Re: [v8-dev] Clever number of parallel jobs for Ninja?

On Tue, Apr 30, 2024 at 3:56 AM Paul Harris wrote: > > Hi, > > I'd like to run Ninja build with as many jobs as I have CPUs. > However, some of the build jobs require a lot of RAM (especially for Debug > builds), so I have to reduce the number of jobs artificially, just to keep > the number of

Re: [v8-dev] performance considerations for well-typed Wasm binaries

Does --no-wasm-bounds-checks imply no guard pages? That could explain the difference when the wasm module is small and executes quickly; the overhead of mmap/mprotect/munmap probably starts to dominate. Profiling with perf(1) should be able to confirm that. On Mon, Apr 8, 2024 at 7:07 AM 'Andreas

Re: [v8-dev] are V8 VM concepts the same to Java VM ones?

On Fri, Mar 22, 2024 at 5:27 PM anlex N wrote: > > V8 is built on top of C++, Java is too. > So I wonder this problem. Your question is too open-ended and vague to answer with yes or no. Yes _and_ no would be a correct answer on the other hand. :-) -- -- v8-dev mailing list

Re: [v8-dev] Fatal process out of memory: DateTimePatternGeneratorCache::CreateGenerator

On Thu, Mar 7, 2024 at 12:16 PM Angelo Dell'Aera wrote: > > I collected a gdb backtrace and it shows something which is potentially > interesting > > (gdb) info stack > #0 0x74d72ae6 in v8::base::OS::Abort() () from >

Re: [v8-dev] Fatal process out of memory: DateTimePatternGeneratorCache::CreateGenerator

On Wed, Mar 6, 2024 at 4:22 PM Angelo Dell'Aera wrote: > > > > Did you also update ICU? That error message indicates V8 couldn't > create a icu::DateTimePatternGenerator instance. > > > The Github Actions workflow I use to build STPyV8 automatically fetches the > V8 repository, > perfoms a

Re: [v8-dev] Fatal process out of memory: DateTimePatternGeneratorCache::CreateGenerator

On Mon, Mar 4, 2024 at 2:54 PM Angelo Dell'Aera wrote: > > Hi all, > starting from a few weeks I started observing these crashes while using the > Python > wrapper STPyV8 [1]. I am the lead developer of STPyV8 and no relevant changes > were > introduced recently other than updating the V8

Re: [v8-dev] FixedArray Creation

On Sat, Jan 13, 2024 at 3:38 PM 'Ronald Fenner' via v8-dev wrote: > > Is there a way to create a FixedArray for the embedder standpoint. Looking > through the coode it looks like it's creation is an internal method and can't > create one to use. I was wanting to unit test m y import assertion

Re: [v8-dev] Design Doc: v8::String::WriteUtf8() without flattening

On Wed, Dec 13, 2023 at 5:37 PM Caitlin Potter wrote: > > I've shared this on Slack already, hoping to get a few more eyes on it via > v8-dev. > https://docs.google.com/document/d/1jn6KFlrnfGwI1f7eLS7QQ7STzN58voONCk4TuCMt7es/edit?usp=sharing > > The primary goal here is to enable the option to

Re: [v8-dev] V8 versions supports

On Sun, Oct 22, 2023 at 11:32 AM Meir Shpilraien wrote: > > Hey, > > According to the documentation, around every 4 weeks a new stable version is > tagged and the maintenance of the previous version is stopped. This is why we > should make sure we are always up to date with the latest stable

Re: [v8-dev] How do I submit a bugfix for multiple branches?

On Wed, Oct 18, 2023 at 12:22 AM Luke Albao wrote: > > Hello! > > I'm a node user, and I've traced a small bug to a change back in v8.x of v8. > I'd like to submit a patch, but I'm unsure of the best way to go about this. > It would be great to have this affect any downstream release that is

Re: [v8-dev] Build failure on mac

FWIW, I don't think the #include of uchar.h is necessary. Looking at commit 182d9c05e78 it was added to get the definition of char16_t but that's an intrinsic type in C++. You only need that header when compiling C code. On Mon, Aug 28, 2023 at 10:54 AM Jakob Kummerow wrote: > > I don't know.

Re: [v8-dev] Is ValueDeserializer secure?

