On Tue, Jun 16, 2015 at 10:03 AM, Michael Hablich habl...@chromium.org
wrote:
Hi have you tried building it with the latest stable release (4.3)? Our
Getting
Started https://developers.google.com/v8/get_started is only compatible
with the latest stable release.
I discussed this with Jochen
On Tue, Jun 16, 2015 at 1:14 PM, treck piotre...@gmail.com wrote:
Ok, Now I've used code from
https://github.com/v8/v8-git-mirror/blob/master/samples/hello-world.cc
and it worked.
See https://groups.google.com/forum/#!msg/v8-users/V1YSPacBZWc/HXWuTZeuDzgJ
Regarding the warnings: v8 itself
On Fri, Jun 12, 2015 at 5:47 AM, solar solarcosm...@gmail.com wrote:
[...]
char* retval = (char*)ToCString(String::Utf8Value(result));
printf(retval = %s, %p \n, retval, retval);
That's a classical use-after-free bug: String::Utf8Value(result) results in
a temporary object which is dead
On Wed, Jun 3, 2015 at 11:44 AM, Deepak Subramanian subudee...@gmail.com
wrote:
1. I meant now that it a boxed value, I have no reason to add more
statements in the GC, right ?. [Atleast from my understanding of the GC]
Yes, the GC should need no changes if the new field is boxed.
2. Cool.
On Wed, Jun 3, 2015 at 10:57 AM, Deepak Subramanian subudee...@gmail.com
wrote:
1. [...] Now that it a boxed value, I have no reason to print it, right ?.
I don't understand that question, could you elaborate?
2. ACCESSORS_SPECIAL is currently the same as ACCESSORS with some print
On Thu, May 28, 2015 at 3:52 PM, Deepak Subramanian subudee...@gmail.com
wrote:
[...]
ACCESSORS_SPECIAL(JSValue ...) //instead of accessors
What is ACCESSORS_SPECIAL?
//Changed the builtins-x64.cc to correct the STATIC_ASSERT(JSValue::kSize
== 4* kPointerSize) to
On Mon, May 25, 2015 at 5:08 PM, Deepak Subramanian subudee...@gmail.com
wrote:
[...] But I am not able to find the implementation of either value() or
set_value(). Am I missing something trivial here ?
The magic is in src/objects-inl.h, of course again behind a macro: :-}
On Tue, May 12, 2015 at 6:11 PM, Jason Josephy jjose...@gmail.com wrote:
I just built using Visual Studio and I am confused by the generated libs.
I read there should be a v8_lib.lib that I can link to but I see that there
are 4 of these in the output they look like this
v8_lib_0.lib
On Tue, Mar 10, 2015 at 7:39 PM, David Xanatos xanatosda...@gmail.com
wrote:
I'm porting a large project right now and ofter I head code the returned
v8::Undefined(),
now we don't have a return like this but a args.GetReturnValue()... whats
the default here, I guess I don't need to explicitly
On Tue, Mar 10, 2015 at 8:32 AM, David Xanatos xanatosda...@gmail.com
wrote:
whats the difference between v8::Handle and v8::Local ?
Handle is a historical artifact and will go away completely soon: There
will be only v8::Local (normal implicitly handled references to JavaScript
object) and
On Tue, Mar 10, 2015 at 9:01 AM, David Xanatos xanatosda...@gmail.com
wrote:
PS: do i still need HandleScope if working with v8::Local ?
Yes, nothing changed regarding this.
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
---
You received this
On Tue, Mar 10, 2015 at 8:58 AM, David Xanatos xanatosda...@gmail.com
wrote:
[...] and what about v8::Persistent is that repalced by v8::Global?
If you're working with a bleeding edge v8, you should use v8::Global
already.
--
--
v8-users mailing list
v8-users@googlegroups.com
On Tue, Mar 10, 2015 at 9:34 AM, David Xanatos xanatosda...@gmail.com
wrote:
Is v8-git-mirror-4.3.26.zip bleeding edge or is that a stabile build?
It's an almost bleeding edge vesion, see
https://github.com/v8/v8-git-mirror/branches: It's currently 30 commits
behind.
