V8 (branches 9.5 and 9.6) build fails on multiple platforms

[Angelo Dell'Aera]

Hi all,

just to want to report an issue I am experiencing in branches 9.5 and 9.6. The 

build fails on multiple platforms (tested on Ubuntu 18.04, Ubuntu 20.04, Gentoo

Linux and MacOS ). The error I get is the same on all the platforms and is reported 

below for convenience. This does not happen with branches <= 9.4

FAILED: obj/v8_base_without_compiler/heap.o

1481../../third_party/llvm-build/Release+Asserts/bin/clang++ -MMD -MF obj/v8_base_without_compiler/heap.o.d -DUSE_UDEV -DUSE_AURA=1 -DUSE_GLIB=1 -DUSE_NSS_CERTS=1 -DUSE_OZONE=1 -DUSE_X11=1 -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_GNU_SOURCE -DCR_CLANG_REVISION=\"llvmorg-14-init-4918-ge787678c-1\" -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D_FORTIFY_SOURCE=2 -DCR_SYSROOT_HASH=95051d95804a77144986255f534acb920cee375b -DNDEBUG -DNVALGRIND -DDYNAMIC_ANNOTATIONS_ENABLED=0 -DV8_TYPED_ARRAY_MAX_SIZE_IN_HEAP=64 -DENABLE_GDB_JIT_INTERFACE -DENABLE_MINOR_MC -DV8_INTL_SUPPORT -DV8_ATOMIC_OBJECT_FIELD_WRITES -DV8_ATOMIC_MARKING_STATE -DV8_ENABLE_LAZY_SOURCE_POSITIONS -DV8_SHARED_RO_HEAP -DV8_WIN64_UNWINDING_INFO -DV8_ENABLE_REGEXP_INTERPRETER_THREADED_DISPATCH -DV8_SNAPSHOT_COMPRESSION -DV8_ENABLE_WEBASSEMBLY -DV8_ALLOCATION_FOLDING -DV8_ALLOCATION_SITE_TRACKING -DV8_ADVANCED_BIGINT_ALGORITHMS -DV8_INCLUDE_RECEIVER_IN_ARGC -DV8_DEPRECATION_WARNINGS -DV8_IMMINENT_DEPRECATION_WARNINGS -DCPPGC_CAGED_HEAP -DV8_TARGET_ARCH_X64 -DV8_HAVE_TARGET_OS -DV8_TARGET_OS_LINUX -DV8_RUNTIME_CALL_STATS -DU_USING_ICU_NAMESPACE=0 -DU_ENABLE_DYLOAD=0 -DUSE_CHROMIUM_ICU=1 -DU_ENABLE_TRACING=1 -DU_ENABLE_RESOURCE_TRACING=0 -DU_STATIC_IMPLEMENTATION -DICU_UTIL_DATA_IMPL=ICU_UTIL_DATA_FILE -I../.. -Igen -I../../include -Igen/include -I../../third_party/icu/source/common -I../../third_party/icu/source/i18n -I../../third_party/zlib -fno-delete-null-pointer-checks -fno-ident -fno-strict-aliasing --param=ssp-buffer-size=4 -fstack-protector -funwind-tables -fPIC -pthread -fcolor-diagnostics -fmerge-all-constants -fcrash-diagnostics-dir=../../tools/clang/crashreports -mllvm -instcombine-lower-dbg-declare=0 -m64 -march=x86-64 -msse3 -Wno-builtin-macro-redefined -D__DATE__= -D__TIME__= -D__TIMESTAMP__= -ffile-compilation-dir=. -no-canonical-prefixes -Wall -Werror -Wextra -Wimplicit-fallthrough -Wunreachable-code-aggressive -Wthread-safety -Wextra-semi -Wno-missing-field-initializers -Wno-unused-parameter -Wloop-analysis -Wno-unneeded-internal-declaration -Wenum-compare-conditional -Wno-psabi -Wno-ignored-pragma-optimize -Wno-builtin-assume-aligned-alignment -Wno-unused-but-set-parameter -Wno-unused-but-set-variable -Wmax-tokens -fno-omit-frame-pointer -g0 -Wheader-hygiene -Wstring-conversion -Wtautological-overlap-compare -Wmissing-field-initializers -Wno-shadow -Wunreachable-code -Wshorten-64-to-32 -O3 -fdata-sections -ffunction-sections -fno-unique-section-names -fvisibility=default -Wexit-time-destructors -std=c++14 -fno-trigraphs -Wno-trigraphs -fno-exceptions -fno-rtti --sysroot=../../build/linux/debian_sid_amd64-sysroot -c ../../src/heap/heap.cc -o obj/v8_base_without_compiler/heap.o