On Thu, Jun 29, 2023 at 4:28 PM 'Kenton Varda' via v8-dev wrote: > > Hi v8-dev, > > We (Cloudflare Workers team) are wondering how V8 feels about the security of > the ValueDeserializer API. Do you believe it's safe to parse > possibly-malicious input with this? My understanding is that Chrome

Re: [v8-dev] Extremely long optimizer pass

On Thu, Jun 29, 2023 at 2:02 AM Andrey Sidorov wrote: > > Hi, > > Crossposting from node, issue https://github.com/nodejs/node/issues/48581 > > We have a CPU spike in a node process while no JS is being executed. The time > is likely spent in an optimiser thread. > > Steps to reproduce: run

Re: [v8-dev] Terminate an async function using TerminateExecution

On Tue, Jun 27, 2023 at 12:57 PM Meir Shpilraien wrote: > > Hey > > I notice that when I terminate an async function using TerminateExecution the > promise that was suppose to be resolved/rejected as a result of this async > function termination are not resolved nor rejected. To demonstrate,

Re: [v8-dev] Limit memory usage by all isolates

On Tue, May 9, 2023 at 3:05 PM Meir Shpilraien wrote: > > I know I can limit the memory used by a single isolate. But assuming my > application creates more than one isolate. Is there a way to put a global > limit such that the memory used by all isolates combined will not bypass this > limit?

Re: [v8-dev] V8 CVE calrigication

On Sun, Mar 12, 2023 at 9:44 AM Meir Shpilraien wrote: > > Hello v8-dev, > > I saw some CVE descriptions which look like this: > > Type confusion in V8 in Google Chrome prior to 111.0.5563.64 allowed a remote > attacker to potentially exploit heap corruption via a crafted HTML page. > (Chromium

Re: [v8-dev] V8 11.2 fails to compile on VS2022, with conflicting C++20 and C++17 compiler arguments

On Tue, Feb 28, 2023 at 3:15 AM Ben Ernst wrote: > > use_custom_libcxx=false > > [snip] > > [exec] cl : Command line warning D9025 : overriding '/std:c++20' with > '/std:c++17' > [exec] C:\88082ece\v8\include/cppgc/allocation.h(8): fatal error C1083: > Cannot open include file:

Re: [v8-dev] Build error: static_assert failed due to requirement 'std::is_standard_layout::value'

On Mon, Dec 19, 2022 at 1:43 PM Immanuel Haffner wrote: > > I am currently facing issues building V8 at version 11.0.226 for embedding in > my application. Compilation of several source files fails with the following > error: > >

Re: [v8-dev] Looking for help understanding StackFrameIterator

On Wed, Sep 14, 2022 at 6:16 PM Nathan Sandum wrote: > > I've been writing some code to trace through the stack in the middle of a > WebAssembly execution. Mostly I've just been using > v8::internal::GetCurrentStackPosition() and sorta just guessing when to stop, > but this is probably

Re: [v8-dev] Building *.dll , *.lib from v8

On Sun, Aug 28, 2022 at 6:20 PM 'Robert Moeller' via v8-dev wrote: > > Hi everyone! I work under Windows. Can I obtain somewhere a bit of > information how to compile v8 into static or (and) dynamic library in order > to embed them in my application? I don't develop on Windows but I believe it

Re: [v8-dev] How to build own v8 fork

On Fri, Jun 3, 2022 at 7:08 PM Simon Koch wrote: > > I have forked the v8 repository and made some changes. Now I want to build it. > However, I cannot find any documentation concerning this use case. > Every documentation starts with using fetch, however, fetch does not offer > the option to

Re: [v8-dev] Would it be possible to have v8 for armv6z -> RaspberryPi 0[W]

On Mon, Feb 21, 2022 at 2:01 AM Gérard Vidal wrote: > > Hi @Camillo, > I have launched a compilation on my PiZeroon the source dowloaded from the > nodejs site, without any modification and after one day and a half it fails > in v8 compilationon this command : > ```shell > g++ -o >

Re: [v8-dev] Getting currently executing script from microtask

On Fri, Jul 23, 2021 at 2:07 AM Shivan wrote: > > (might be better suited to v8-dev) > > For a research project, I'm trying to get the currently executing script's ID > from a microtask in V8 isolate.cc. I've thought of the following two > approaches: > > if (!this->context().is_null()) { >

Re: [v8-dev] SharedFunctionInfo::IsUserJavaScript() coredump on native function

On Tue, May 18, 2021 at 3:03 AM Sam Cao wrote: > > Hi Camillo, > > Thank you for the quick response. I failed to find any public API that could > meet my need, so I had to use the internal one. > > The requirement I got is: To tell whether a function is a user defined JS > function or not. > >

Re: [v8-dev] Module Evaluate() return value

On Sat, May 15, 2021 at 2:19 AM stuta wrote: > > New v8 return promise from module evaluate. My problem is that the promise > result is always "undefined". This code worked before the new promise return. > > Here's a simplified example. > > auto result =