--
--
v8-users mailing
On Thu, Nov 20, 2014 at 12:48 AM, Robert Jenks rje...@universalbits.com
wrote:
[...]class Jav8Context {
public:
Jav8Context() {
// Create a new Isolate and make it the current one.
isolate_ = v8::Isolate::New();
v8::Isolate::Scope
On Tue, Oct 14, 2014 at 2:53 AM, Mark Tarrabain markt1...@gmail.com wrote:
[...] It appears to be impossible for different threads to use different
isolates at the same time,
That't not correct: 10 different threads could happily each use a different
Isolate at the same time, that's how web
Following Chrome's lead, v8 will require a C++11-capable compiler now. Note
that we don't require a *full* C++11 implementation, basically we rely on
language features, not library/header ones. For details see
http://chromium-cpp.appspot.com/.
Some features already sneaked in earlier and there
On Mon, Sep 22, 2014 at 8:16 AM, ksakam...@chromium.org wrote:
[...] V8 has --use_strict command line flag which enables strict mode
globally, but we want to enforce strict mode only on worker scripts. Is
there any way to achieve this?
Currently, no: All our flags are global variables, and
On Fri, Sep 12, 2014 at 2:27 AM, Mark Tarrabain markt1...@gmail.com wrote:
[...] I intend to wrap each thread's use of the isolate in a Locker
anyways, but will using a v8::Locker in each thread that needs to use the
isolate pointer be sufficient to block simultaneous access, or do I
You need to initialize the platform, see e.g.
https://code.google.com/p/v8/source/browse/branches/bleeding_edge/samples/shell.cc#82.
No fancy locking etc. involved here, the platform is internally just a
global variable. Not nice, but OK for now...
--
--
v8-users mailing list
On Thu, Sep 4, 2014 at 6:01 AM, Srinivas sreeni.ed...@gmail.com wrote:
[...] Can Some one help me to understand whether it is possible to Unlock
while in the middle of some JS function execution and entered by another
thread.?
This should work, e.g. Chrome's proxy resolver is working like
On Fri, Aug 29, 2014 at 1:14 PM, Ben Noordhuis i...@bnoordhuis.nl wrote:
Try `make builddeps` first. That should probably be in a README or
something.
As surprising as it is for the v8 project, ;-) it *is* actually documented
on https://code.google.com/p/v8/wiki/UsingGit (linked from
On Thu, Aug 28, 2014 at 4:02 AM, Liu Xiao liuxiao186...@gmail.com wrote:
We run it on basically every commit in our waterfall: See the V8 Linux
- gcmole column on http://build.chromium.org/p/client.v8/console.
Sorry I can't find anything about GCMole.
Next to last column in the Linux
On Thu, Aug 28, 2014 at 10:17 AM, Liu Xiao liuxiao186...@gmail.com wrote:
I have ran bootstrap.sh, but error occurred when it compiled the
libgcmole.so [...]
Perhaps the llvm/clang API changed once again, that happens quite often,
and that's the reason why we update our gcmole installation
On Thu, Aug 21, 2014 at 5:52 AM, Louis Santillan lpsan...@gmail.com wrote:
Thanks Jakob. That answers my question on how to link it. The
following line works the way I intend it to.
g++ -Iinclude myshell.cpp -o myshell -Wl,--start-group
On Thu, Aug 7, 2014 at 11:50 PM, Prince Wall realprincew...@gmail.com
wrote:
[...] However, from [v8]\src\arm64\constants-arm64.h, it says Assert that
this is an LP64 system and uses type long a lot, assuming it is 64-bit.
However, Windows x64 is LLP64 system and has long as 32-bit. Does this
Just a quick heads up: Following Chrome and making our life as developers
easier, we intend to switch to C++11 in v8 soon. For now we will do this
only partially: Usage of a compiler-only features will be allowed, but not
something based on headers/libraries. This means in detail:
* Usage of
On Mon, Jul 28, 2014 at 6:19 PM, fabio...@gmail.com wrote:
How can I change (and dispose) an object allocate with External::New ?
[...]
