to work fine until I
updated the version of V8 used from 3.15.0 to the last 3.15 revision. Now I
get the following error:
#
# Fatal error in ../src/platform.h, line 604
# CHECK(mutex_ != __null) failed
#
This happens before anything in main(),
I'm using Linux, Fedora 17 x86_64, GCC 4.7.2, V8 3.15.11, x64.debug build.
I tried a bit to make a small program that would cause this same problem,
but I couldn't find a simple way to do it without just using large portions
of the source of the program I'm working on.
The call stack (that's
I recently updated the version of V8 I build my project against, and I was
greeted with a new warning:
warning: ‘void* v8::Object::GetPointerFromInternalField(int)’ is deprecated
I've seen the 'aligned' version of this function,
GetAlignedPointerInInternalField. How would I go about replacing
I routinely use Valgrind to debug and test my own program that embeds V8,
and I have never seen this problem with respect to V8.
Are you sure the problem isn't related to libclnt being compiled to use AVX
or another new instruction set extension that valgrind doesn't fully
support? The report
What switches did you use to compile it? Does it have a snapshot built in?
Is it a shared or static library? Also, why is the download so very large?
I think there may be reasons why it is not especially economical to provide
binary distributions of V8, except with the application that needs
You could try (sort of a hack) blacklisting libcmt.lib in the link - input
-ignore specific default libraries section.
Really, this sounds like a multi-threaded/multi-threaded DLL problem,
though.
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
Not V8, I mean the standard library for your project.
The selection is on the first page of the property menu.
Alternatively, you could try setting the default library ignore settings to
ignore one of those two libraries (libcmt and msvcrt). The problem is that
both are being linked in one way
Linker - input -ignore specific default libraries section.
Or at least they were in MSVC 2010.
On Thursday, October 24, 2013 5:31:53 PM UTC-8, Spencer Killen wrote:
K sweet, where are the library ignore settings?
--
--
v8-users mailing list
v8-users@googlegroups.com
Let's say I have an object on the C++ side that is wrapped for use in JS.
It has several member functions attached to its prototype. Is there a way
to allow this object to be used, for instance, just like an array as well
on the JS side, which will access values specified on the C++ side?
Like
That sounds like the optimizer kicking in.
You can improve startup performance (or so I'm told) by keeping a copy of
its script data (v8::ScriptData), which supposedly holds this information.
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
---
Ah yes, my mistake. I thought the numbers were latencies, not interations.
On Friday, November 15, 2013 1:09:58 AM UTC-9, Jakob Kummerow wrote:
No, a drop in performance after a while does not at all sound like the
optimizer kicking in.
On Thu, Nov 14, 2013 at 7:59 PM, Flying Jester
The first step would be to interface with the service in C++. Then you
would have to expose it to JS.
There is no specific allowance for what you are asking built in to V8, or
any JS. What you want to use is, from what you've said, part of Chrome, and
not part of V8.
--
--
v8-users mailing
Can we see the code?
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
---
You received this message because you are subscribed to the Google Groups
v8-users group.
To unsubscribe from this group and stop receiving emails from it, send an email
to
I'm updating the version of V8 I embed in a program from 3.21 to 3.24, and
I see two new libraries are being built: libcuuc, and libicui18n.
What are these?
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
---
You received this message because you
for sure o.o
but I never saw libcuuc…just try to google it, maybe o.o
Am Di. Feb. 04 2014 02:16:29 schrieb Flying Jester:
I'm updating the version of V8 I embed in a program from 3.21 to 3.24,
and I see two new libraries are being built: libcuuc, and libicui18n.
What
I'm having trouble getting working accessors for JS objects that represent
native data.
You can see how I define the objects here:
https://github.com/FlyingJester/TurboSphere/blob/galileo/plugins/common/plugin.h#L181
How I set accessors here:
On Sunday, April 20, 2014 4:37:28 AM UTC-8, Jos Kuijpers wrote:
Found it: -Dv8_enable_gdbjit=1 on GYP_DEFINES. I missed it in the manual
on google code.