Re: [v8-dev] CallSite enhancements

On Wed, Feb 17, 2021 at 5:42 AM Alex Kodat wrote: > > It was pretty trivial to add these so it doesn't seem like there's a major > downside, so does it seem reasonable to submit a CL for this? I'd offer to > fix up the doc at https://v8.dev/docs/stack-trace-api if it does, but, of > course, I

Re: [v8-dev] fly.rs new build

On Wed, Feb 10, 2021 at 2:28 PM saber mesgari wrote: > > Hi, > > Project fly.rs is a rust project using V8 to run JavaScript code as > applications. It support HTTP requests and methods like: "addEventListener". > But now the project has been archived and the last valid build is for two >

Re: [v8-dev] x87 Support: assertion failure in register-allocator

On Mon, Oct 5, 2020 at 10:19 AM Konrad Schwarz wrote: > > To support a product based on the Intel Quark, which only has an x87 FPU and > no MMX/SSE/AVX, a colleague of mine resurrected the x87 port. > > This was done successfully for node.js version 8 -- unfortunately, I don't > know to which

Re: [v8-dev] Problems with Mac Compilation

On Mon, Sep 28, 2020 at 12:55 AM Artur Ventura wrote: > > Hi there > > I've been trying to compile and run v8 on a mac but I'm having problems. > > I used the basic compilation args.gn generated by `gn args > out.gn/x64.release.sample`: > > ``` > is_component_build = false > is_debug = true >

Re: [v8-dev] Building V8 then linking to it fails (OS X)

On Tue, Jun 2, 2020 at 1:14 PM Al Mo wrote: > > Hi all, > > I'm trying to build v8 from source in order to test/submit a cl later on, > > I followed this: https://v8.dev/docs/build > > It builds fine and in the end I have an 'out/x64.release' directory, which in > turn has the 'libv8_libbase.a'

Re: [v8-dev] "Internal error. Icu error" when upgrading from v8 7.3 to 7.8

On Wed, May 6, 2020 at 5:15 AM 'nightpool' via v8-dev wrote: > > Thanks Ben! I'm using 7.8 since that's what the last contributor who was > working on the upgrade upgraded to, in December. I'll try and figure out what > releases the bug you linked made it into. How can I tell if I'm using the

Re: [v8-dev] "Internal error. Icu error" when upgrading from v8 7.3 to 7.8

On Mon, May 4, 2020 at 6:28 PM 'nightpool' via v8-dev wrote: > > Hi, i'm trying to update the mini_racer ruby extension to use a more modern > version of v8, and i'm running into a problem where Number.toLocaleString is > throwing "TypeError: Internal error. Icu error." > > The error is

Re: [v8-dev] Re: Building with MSVC

On Sat, Mar 14, 2020 at 3:03 PM Chris Fischer wrote: > > I'd like to chime in for continued support of MSVC (with build regression > automated tests). We embed V8 in a large project with many third party code > components. Many of these components are not available as source so building >

Re: [v8-dev] Port Node N-API addon to V8 standalone

On Tue, Jan 21, 2020 at 6:40 AM Dominic Cerisano wrote: > > Hi there, > > I just built a N-API addon for Node, and now have a requirement to run it in > a V8 standalone build (without Node). > > From what I understand N-API is a wrapper around V8, so hoping a V8 > standalone addon should be

Re: [v8-dev] Sharing Memory between Host and WebAssembly Module

On Wed, Jan 15, 2020 at 5:05 PM Immanuel Haffner wrote: > To make my scenario more concrete: > I habe data in main memory, easily tens or hundreds of gigabytes. Programs > fly in requiring to access the data. Some programs are short-lived, some will > take a long time to compute. Some programs

Re: [v8-dev] IsUndefined and IsNull segfault when compiled without V8_ENABLE_CHECKS

On Thu, Dec 26, 2019 at 2:53 PM Jeroen Ooms wrote: > > On Thu, Dec 26, 2019 at 10:36 AM Ben Noordhuis wrote: > > > > On Mon, Dec 23, 2019 at 11:42 PM Jeroen Ooms wrote: > > > > > > Hi! > > > > > > I maintain v8 bindings for the R progra

Re: [v8-dev] Re: Looking for people to maintain the MIPS port

On Tue, Dec 24, 2019 at 11:46 AM wrote: > > Hi yuyin: > > I just build nodejs@12.14.0 on mips64el (loongson 3A3000) with > --with-snapshot ，the result was successed. When I run nodejs, it crashed. The > messages are as follows: > > # > # Fatal error in ../deps/v8/src/execution/isolate.cc,