The External objects are immutable wrappers around void*, so you can't
change these. Using SetInternalField to overwrite the previous field with a
new
On Mon, Jul 28, 2014 at 10:11 PM, Jacob G kurtext...@gmail.com wrote:
https://blog.mozilla.org/javascript/2014/07/21/slimmer-and-faster-javascript-strings-in-firefox/
Honestly, I'm not all into this and I'm therefore not sure if this is a
thing in v8 already, or could be a thing. Still, I
On Wed, Jul 16, 2014 at 12:29 PM, Oliver Bock oli...@g7.org wrote:
Hi Jakob,
This is a good idea, but unfortunately my existing code relies on the
eval-like nature of fragments, in that they return whatever the last line
in the script evaluates to. [...]
Easily fixable via the highly
On Sun, Jun 29, 2014 at 12:07 PM, Peter dietersk...@gmail.com wrote:
[...] Error while loading operator: ../build/libNodeJSOperatorWrapper.so:
undefined symbol:
_ZN2v811HandleScope12CreateHandleEPNS_8internal7IsolateEPNS1_6ObjectE -
TxToken: 5
I don't know in detail what you're doing, but
On Sat, Jun 14, 2014 at 11:45 AM, Isiah Meadows impinb...@gmail.com wrote:
Actually, a fairly similar hack appears to already be in place, and likely
more efficient as well.
Hmmm, unless I misunderstand things, we are less efficient in
StringStartsWith: Using '%StringIndexOf(s, ss, start) ===
On Fri, Jun 13, 2014 at 2:15 PM, juu julien.vouilla...@gmail.com wrote:
But if I use multiples isolates, I won't be able to share objects across
Isolates' contexts ? No ?
Nope, hence the term Isolate... ;-)
In my case, (AMD Require implementation) the result of my second thread
execution
On Thu, Jun 12, 2014 at 1:51 PM, Jakob Kummerow jkumme...@chromium.org
wrote:
Right. Closure is a bit of an overloaded term. I meant function objects
(or instances if you prefer). Even if they don't access variables from
their outer scope, by executing foo = function() {...} every time the
On Fri, May 16, 2014 at 3:26 AM, Yong Wang wangyong...@gmail.com wrote:
try bash tools/plot-timer-events benchmarks/v8.log
Hmmm, that's bad: We start with '#!/bin/sh' and actually depend on a few
bash-isms later. One easy fix would be using '#!/bin/bash' instead, but an
IMHO nicer one would
On Tue, May 13, 2014 at 11:15 PM, Ben Noordhuis i...@bnoordhuis.nl wrote:
Yang, isn't that an implementation detail leaking into the API? [...]
Yup, it is... :-} Fix reverting the external API change under way:
https://codereview.chromium.org/285643008/
--
--
v8-users mailing list
On Fri, May 9, 2014 at 2:50 PM, Ben Noordhuis i...@bnoordhuis.nl wrote:
[...] Jokes aside, a while ago I did some non-scientific benchmarks on cmov
vs. cmp + a conditional jump (trying to speed up polymorphic functions
in OCaml) and cmov was always slower on an i7. Maybe it performs
better
On Fri, May 9, 2014 at 5:11 PM, Louis Santillan lpsan...@gmail.com wrote:
Intel [1] and DMP [2] are still producing non-SSE (let alone SSE2) x86-32
chips. You may wish to reconsider.
[1]
http://linuxgizmos.com/intel-launches-arduino-compatible-galileo-board/
[2]
On Fri, Mar 28, 2014 at 7:20 PM, Jane Chen jxche...@gmail.com wrote:
[...] I got it just after I wrote my own book-keeping for them. :-) Will
check it out anyway.
Just as a reminder (this comes up roughly once a month at least ;-) : Don't
make any assumptions when or even *if* the callback
On Fri, Mar 28, 2014 at 2:46 AM, Jane Chen jxche...@gmail.com wrote:
With a model where I always have one isolate per thread, and a thread may
serve multiple requests and live around for some time, is there any reason
why I want to Enter the isolate every time I make v8 calls and Exit
On Fri, Mar 28, 2014 at 2:50 AM, Jane Chen jxche...@gmail.com wrote:
May I know what would be the svn repository branch version for this?
http://v8.googlecode.com/svn/branches/3.24 is the SVN branch for the v8
version in the upcoming Chrome 34,
http://v8.googlecode.com/svn/branches/3.25 will
With the Chrome 34 stable release not too far in the future and the branch
for 35 already looming, it is a good time to give a small overview of the
recent and upcoming v8 API changes:
* In the 34 branch, we removed basically all previously deprecated
things to wipe the slate clean.