The Wiki page says it only works on Linux btw. is that out-dated?(like
everything else?)
It most definitely is out of date. I am
Basically, at startup I make a function template for each JS class. I make
an instance template, a prototype template, and a constructor from that
template.
I add accessors as follows:
InstanceTemplate-SetAccessor(v8::String::NewFromUtf8(v8::Isolate::
GetCurrent(), name), Getter, Setter);
You could try using a local, newer version of just ar. It is an
exceptionally simple tool, and I believe that the -T option is specifically
for emulating the even simpler, older version of it.
But building V8 using older tools sounds like a recipe for disaster in the
first place. And newer
It looks like you haven't linked V8 in.
Look at your build tools' documentation for how to link with a library/
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
---
You received this message because you are subscribed to the Google Groups
v8-users
Looking at my build (which specifically uses V8 as a shared library and
lacks il8n support):
The old WinSock library no longer needs to be expressly linked to the final
executable. I also haven't got v8_base linking in either, just v8.lib.
It might be useful to just start a new project file,
V8 replaces exactly the component of webkit you are talking about.
Unless you are wondering about a reactionary response. Which I doubt will
happen, the only other serious contender with V8 is SpiderMonkey.
--
--
v8-users mailing list
v8-users@googlegroups.com
to b) let's say I compile my d8 binary, one time for x86. Now I am running
the benchmarks above on 2 different x86 processors (binary is the same!).
Would you expect the same executed/retired instruction count? One reason
why you wouldn't have the same count would be for instance, if
At that age of V8, it's basically a crap shoot. You'd be better off to
rewrite the JS parts altogether.
Off the top of my head:
Callbacks work completely differently.
Non-local handles work completely differently.
Context and isolation and lockers work completely differently.
The exact template
Are there n32 compatible Android devices? Are there any n32 incompatible
Android devices?
V8 is really here to power Chrome, and Chrome on anything except x86 is
really there to power Android.
On Thursday, July 17, 2014 4:02:45 PM UTC-7, Jacob Burkholder wrote:
Hi, thanks for your reply. We
How would get all the accessors of an InstanceTemplate?
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
---
You received this message because you are subscribed to the Google Groups
v8-users group.
To unsubscribe from this group and stop receiving
Is there a way to get the value of a ReturnValue after being set with Set()?
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
---
You received this message because you are subscribed to the Google Groups
v8-users group.
To unsubscribe from this
Since V8 does not know how the memory backing an ArrayBuffer has been
allocated in this case, there is now good way for V8 to free it.
This doesn't make sense.
A part of v8::ArrayBuffer::Allocator is a Free method. When you pass an
allocator to v8::V8::SetArrayBufferAllocator, you are
Will the related repos (gyp, icu, etc.) that are retrieved to build V8 also
be hosted on git at that point?
On Monday, September 15, 2014 11:09:42 PM UTC-8, Michael Achenbach wrote:
Hi!
We're planning to migrate the V8 sources from Svn to Git. The estimated
timeline for this is 1-2 month.
I don't believe you should be able to pass any old pointer into it. I'm
saying it would vastly simplify things to be able to re-internalize a
previously externalized pointer.
On Thursday, September 18, 2014 11:00:10 AM UTC-8, Dmitry Lomov wrote:
On Fri, Sep 12, 2014 at 11:38 PM, Flying
, Sep 18, 2014 at 9:20 PM, Flying Jester foolki...@gmail.com
javascript: wrote:
I don't believe you should be able to pass any old pointer into it. I'm
saying it would vastly simplify things to be able to re-internalize a
previously externalized pointer.
How this restriction solves anything
You should ask on the Chromium (or very maybe the V8-dev) lists.
It will likely involve understanding how to write GYP files. Which is kind
of a mess, especially in Chrome and V8.