Re: [v8-dev] IsUndefined and IsNull segfault when compiled without V8_ENABLE_CHECKS

On Mon, Dec 23, 2019 at 11:42 PM Jeroen Ooms wrote: > > Hi! > > I maintain v8 bindings for the R programming language. As of recently (I > think 7.9) we started getting segfaults at calls to IsUndefined() and > IsNull(). The problem has gotten more prevalent in 8.1. We're encountered > this

Re: [v8-dev] State of MapWord on HeapObjects during a full GC

On Wed, Jun 19, 2019 at 6:17 PM billti via v8-dev wrote: > > Trying to clarify how the Map tagged pointer on objects is modified during > GC. The MapWord class (objects.h) has the below comment I can't quite parse - > specifically in the parens at the end. Am I reading right that the Map word

Re: [v8-dev] Behavior of GetEmbedderData() for unset fields (how to check if SetEmbedderData() already called)

On Wed, Jun 5, 2019 at 2:46 AM wrote: > > I have some code that sets data using Context::SetEmbedderData(). I'd like > to have an assert verifying that this is only done once on a given V8 > Context, i.e. assert that the embedder data field at a given index is > currently unset. > > Is it

Re: [v8-dev] Rewrite code in V8

On Sat, May 25, 2019 at 12:01 AM wrote: > > Hi all, > > I'm trying to create some automatic rewriting of all the JS code that is > being parsed by Chrome/V8. For that I went to ParseInfo::CreateScript and > just replaced the source string with my rewritten one: > > std::string s =

Re: [v8-dev] Support for retrieving type name on a Stackframe

On Thu, May 16, 2019 at 2:10 PM 'Simon Zünd' via v8-dev wrote: > AFAIK Node 12 will stay at V8 API compatibility of 7.5 (the current V8 > version being worked on is 7.6). Meaning your change won't make it into Node > 12. Note that the contribution process for Node itself differs from V8.

Re: [v8-dev] Tracking Script which are executing microtasks by ScriptId

On Tue, Apr 2, 2019 at 8:38 PM wrote: > I am trying to track which scripts are responsible for executing microtasks > (by ScriptId). > > I am instrumenting MicrotaskQueue::RunMicrotasks to show when microtask > execution begins and ends, but I want to track which script is responsible > for

Re: [v8-dev] Re: Looking for people to maintain the MIPS port

On Tue, Feb 26, 2019 at 12:34 PM Michael Hablich wrote: > great news! Loongson (on CC) is indeed willing and ready to maintain V8's > MIPS ports, not including Big Endian though. Thanks! For my understanding, that means mipsel and mipsel64 but no more mips and mips64? -- -- v8-dev mailing

Re: [v8-dev] V8 Embeded Application throws Exception randomly. [RangeError: Maximum call stack size exceeded]

On Fri, Nov 9, 2018 at 7:29 PM Sad al Abdullah wrote: > Is there any way to get the current stack_size ? On Windows? https://docs.microsoft.com/en-us/windows/desktop/api/processthreadsapi/nf-processthreadsapi-getcurrentthreadstacklimits If you mean V8's default value on your platform, `d8

Re: [v8-dev] V8 Embeded Application throws Exception randomly. [RangeError: Maximum call stack size exceeded]

On Thu, Nov 8, 2018 at 10:31 PM wrote: > Hi, > > I am using V8 in an asp.net application. I wrote some c++ code to expose some > V8 functionality to C# via PInvoke. I create V8 Isolate object (in future > we will use isolate pool for reuse.) for every request and dispose end of > the

Re: [v8-dev] Generating machine code for typed javascript

On Wed, Oct 24, 2018 at 9:39 AM Hayden Livingston wrote: > I want to use V8 as a scripting language for my game but want to generate > machine code directly without any type profiling or interpreter. > > Is it possible to tell v8 the types it is seeing in the source are 100% typed > like

Re: [v8-dev] Re: Where "locale" variable comes from? Having "Failed to create ICU number format, are ICU data files missing?"

On Thu, Jun 7, 2018 at 9:16 AM, wrote: > That's weird. I've checked udata_setCommonData to be called succesfully. > However in Release build i can see default_locale is received wrong: > "en_US_POSIX" and then converted to "en-US". ICU queries `setlocale(LC_MESSAGES, NULL)` and the LANG,

Re: [v8-dev] Guidance required

On Wed, May 23, 2018 at 12:12 PM, wrote: > hello my name is krishna satya . I want to contribute to v8. Please guide me > in contributing See https://github.com/v8/v8/wiki/Contributing and if you're looking for bugs to work on, visit

Re: [v8-dev] Where do I change a compiler flag?