* The
Your problem is that there is no current Isolate in the newly created
thread, as you can see from the isolate=0x lines in the
backtrace. The current Isolate is thread-local state, and a big PITA,
especially when combined with the default Isolate, as in your example. I
don't know
On Thu, Jan 23, 2014 at 11:55 PM, Niklas Voss niklas.v...@gmail.com wrote:
[...] Are persistent values bound to isotopes? And if, how can I pass an
arbitrary value handle to another isotope/thread?
Everything created in one Isolate is tied to that Isolate and using it
somehow in another
On Fri, Jan 17, 2014 at 6:56 PM, Niklas Voss niklas.v...@gmail.com wrote:
[...]
* How can I successfully register an extension?
Your code fragment looks OK *if* the strings you pass to the Extension
constructor are still alive at the point where the Context is created,
there is even a warning
One thing I forgot in my previous email: We want to remove the whole
extension mechanism relatively soon, anyway. It is quite complicated
internally, has an ugly API for various reasons and is basically useless:
* To install native functions, we have FunctionTemplate.
* To share code
On Mon, Jan 20, 2014 at 2:55 PM, Niklas Voss niklas.v...@gmail.com wrote:
[...] If the program is executed I get the following error:
#
# Fatal error in CALL_AND_RETRY
# Allocation failed - process out of memory
#
The problem is most likely related to me using Script::New wrong or
On Fri, Jan 17, 2014 at 4:20 PM, Jakob Stoeck stoeckja...@googlemail.comwrote:
Hi, when running the example from the testing wiki
pagehttps://code.google.com/p/v8/wiki/Testing
tools/run-tests.py --arch=ia32 cctest/test-heap/SymbolTable
mjsunit/delete-in-eval
I get the following output:
On Thu, Jan 16, 2014 at 11:16 AM, Александр Квашнин
alexander.kvash...@gmail.com wrote:
[...]t
File build/gyp/gyp_main.py, line 12
except ImportError, e:
^
SyntaxError: invalid syntax
make: *** [out/Makefile.native] Error 1
What's wrong?
That's a Python 2
On Tue, Jan 7, 2014 at 9:11 PM, Alejandro F. Reimondo
alereimo...@smalltalking.net wrote:
[...] If it is required to put gyp under SVN, please,
can you give me the SVN server URL (or link
to docummentation where I need to read to
upgrade my tools)
See the first section of
On Wed, Dec 25, 2013 at 7:16 AM, Richard S blakpi...@gmail.com wrote:
So, for Spencer, myself, and anyone else using VS2012, would the solution
be to link against all of the lib files produced after building? (To be
specific, v8_{base.ia32, nosnapshot.ia32, snapshot}.lib and icu{i18n,
And another (performance) hint: If you can ensure that the native pointer
is always 2-byte-aligned, you can use
void* v8::Object::GetAlignedPointerFromInternalField(int index);
void v8::Object::SetAlignedPointerInInternalField(int index, void* value);
to avoid the wrapping of the pointer via
On Tue, Dec 17, 2013 at 8:54 AM, Venkata Savilla svn...@gmail.com wrote:
Hi Sven Panne ,
V8 snapshot stores the context of compiled javascript code .Our
requirement is to store the runtime javascript state and then resore the
state dynamically.
What exactly should runtime javascript
On Tue, Dec 17, 2013 at 10:10 AM, Venkata Savilla svn...@gmail.com wrote:
The use case is follows.
when the users browses the website , some times browser crashes .
I want to restore the application to the previous state by saving
javascrpt context and DOM node data.
i want to launch the
This has already been discussed in
https://code.google.com/p/v8/issues/detail?id=3057, but our external web
pages have not been updated. Stefano?