On Sunday, September 21, 2014 8:52:58 PM UTC-8, Liu Xiao wrote:
Dear all:
I'm sorry, it isn't a v8 specified
How would I set the contents of an ArrayBuffer from native code?
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
---
You received this message because you are subscribed to the Google Groups
v8-users group.
To unsubscribe from this group and stop
Very simply, V8 is incapable of handling memory efficiently. If you want it
done right, don't even begin to trust that V8 can clean up after itself.
V8 is why people make fun of garbage collection.
On Thursday, October 30, 2014 11:15:24 PM UTC-8, Sebastian Herrlinger wrote:
Did not get to the
No joke it isn't fast...using a brand new MacBook pro, it takes almost an
entire second to copy three megabytes.
But it does work. Thanks!
On Monday, November 10, 2014 2:50:49 PM UTC-9, Ben Noordhuis wrote:
On Tue, Nov 11, 2014 at 12:38 AM, Flying Jester foolki...@gmail.com
javascript
, and use this
trick then. I can still fall back to the slow and terrible method that is
officially supported if that fails.
On Wednesday, November 26, 2014 1:56:48 PM UTC-9, Flying Jester wrote:
I'd like to, but GetIndexedPropertiesExternalArrayData() always returns
NULL.
On Saturday, November
This isn't really the right group for this question.
On Wednesday, November 26, 2014 6:14:16 AM UTC-9, Alex Sakhartchouk wrote:
Hi All,
Hope you can help. Last night, I started seeing this message on my dev
Chromebook pixel, and it's stuck at boot screen. Things were fine yesterday
I've been having some issues with certain native code constructors not
being called.
Usually, this manifests itself when I try to call a member function of the
created object, in which case I get an error about undefined not being a
function.
To demonstrate the root issue, I have tried to
I am not using Node.js
When I say 'wrap the function', I mean that I create a new function
template, and then use Set from the current context's global object with
the intended script-side name of the function, and the function template's
function.
v8::Handlev8::FunctionTemplate Ctortempl =
What I do is to embed the function's name in the exception being thrown
from native. It's not the best, but it gets the job done.
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
---
You received this message because you are subscribed to the
I believe you should be using CXXFLAGS. g++ -std=c++11 is, in fact, NOT a
valid command.
On Tuesday, December 9, 2014 12:41:43 AM UTC-9, kunjie Chen wrote:
Hi guys.
I'm gonna build V8 on centos4.3 and use the version 3.30.10.
I use the follow exports
export CXX=g++ -std=c++11
export
Almost all the crashes I get in garbage collection are either some logic
error in my callback, or because I set up the weak reference callback wrong.
As an example: https://groups.google.com/forum/#!topic/v8-users/luJN258fHMg
Be sure that your callback really works fine. Maybe something is a
I am wondering if they have been taking old svn as bleeding edge and thus
reporting wrong data, or the gap between turbofan and crankshaft is really
that much. Any comments?
That would make sense to me. I have done a lot of work with TurboFan, and I
find it MUCH faster than older V8.
--
--
It might be that the old ISA the rPi uses is just too old to run V8.
Generally, the devs only care about the devices that can also run Android.
On Friday, January 9, 2015 at 5:40:30 AM UTC-9, Milan Pandurov wrote:
I am having a problem with cross compiling google v8 libraries for
raspberry
Unless you need to change the semantics or syntax of the underlying
language, I see absolutely no reason to modify V8. That would be much, much
more work.
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
---
You received this message because you
What exactly are you trying to accomplish with this?
Copying a context would be a huge operation, if it's possible at all (I
don't think it is).
You are almost certainly better off encapsulating the state you want in
JSON or through some API you come up with yourself, and then loading that
in
The `console' object is not a part of JavaScript. It's standard in Web JS,
but no JS engine on its own provides that. If you want some way to log from
JS, you will have to do it yourself.
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
---
You
Perhaps if you iterated the global object? That would give you access to
all objects that exist in the current scope, if you recursively iterated
the objects that the global object has handles to.
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
52 matches
Mail list logo