On Mon, Mar 26, 2018 at 1:27 AM, Radhesh Krishnan K wrote: > How do I disable -Werror=sign-compare flag so that I can compile the V8 > successfully? I believe you can pass treat_warnings_as_errors=false to tools/dev/v8gen.py to disable the -Werror flag. -- --

Re: [v8-dev] Building with GCC

On Wed, Feb 14, 2018 at 12:59 AM, wrote: > I'm working on a RISC-V target for V8. Currently the LLVM toolchain for > RISC-V is less mature and accessible than the GCC one (which is available > upstream as of GCC 7.1 or thereabouts), so I'm wondering if and how I might >

Re: [v8-dev] Side effect of increasing max_old_space_size

On Tue, Feb 6, 2018 at 8:43 PM, wrote: > Hi v8 devs, > I'm working on VSCode and looking into on large file (Gigabyte files) > improvement. As VSCode is Electron based and currently we load the whole > file into memory (we can do mmap), we easily run into heap size limit.

Re: [v8-dev] Threading inside V8

On Sat, Oct 21, 2017 at 8:28 AM, Vaibhav Pande wrote: > From what I have seen till now the there is a thread pool created which > waits on the DefaultPlatform::queue_ and a task is popped by taking a > lock(work sharing). That's the default Platform implementation;

Re: [v8-dev] Sampler - pthread_kill(platform_data()->vm_tid(), SIGPROF);

On Wed, Oct 18, 2017 at 5:04 PM, wrote: > Hello All, > > I have a multi-threaded application that embeds v8 and starts up quite a few > isolates. Each isolate is locked by a global locker (not the same global > locker, each has it's own) that works without

Re: [v8-dev] ReferenceError: arguments is not defined

On Mon, Oct 9, 2017 at 11:32 PM, wrote: > Hi, I'm getting errors are a pretty basic use case, trying to run the > attached v8.log: > > $ ./tools/plot-timer-events v8.log > > /Users/colbyblair/code/v8/tools/profviz/stdio.js:28: ReferenceError: > arguments is not defined > >

Re: [v8-dev] v8 postmortem debug: get Context from Isolate using offsets

On Mon, Aug 21, 2017 at 10:38 PM, wrote: > I'm developing some features for llnode, a LLDB plugin that allows better > postmortem debugging for NodeJS, and while developing one of those features > I needed to access a Context object from an Isolate object, which is very >

Re: [v8-dev] Undefined backtraces when debugging with D8 and GDB JIT enabled

On Mon, Jul 31, 2017 at 10:59 AM, Patrick Spiegel wrote: > When debugging D8 with GDB (gdbjit enabled) major parts of the backtrace are > still marked with ??. > > #0 0x01eab2ef in StringMark::StringMark (this=0x7fffdc48, > other=...) at ../src/mark.cc:214

Re: [v8-dev] A fatal error of wasm with my own built node with v8 6.1.0

On Thu, Jul 27, 2017 at 9:34 AM, aseaday lonsdale wrote: > I replace the directory deps/v8 with the node 6.1.0 source code and build a > node. When I try to test new features of WebAssembly I met an error > > ```javascript >> WebAssembly.compile(arrayBuffer) > Promise { > ,

Re: [v8-dev] Is it possible to serialize a funciton pointer or a generic JS object to a string?

On Mon, Jul 17, 2017 at 3:27 PM, wrote: > Hello, > > I am trying to pass a function pointer through a string pipe and recover it > on the other end. It seems that if serialization to a string is possible and > I can deserizalize the string in C++ to a v8::Object or a

Re: [v8-dev] Bug Report: Low level string function does not check for null pointer argument

On Thu, Jul 13, 2017 at 3:21 AM, Sam Weaver wrote: > > If this isn't the right place for this, please direct me to where I can > submit a proper bug report. Keep in mind I'm a web developer with limited > knowledge of lower level systems programming, so please forgive me if

Re: [v8-dev] Requesting try job access

On Thu, Jul 13, 2017 at 1:42 AM, Adam Klein wrote: > Hi Ben, > > Indeed this is a change in how Gerrit works compared to Rietveld, and my > understanding is that all contributors who wish to be able to land their own > changes need to request V8 tryjob access (CCing machenbach