2 remarks:
* The ICU libraries have circular dependencies, that's why you should
use --start-group/--end-group.
* Using werror=no should not be
On Fri, Dec 13, 2013 at 6:41 PM, shyam kumar shyamu...@yahoo.co.in wrote:
thanks for the reply.I am new to v8 .Can you explain more in detail about
per -Isolate and per-Context.
https://developers.google.com/v8/embed
--
--
v8-users mailing list
v8-users@googlegroups.com
On Wed, Dec 4, 2013 at 11:11 AM, dan wu wudan@gmail.com wrote:
I have called V8:Dispose() at last,but still has memory leaks !
Does it still leak with a v8 from bleeding_edge? The last 3.20 version is
more than 1500 commits behind, there might very well be some fixes since
then...
--
--
On Wed, Dec 4, 2013 at 10:33 AM, Steve Jonghee Yun yun.jong...@gmail.comwrote:
From the example code, I think that all of tasks are done.
In general you don't care about leaks when the process is about to be
terminated, anyway, that's probably why lots of example code doesn't call
On Wed, Dec 4, 2013 at 4:01 PM, dan wu wudan@gmail.com wrote:
I upgrade V8 version to 3.21,but has leaks also!
That's still more than 1000 commits behind bleeding_edge, which is at
3.23.19...
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
On Thu, Nov 28, 2013 at 3:51 AM, Hoping White baihaop...@gmail.com wrote:
[...] when calling function2, there is memory leak or not?
No, there is no leak, the new Handle in function1 lives in the
HandleScope of function2.
--
--
v8-users mailing list
v8-users@googlegroups.com
On Thu, Nov 28, 2013 at 4:15 AM, Hoping White baihaop...@gmail.com wrote:
Thanks for the reply. Then any object which has
AlignedPointerInInternalField must be hold by myself and release by myself.
Is there any interceptor to do this?
So this can be easier with v8 gc.
You can use
On Fri, Nov 15, 2013 at 6:58 AM, Steve Jonghee Yun yun.jong...@gmail.comwrote:
Q: Are there any differences between (**reinterpret_cast*Object(F
IELD_ADDR(p, offset))) and (*reinterpret_cast*Object***(FIELD_ADDR(p,
offset))) ?
Let's rename Object* to Foo and FIELD_ADDR(p, offset) to bar,
https://code.google.com/p/v8/source/browse/branches/bleeding_edge/src/objects-inl.h#5099
--
--
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
On Wed, Nov 13, 2013 at 6:17 AM, Steve Jonghee Yun yun.jong...@gmail.comwrote:
[...] 1. Why does subtract kHeapObjectTag in FIELD_ADDR() ?
We use a tagged representation to distinguish pointers and integral values,
so all our pointers are off by 1 and the integers are scaled by a factor of
2
On Mon, Nov 11, 2013 at 12:51 PM, Dmitri S dsavi...@gmail.com wrote:
On Monday, November 4, 2013 8:26:59 AM UTC+1, Sven Panne wrote:
Unless there is a bug, they will always be the same. But note that
Isolate::GetCurrent() is on death row and will be removed as soon as we can.
Who could
Unless there is a bug, they will always be the same. But note that
Isolate::GetCurrent() is on death row and will be removed as soon as we can.
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
---
You received this message because you are subscribed
There are no hard and fast rules in our test suite, it is basically a
slightly confusing mixture of several naming schemes. We have, as you
already mentioned, regression tests with v8 issue numbers, we some with
Chrome issue numbers (of course with slightly varying numbering schemes),
we have
On Thu, Oct 24, 2013 at 3:32 AM, Steve Jonghee Yun yun.jong...@gmail.comwrote:
[...] I have spent 3 weeks but I couldn't find that how 8v call DOM's
function.