1482In file included from ../../src/heap/heap.cc:5:

1483../../src/heap/heap.h:1751:22: error: definition of implicit copy constructor for 'GCCallbackTuple' is deprecated because it has a user-provided copy assignment operator [-Werror,-Wdeprecated-copy-with-user-provided-copy]

1484 GCCallbackTuple& operator=(const GCCallbackTuple& other) V8_NOEXCEPT;

1485 ^

1486../../build/linux/debian_sid_amd64-sysroot/usr/lib/gcc/x86_64-linux-gnu/10/../../../../include/c++/10/ext/new_allocator.h:150:23: note: in implicit copy constructor for 'v8::internal::Heap::GCCallbackTuple' first required here

1487 { ::new((void *)__p) _Up(std::forward<_Args>(__args)...); }

1488 ^

1489../../build/linux/debian_sid_amd64-sysroot/usr/lib/gcc/x86_64-linux-gnu/10/../../../../include/c++/10/bits/alloc_traits.h:512:8: note: in instantiation of function template specialization '__gnu_cxx::new_allocator<v8::internal::Heap::GCCallbackTuple>::construct<v8::internal::Heap::GCCallbackTuple, v8::internal::Heap::GCCallbackTuple>' requested here

1490 __a.construct(__p, std::forward<_Args>(__args)...);

1491 ^

1492../../build/linux/debian_sid_amd64-sysroot/usr/lib/gcc/x86_64-linux-gnu/10/../../../../include/c++/10/bits/stl_uninitialized.h:976:17: note: in instantiation of function template specialization 'std::allocator_traits<std::allocator<v8::internal::Heap::GCCallbackTuple>>::construct<v8::internal::Heap::GCCallbackTuple, v8::internal::Heap::GCCallbackTuple>' requested here

1493 __traits::construct(__alloc, __dest, std::move(*__orig));

1494 ^

1495../../build/linux/debian_sid_amd64-sysroot/usr/lib/gcc/x86_64-linux-gnu/10/../../../../include/c++/10/bits/stl_uninitialized.h:1014:7: note: in instantiation of function template specialization 'std::__relocate_object_a<v8::internal::Heap::GCCallbackTuple, v8::internal::Heap::GCCallbackTuple, std::allocator<v8::internal::Heap::GCCallbackTuple>>' requested here

1496 std::__relocate_object_a(std::__addressof(*__cur),

1497 ^

1498../../build/linux/debian_sid_amd64-sysroot/usr/lib/gcc/x86_64-linux-gnu/10/../../../../include/c++/10/bits/stl_uninitialized.h:1028:14: note: in instantiation of function template specialization 'std::__relocate_a_1<v8::internal::Heap::GCCallbackTuple *, v8::internal::Heap::GCCallbackTuple *, std::allocator<v8::internal::Heap::GCCallbackTuple>>' requested here

1499 return __relocate_a_1(std::__niter_base(__first),

1500 ^

1501../../build/linux/debian_sid_amd64-sysroot/usr/lib/gcc/x86_64-linux-gnu/10/../../../../include/c++/10/bits/stl_vector.h:456:14: note: in instantiation of function template specialization 'std::__relocate_a<v8::internal::Heap::GCCallbackTuple *, v8::internal::Heap::GCCallbackTuple *, std::allocator<v8::internal::Heap::GCCallbackTuple>>' requested here

1502 return std::__relocate_a(__first, __last, __result, __alloc);

1503 ^

1504../../build/linux/debian_sid_amd64-sysroot/usr/lib/gcc/x86_64-linux-gnu/10/../../../../include/c++/10/bits/stl_vector.h:469:9: note: in instantiation of member function 'std::vector<v8::internal::Heap::GCCallbackTuple>::_S_do_relocate' requested here