[v8-dev] Requesting try job access

Hi, I'd like to request try job access so that I can land my CLs (again - used to work before the switch.) CLAs are on file for me personally (i...@bnoordhuis.nl) as well as from my employer, IBM. Should other Node.js contributors request access individually or can we get some blanket stamp of

Re: [v8-dev] Re: Contributing to V8

On Sat, May 27, 2017 at 10:25 PM, wrote: > Hi Ben, > > Thanks a lot for your reply. > > First of all, how will I get to know which patches need to be ported. > Second, how will I test if the port worked fine? > > A very simple explanation would be really appreciated. :) >

Re: [v8-dev] Re: Contributing to V8

On Fri, May 26, 2017 at 3:03 PM, wrote: > Hi, > > Can anyone point in some direction to work on v8. > Please. > I did have a look a lot of issues. > > For e.g. how about something like this - >

Re: [v8-dev] V8 Library size in 5.3 version.

On Fri, May 12, 2017 at 12:07 PM, wrote: > We are working on the V8 version 5.3.332 & we choose to compile it as static > library. We compiled the library with the following configuration. > > MAKE_TARGET="android_$BUILD_LIB_VERSION.$BUILD_MODE" > make $MAKE_TARGET -j1

Re: [v8-dev] What is the difference between JSConstructEntryStub and JSEntryStub?

On Fri, Mar 3, 2017 at 11:57 AM, Dmitriy - wrote: > What is the difference between JSConstructEntryStub and JSEntryStub? One is a regular function or method call, the other a `new ClassName(...)` call. The calls have different semantics, hence different stubs. -- --

Re: [v8-dev] How to extract function address from a v8::Local object

On Sun, Feb 26, 2017 at 8:50 PM, wrote: > Thank you Ben! > > So I can use the GetIdentityHash() function to obtain an ID of the object. > But will this ID be different, if I have two different handlerFunction > object but they are trying to call the same function but with

Re: [v8-dev] How to extract function address from a v8::Local object

On Sun, Feb 26, 2017 at 2:18 AM, wrote: > Dear all, > > I'm doing research on web java-script event execution with chromium, and I'm > trying to log the function address of javascript listener function. > > Currently I have found the v8::Local object that holds the > listener

Re: [v8-dev] Backporting a GC Crash Fix from 5.0 to 4.9

On Tue, Jan 24, 2017 at 4:43 PM, Eric Lawless wrote: > Thanks Ben! This app supports a few million users still on Windows XP, so > upgrading isn't an option for us quite yet. I'll investigate using V8 from > one of the Node LTS releases. > > It looks like the latest Node v4

Re: [v8-dev] Backporting a GC Crash Fix from 5.0 to 4.9

On Tue, Jan 24, 2017 at 4:25 AM, Eric Lawless wrote: > Hello! > > I've been investigating a frequent crash within the garbage collector of V8 > v4.9.385 in our CEF 3 application (on Chromium v49.0.2623.110, the last > version with Windows XP and Vista support). It looks like

Re: [v8-dev] Efficient property fetch from C++

On Sat, Dec 31, 2016 at 10:44 AM, wrote: > I'm wondering how to efficiently Get a property of an Object from within > C++. Let's say that I have a Local that refers to a JS object with > property "method". In JavaScript land you would access this property > efficiently

Re: [v8-dev] Why sort slower in v8 vs js?

On Wed, Oct 26, 2016 at 3:42 AM, Samantha Krieger wrote: > OK I see. My bad on that assumption. > > The truth is I actually wanted to write a native node module to do a lot of > manipulation of dynamic JS objects (from a few to many thousand properties > of varying types). I

Re: [v8-dev] What is purpose of Builtins::Generate_Adaptor?

On Wed, Oct 12, 2016 at 1:25 PM, Dmitriy - wrote: > Hi, > I try to run "print(1)" on d8 and I get following calls: > > (in runtime) > ... -> LoadICStub -> CallICStub -> Builtin::Generate_Call -> ... -> > Builtins::Generate_Adaptor > > I don't understand who and why calls

Re: [v8-dev] What is an extra frame when I try to call CallICStub?

On Mon, Oct 10, 2016 at 7:33 AM, Dmitriy - wrote: > In ic.cc: > > // The IC code is either invoked with no extra frames on the stack > // or with a single extra frame for supporting calls. > enum FrameDepth { NO_EXTRA_FRAME = 0, EXTRA_CALL_FRAME = 1 }; > > And CallIC: > >

Re: [v8-dev] What is handler chain in jsentry stub?