All the DOM stuff is not part of v8 itself, but part of Chrome (Blink, to
be more exact). Blink registers all DOM attributes and
On Thu, Oct 10, 2013 at 1:24 PM, Daniel Clifford da...@chromium.org wrote:
In order to build the shared library configuration on Mac (actually, any
platform expect linux with make), you need to manually set up the path to
find intermediate shared library results (thanks for Jochen Eisinger for
On Thu, Sep 19, 2013 at 6:07 PM, Luke l...@webconnex.com wrote:
On Wednesday, September 18, 2013 11:16:18 PM UTC-7, Sven Panne wrote:
Just a quick drive-by warning: Try to avoid Isolate::GetCurrent(), we
work hard on removing the whole notion of current Isolate. There is no
exact time frame
Just a quick drive-by warning: Try to avoid Isolate::GetCurrent(), we work
hard on removing the whole notion of current Isolate. There is no exact
time frame for this, but we will do it as soon as we can, so don't say you
haven't been warned... ;-)
--
--
v8-users mailing list
Hmmm, this doesn't look right. There seems to be some API nonsense/bug,
investigating...
--
--
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
On Mon, Jul 29, 2013 at 9:41 AM, Sven Panne svenpa...@chromium.org wrote:
Hmmm, this doesn't look right. There seems to be some API nonsense/bug,
investigating...
Actually things are correct, but very perplexing. Consider the following
session in d8
On Mon, Jul 29, 2013 at 2:13 PM, ioannis ioanni...@gmail.com wrote:
So if i understand correctly, in order to obtain the underlying value of a
BooleanObject, i need to first cast the Value to BooleanObject and then call
BooleanValue()
Yes, and this is caused by the fact that polymorphism
Before we continue this useless discussion any further: Has anybody
actually *read* Boehm's paper/slides?
--
--
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.
On Sat, May 25, 2013 at 4:36 AM, Stephan Beal sgb...@googlemail.com wrote:
On Fri, May 24, 2013 at 9:58 PM, Jim Acquavella loudersp...@gmail.comwrote:
Thanks Stephan! I'm surprised and disappointed I can't guarentee my C++
objects will be destroyed. Can anyone from Google comment on this?
Further note: When your test on IE10 shows that it is almost twice as fast
as the plain old for loop, check if your test contains dead code. IE10 is
very good at removing this, it easily gives you good benchmark scores
without actually doing anything fantastic, because a lot of benchmarks
contain
If you use v8::Locker anywhere, you have to use it everywhere. If you
that's not practicable for you, you'll have to ensure that at most one
thread is executing within each given Isolate by other means, i.e. locking
on your application side.
--
--
v8-users mailing list
v8-users@googlegroups.com
Just a quick drive-by comment: You *can* use an Isolate from various
threads, but you'll have to use a v8::Locker in all threads then. This way,
you inform v8 about your thread switches and make sure that at any given
point in time at most one thread is actively using an Isolate.
--
--
v8-users
I would propose that you build a debug version of d8 (make x64.debug, in
your case), add -g to your compile command above, and start the resulting
program in gdb (gdb --eval-command=run --args programm). Perhaps you
trigger an assertion in v8 and/or the backtrace in gdb will give you a
hint. In
v8::Locker::IsActive() is just telling you that a v8::Locker has been used
before *somewhere* in your program, not necessarily in your thread or for
your current Isolate. This means: IsActive is basically useless, and I
consider it to be on death row for deprecation/removal. Apart from that API
Re-reading your original mail, I am not 100% sure what you are trying to
achieve by using separate Isolates and threads: Sharing *anything*
JavaScript-ish between different Isolates is by definition impossible
(hence the name ;-). An Isolate is a complete separate instance of the VM,
and of course
What you are describing is again not a statement of the problem in itself,
but an attempt to solve something. What in detail that is, is still unclear
to me. I think you are mixing up the concepts of Context and Isolate: An
Isolate a concept for strictly, well, isolating several V8 instances
The description in the Wiki is not really accurate: We copy bleeding_edge
to trunk on a per-need basis, and things have been mildly tested by our own
v8 build bots. Depending on what you need, the branch for the next upcoming
Chrome version (currently http://v8.googlecode.com/svn/branches/3.16/)
On Fri, Feb 22, 2013 at 11:26 PM, Flying Jester foolkingcr...@gmail.comwrote:
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
v8's external API is currently
What are the detailed steps to reproduce this?