1505 return _S_do_relocate(__first, __last, __result, __alloc, __do_it{});

1506 ^

1507../../build/linux/debian_sid_amd64-sysroot/usr/lib/gcc/x86_64-linux-gnu/10/../../../../include/c++/10/bits/vector.tcc:461:23: note: in instantiation of member function 'std::vector<v8::internal::Heap::GCCallbackTuple>::_S_relocate' requested here

1508 __new_finish = _S_relocate(__old_start, __position.base(),

1509 ^

1510../../build/linux/debian_sid_amd64-sysroot/usr/lib/gcc/x86_64-linux-gnu/10/../../../../include/c++/10/bits/vector.tcc:121:4: note: in instantiation of function template specialization 'std::vector<v8::internal::Heap::GCCallbackTuple>::_M_realloc_insert<void (*&)(v8::Isolate *, v8::GCType, v8::GCCallbackFlags, void *), v8::GCType &, void *&>' requested here

1511 _M_realloc_insert(end(), std::forward<_Args>(__args)...);

1512 ^

1513../../src/heap/heap.cc:6039:26: note: in instantiation of function template specialization 'std::vector<v8::internal::Heap::GCCallbackTuple>::emplace_back<void (*&)(v8::Isolate *, v8::GCType, v8::GCCallbackFlags, void *), v8::GCType &, void *&>' requested here

1514 gc_prologue_callbacks_.emplace_back(callback, gc_type, data);

1515 ^

15161 error generated.

Ciao,

Angelo

[Jakob Kummerow]

Have you run gclient sync after switching branches?

What are your GN args?

--
--
v8-dev mailing list
v8-...@googlegroups.com
http://groups.google.com/group/v8-dev
---
You received this message because you are subscribed to the Google Groups "v8-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to v8-dev+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/v8-dev/d1777904-00b3-4b9a-ba58-c3aa2ab02061n%40googlegroups.com.

[Angelo Dell'Aera]

FYI I maintain a Python V8 wrapper (available at https://github.com/area1/stpyv8) and use 

Github Actions to automatically build new releases on multiple platforms and never experienced 

this issue since the first release (v7.9.317.31). 

To answer your questions, yes "gclient sync" is executed every time the workflow runs and these 

are the GN args I use

gn_args = {

  "v8_deprecation_warnings" : "true",

  "v8_imminent_deprecation_warnings" : "true",

  "v8_enable_disassembler" : "false",

  "v8_enable_i18n_support"  : "true",

  "is_component_build" : "false",

  "is_debug" : "true" if os.environ.get("STPYV8_DEBUG") else "false",

  "use_custom_libcxx" : "false",

  "v8_monolithic" : "true",

  "v8_use_external_startup_data" : "false",

  "v8_enable_pointer_compression" : "false",

  "v8_enable_31bit_smis_on_64bit_arch" : "false"

}

Regards,

Angelo

[Jakob Kummerow]

I'd try to use either both of:

is_clang = false

use_custom_libcxx = false

or neither of them. Looks like the bundled Clang isn't happy with the sysroot's libcxx.

A possible alternative solution for making things work with non-default toolchains is treat_warnings_as_errors = false (and possibly fatal_linker_warnings = false as well).

[Angelo Dell'Aera]

In the last couple of days I tried playing with the GN args as you suggested but no luck. 

Please note that I was able to successfully compile removing both of them as you suggested but I got the following error when I tried running the test suite

ImportError: /usr/lib/python3.9/site-packages/_STPyV8.cpython-39-x86_64-linux-gnu.so: undefined symbol: _ZNSt3__119__shared_weak_countD2Ev

Moreover please be aware that I was able to successfully compile V8 v9.4.146.21 today using the GN args I showed in my previous email and no issues at all with the test suite.

Regards,

Angelo

[Angelo Dell'Aera]

Just an update.

I figured out what the issue was and solved it. The error I got is actually a warning which was treated as 

an error until I added "treat_warnings_as_errors = false" to the GN args. This was enough to successfully

build V8 and the Python wrapper. At a first glance it seems to me the V8 libcxx was updated in branch 9.5+ 

and the updated version emits this warning now.

Thanks for the support and regards,

Angelo