On Thu, Oct 6, 2016 at 10:19 AM, Dmitriy - wrote: > Hi, I can't understand what is it? > > > // Invoke: Link this frame into the handler chain. > __ bind(); > __ PushStackHandler(); > > > void MacroAssembler::PushStackHandler() { > // Adjust this code if not the case.

Re: [v8-dev] Where does d8 set up builtin functions?

On Thu, Oct 6, 2016 at 12:17 PM, Dmitriy - wrote: > For example d8 has "print" function. > > I know that all bootstrapping business is in bootstrapper.cc > > Where v8 adds "print" function to global script context so that it can call > it? print() is defined in d8.cc, it's

Re: [v8-dev] How v8 uses return addres that stores in frame in ICs?

On Thu, Sep 29, 2016 at 10:53 AM, Dmitriy - wrote: > Hi, > > The v8 frame for callee (all for x64 platform) for full-codegen is: > > -- > return_addr > saved frame ptr > context > JsFunction > -- > > You can see description in frames.h file -

Re: [v8-dev] VariableProxy ast class

On Tue, Sep 13, 2016 at 9:54 AM, Dmitriy - wrote: > Hello all, > > I see at FullCodegen and VariableProxy class. > What does it mean? > > It calls EmitVariableLoad in FC but I can't understand for what purposes it > has made. VariableProxy is a layer of indirection between

Re: [v8-dev] When we call global function in d8 what does inline caches do?

On Fri, Sep 9, 2016 at 11:37 AM, Dmitriy - wrote: > For example: > > d8> print(1) > > Fullcodegen call EmitCallWithLoadIC. > > print is a global function. Why v8 emit IC for this case? And how IC works > in this case? 'print' is looked up on the global object. Your example

Re: [v8-dev] Adding compile/run-time marks to functions

On Tue, Aug 30, 2016 at 5:47 AM, wrote: > Hi, > I am working on a Chromium project, for which I need to mark JavaScript > functions The main goal is to be able to determine what JS functions are > being run in real time (i.e. what source code is being run in real time). >

Re: [v8-dev] What is a root register and root array?

On Tue, Aug 9, 2016 at 12:54 PM, Dmitriy - wrote: > Hi, > > in x64 implementation I can see that code > > const Register kRootRegister = { 13 }; // r13 (callee save). > // Actual value of root register is offset from the root array's start > // to take advantage of

Re: [v8-dev] What is the correct procedure for forking a process with a running isolate?

On Sat, Jul 16, 2016 at 7:18 AM, wrote: > Is there a correct procedure or methods I need to run on my context after a > process is forked? > > I am thinking something needs to re-establish the internal threads v8 > expects are running. > > Is forking a process with an

Re: [v8-dev] How v8's x64 implementation calls function

On Sat, Jul 9, 2016 at 12:00 PM, Dmitriy - wrote: > Hello, > > I try to understand what happens in this code: > When you call something you are writing: > __ Call(builtin, RelocInfo::CODE_TARGET); > > What how it looks: > void MacroAssembler::Call(Handle code_object, >

Re: [v8-dev] What does JSEntryStub do?

On Tue, Jun 7, 2016 at 12:35 PM, wrote: > Thx Ben. > Where can I find V8 internal calling conventions description? In the code. :-) (Only serious. I don't think it's documented anywhere.) -- -- v8-dev mailing list v8-dev@googlegroups.com

Re: [v8-dev] What does JSEntryStub do?

On Tue, Jun 7, 2016 at 12:08 PM, wrote: > Hi all! > > I'm continuing to port v8 on my architecture. > I implement some stuff and now I can insert call C printf function into > stubs code and run d8. > Something like this: > void JSEntryStub::Generate(MacroAssembler* masm) {

Re: [v8-dev] constant folding in regular expression

On Fri, Jun 3, 2016 at 7:34 AM, Yun Ji Kim wrote: > Does v8 support constant folding in regular expression? > > /foo/.exec("foo"); shouldn't be folded to "foo"? I don't think V8 does that or even can. RegExp.prototype.exec can be monkey-patched by user code and the default

Re: [v8-dev] What address in code->entry?

On Thu, Jun 2, 2016 at 3:37 PM, wrote: > Hello guys. > > src/execution.cc:57 > MUST_USE_RESULT MaybeHandle Invoke(Isolate* isolate, bool > is_construct, >Handle target, >Handle receiver,

Re: [v8-dev] How can I disable installNatives in Genesis?