--
--
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
OK, my apps are d8 and Chrome, and both work. :-) Seriously, we need more
detailed steps and not just some fuzzy prose. It could very well be the
case that some applications embedding v8 worked only by accident, so a
stand-alone reproduction is really needed, e.g. a single small C/C++ file
I don't even know what to compile from
https://code.google.com/p/v8-juice/source/browse/, the trunk seems to be
out of date, it uses e.g. GetPointerFromInternalField and friends which are
not part of v8's external API anymore. So until I see some stand-alone
repro, I'll ignore this issue from now
On Thu, Feb 7, 2013 at 8:22 PM, Stuart Allen stu...@lttlrck.com wrote:
This error occurs if the pointer is 0. That's an even number last time I
looked.
I have been using this facility to storing flags to avoid Object overhead,
now all my flags have to be 0 and even :)
I don't think so, we
On Mon, Feb 4, 2013 at 3:58 PM, Andrii Melnykov andy.melni...@gmail.comwrote:
http://hpaste.org/81784 contains a benchmark - slow_test() is twice as
slow as fast_test() [...]
The way Array.prototype.map is specified (see section 15.4.4.19 in the ECMA
spec) makes it very hard to implement
On Mon, Jan 21, 2013 at 9:20 AM, Flying Jester foolkingcr...@gmail.comwrote:
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
On Thu, Jan 10, 2013 at 7:58 AM, Joe Millenbach jmillenb...@gmail.comwrote:
If I make a patch that added the option, could it be accepted? Or is
this functionality that is not desired?
It depends on how intrusive the patch is: If it is a relatively small one
which is not interfering with our
From what I understand about the example above, it looks like an API abuse:
The GC decides that a given object is dead and gives its associated
callback a chance to revive it (via MakeWeak) or say explicitly that it's
OK to really kill it (via Dispose). In the latter case one should not call
*any*
You return a Handle from a dead HandleScope from Shell::Load. Using a debug
build + gdb can figure this out easily, because we explicitly zap several
data structures with special values to catch such things, in this case it's
kHandleZapValue (0xbaddeaf on ia32).
On Mon, Jan 7, 2013 at 2:11 AM,
Could you give detailed steps to reproduce your problem? Have you checked
that the pointer in question at the point of the assertion is *really*
aligned (i.e. has 0 as its last bit)?
If you can't guarantee 2-byte alignment, use External to (un-)wrap your
pointers and get/set them via the
On Wed, Nov 7, 2012 at 6:29 PM, Vyacheslav Egorov vego...@chromium.orgwrote:
[...] Sven recently was changing things in that neighborhood. I am CCing
him. I hope he will be able to help.
I am not sure which neighborhood we are talking about, but I'll take a
look... ;-)
--
v8-users mailing
A Foreign object is *not* a JavaScript object it is intended to be used
only internally. Furthermore, v8's GC needs to distinguish the objects
under its control from the external ones, so we can't simply put any void*
into an internal field. Currently this part of the API is in a poor and
slightly
Yep, we diagnosed the same yesterday. What happens is this: When trying to
construct an External value, we check if the given void* fits in a Smi. If
it does, we return a Smi, otherwise we create a Foreign object which wraps
the void*. The latter case is problematic, and this is exactly what is
After several discussions, it is not so clear anymore what to do. First of
all, SilkJS does not follow https://developers.google.com/v8/embed#dynamic on
how to handle foreign (i.e. C/C++) pointers when embedding v8. The return
value of External::New is supposed to live in an internal field, but it
Shouldn't use use GetPointerFromInternalField and SetPointerInInternalField
for this? For SetInternalField one could change the signature to use Data
or provide an overloaded version. Anyway, saying External is-a Value in its
current state is plainly wrong... :-)
On Fri, Sep 7, 2012 at 3:02 PM,
The beating continues... ;-)
Regarding Boehm's finalization is rarely needed: I think statement
generally holds, *except* for the case we are discussing, namely writing a
binding for a library with explicit allocation/deallocation where the
deallocation part is not exposed in the binding. While
1 - 100 of 110 matches
Mail list logo