On Thu, May 26, 2016 at 9:32 AM, wrote: > What is about InstallExtraNatives? That only does something when you have v8_extra_library_files set. -- -- v8-dev mailing list v8-dev@googlegroups.com http://groups.google.com/group/v8-dev --- You received this message because

Re: [v8-dev] How can I disable installNatives in Genesis?

On Thu, May 26, 2016 at 9:03 AM, wrote: > Hi all. > > I see that that v8 install some codes during the genesis process: > > InitializeNormalizedMapCaches(); > > if (!InstallNatives(context_type)) return; > > MakeFunctionInstancePrototypeWritable(); > > if

Re: [v8-dev] Why v8 call lithium's code before full-codegen?

On Tue, May 24, 2016 at 10:19 AM, wrote: > How can I disable that behaviour? > I'm porting v8 to new architecture and I don't want to implement > Hydrogen/Lithium before full-codegen. I don't think you can, or at least not easily. A lot of the internal infrastructure

Re: [v8-dev] Why v8 call lithium's code before full-codegen?

On Tue, May 24, 2016 at 9:59 AM, wrote: > Hi all. > > My full-codegen.cc: > > bool FullCodeGenerator::MakeCode(CompilationInfo* info) { > UNREACHABLE(); > Isolate* isolate = info->isolate(); > ... > > I run d8 and I see in call stack > > #5 0x0258c1f0 in >

Re: [v8-dev] Did something change in V8 46 (from 45) that would cause weak references to behave differently?

On Tue, Apr 12, 2016 at 7:11 PM, wrote: > Hi, > > we (Microsoft VS Code team) are tracking down a very weird native crash in > our use of node.js (5.10.0, V8 46) that only ever shows up since we updated > from node.js 4.x (V8 45). It seems that changes (around the

Re: [v8-dev] Is there any bitmap class in v8 utils?

On Thu, Apr 7, 2016 at 10:55 AM, wrote: > Hi guys! > > I'm writing my own implementation of assembler for v8. > I need to have bitmap class with following methods: > > void clear() > > bool at(uint32_t) > > void set_bit(uint32_t) > > > and constructor bitmap(size) > > > Is

Re: [v8-dev] Type inference in V8

On Sun, Mar 20, 2016 at 9:55 AM, wrote: > Thanks for the answer. > I know about inline cache, but I'm more curious about "polymorphism type" > variables within one function. > For example, there is a variable changes its type in a function. > Is there any type inference

Re: [v8-dev] Type inference in V8

On Sun, Mar 20, 2016 at 9:14 AM, wrote: > Hi, > I want to do some researches about type inference system and related > optimizations for dynamic types in V8. > Can someone give me some hints about where or which file I should start > looking for? > Thanks, > > Best

Re: [v8-dev] how to build V8 on windows with post mortem debugging info?

On Wed, Mar 16, 2016 at 9:47 AM, wrote: > Hey Guys, > I need to do post mortem debugging on windows, but can't find a way to build > V8 with post mortem debugging info. > > I tried to change follow setting to true in every place I know, but still > doesn't help. > >

Re: [v8-dev] Undestading V8 full-codegen compilation pipeline

On Thu, Mar 3, 2016 at 8:17 AM, wrote: > Ok, that is I can implement full-codegen-MY-PLATFORM.cc and I don't have to > implement codegen-MY-PLATFORM.cc to port full-codegen on my platform? > > I'm sorry for this obvious things :) $ find src/ -name \*-x64.cc

Re: [v8-dev] Oddball::kArgumentMarker & out/Release/obj/gen/debug-support.cc

On Wed, Feb 17, 2016 at 9:04 PM, Ben Noordhuis <i...@bnoordhuis.nl> wrote: > On Wed, Feb 17, 2016 at 6:59 PM, Peter Schow <psc...@gmail.com> wrote: >> I'm seeing a V8 build failure with one of the files generated at build >> time: >> >> out/Releas

Re: [v8-dev] Oddball::kArgumentMarker & out/Release/obj/gen/debug-support.cc

On Wed, Feb 17, 2016 at 6:59 PM, Peter Schow wrote: > I'm seeing a V8 build failure with one of the files generated at build > time: > > out/Release/obj/gen/debug-support.cc:45:35: error: 'kArgumentMarker' is > not a member of 'v8::internal::Oddball' > int

Re: [v8-dev] v8 understanding full-codegen

On Mon, Feb 15, 2016 at 1:57 PM, wrote: > Hi all! > > I'm trying to understand the full-codegen compiler in v8. > > Ok, I've taken the full-codegen code for x64 architecture. > > As I can see in code, it does following things: > Build x64 frame > Allocate locals > And then

  1   2   >