Package Details: v8 8.7.220.25-1

Git Clone URL: https://aur.archlinux.org/v8.git (read-only, click to copy)
Package Base: v8
Description: Google's open source JavaScript engine
Upstream URL: https://v8.dev/
Keywords: compiler interpreter javascript javascript-engine virtual-machine
Licenses: BSD
Conflicts: v8-3.14, v8-3.14-bin, v8-6.7-static, v8-6.8, v8-r, v8-static-gyp, v8-static-gyp-5.4
Submitter: anatolik
Maintainer: None
Last Packager: relrel
Votes: 15
Popularity: 0.000008
First Submitted: 2015-07-01 16:14 (UTC)
Last Updated: 2020-11-26 11:22 (UTC)

Dependencies (4)

Required by (4)

Sources (2)

Pinned Comments

relrel commented on 2020-10-30 11:56 (UTC) (edited on 2021-02-01 20:07 (UTC) by relrel)

Recommendations
  1. Before installing this package, install depot-tools-git and follow its instructions.

  2. To speed-up compilation, build on tmpfs and disable tests:

BUILDDIR=/tmp/makepkg makepkg --nocheck

Or using yay:

BUILDDIR=/tmp/makepkg yay -S --mflags --nocheck v8

Latest Comments

teddy_coder commented on 2021-06-08 08:24 (UTC) (edited on 2021-06-08 08:35 (UTC) by teddy_coder)

I got the same problem as bashee_wang in previous comment. I solved it by chown for directory /opt/depot_tools/. But then I got an error while compiling:

[666/1044] CXX obj/v8_base_without_compiler/js-list-format.o
FAILED: obj/v8_base_without_compiler/js-list-format.o 
../../third_party/llvm-build/Release+Asserts/bin/clang++ -MMD -MF obj/v8_base_without_compiler/js-list-format.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-12-init-5035-gd0abc757-3\" -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D_DEBUG -DDYNAMIC_ANNOTATIONS_ENABLED=1 -DENABLE_DISASSEMBLER -DV8_TYPED_ARRAY_MAX_SIZE_IN_HEAP=64 -DENABLE_GDB_JIT_INTERFACE -DENABLE_MINOR_MC -DOBJECT_PRINT -DVERIFY_HEAP -DV8_TRACE_MAPS -DV8_ENABLE_ALLOCATION_TIMEOUT -DV8_ENABLE_FORCE_SLOW_PATH -DV8_ENABLE_DOUBLE_CONST_STORE_CHECK -DV8_INTL_SUPPORT -DENABLE_HANDLE_ZAPPING -DV8_SNAPSHOT_NATIVE_CODE_COUNTERS -DV8_CONCURRENT_MARKING -DV8_ENABLE_LAZY_SOURCE_POSITIONS -DV8_CHECK_MICROTASKS_SCOPES_CONSISTENCY -DV8_WIN64_UNWINDING_INFO -DV8_ENABLE_REGEXP_INTERPRETER_THREADED_DISPATCH -DV8_SNAPSHOT_COMPRESSION -DV8_ENABLE_CHECKS -DV8_COMPRESS_POINTERS -DV8_31BIT_SMIS_ON_64BIT_ARCH -DV8_DEPRECATION_WARNINGS -DV8_IMMINENT_DEPRECATION_WARNINGS -DV8_REVERSE_JSARGS -DV8_TARGET_ARCH_X64 -DV8_HAVE_TARGET_OS -DV8_TARGET_OS_LINUX -DDEBUG -DDISABLE_UNTRUSTED_CODE_MITIGATIONS -DV8_ENABLE_CHECKS -DV8_COMPRESS_POINTERS -DV8_31BIT_SMIS_ON_64BIT_ARCH -DV8_DEPRECATION_WARNINGS -DV8_IMMINENT_DEPRECATION_WARNINGS -DV8_REVERSE_JSARGS -DCPPGC_CAGED_HEAP -DUSING_SYSTEM_ICU=1 -DICU_UTIL_DATA_IMPL=ICU_UTIL_DATA_STATIC -DUCHAR_TYPE=uint16_t -DU_IMPORT=U_EXPORT -DV8_ENABLE_CHECKS -DV8_COMPRESS_POINTERS -DV8_31BIT_SMIS_ON_64BIT_ARCH -DV8_DEPRECATION_WARNINGS -DV8_IMMINENT_DEPRECATION_WARNINGS -DV8_REVERSE_JSARGS -I../.. -Igen -I../.. -I../../include -Igen -Igen/shim_headers/icui18n_shim -Igen/shim_headers/icuuc_shim -I../../include -I../../third_party/zlib -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 -mllvm -enable-dse-memoryssa=false -m64 -march=x86-64 -Wno-builtin-macro-redefined -D__DATE__= -D__TIME__= -D__TIMESTAMP__= -Xclang -fdebug-compilation-dir -Xclang . -no-canonical-prefixes -Wall -Werror -Wextra -Wimplicit-fallthrough -Wunreachable-code -Wthread-safety -Wextra-semi -Wno-missing-field-initializers -Wno-unused-parameter -Wno-c++11-narrowing -Wno-unneeded-internal-declaration -Wno-undefined-var-template -Wno-psabi -Wno-ignored-pragma-optimize -Wno-implicit-int-float-conversion -Wno-final-dtor-non-final-class -Wno-builtin-assume-aligned-alignment -Wno-deprecated-copy -Wno-non-c-typedef-for-linkage -Wmax-tokens -fno-omit-frame-pointer -g2 -Xclang -debug-info-kind=constructor -gsplit-dwarf -ggnu-pubnames -ftrivial-auto-var-init=pattern -Wheader-hygiene -Wstring-conversion -Wtautological-overlap-compare -Wmissing-field-initializers -Wunreachable-code -Wshorten-64-to-32 -O3 -fno-ident -fdata-sections -ffunction-sections -fvisibility=default -Wexit-time-destructors -Wno-undefined-bool-conversion -Wno-tautological-undefined-compare -std=c++14 -fno-trigraphs -Wno-trigraphs -fno-exceptions -fno-rtti -c ../../src/objects/js-list-format.cc -o obj/v8_base_without_compiler/js-list-format.o
../../src/objects/js-list-format.cc:145:55: error: 'createInstance' is a private member of 'icu_69::ListFormatter'
  icu::ListFormatter* formatter = icu::ListFormatter::createInstance(
                                  ~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~
/usr/include/unicode/listformatter.h:267:27: note: declared private here
    static ListFormatter* createInstance(const Locale& locale, const char* style, UErrorCode& errorCode);
                          ^
1 error generated.
[675/1044] CXX obj/v8_base_without_compiler/elements.o
ninja: build stopped: subcommand failed.
==> ERROR: Произошел сбой в build().
    Прерывание...

bashee_wang commented on 2021-04-19 03:55 (UTC)

Get below error msg, please help: :: (1/1) Parsing SRCINFO: v8 ==> Making package: v8 8.7.220.25-1 (Mon 19 Apr 2021 11:50:19 AM CST) ==> Retrieving sources... -> Found v8.pc -> Found d8 ==> Validating source files with sha512sums... v8.pc ... Passed d8 ... Passed ==> Making package: v8 8.7.220.25-1 (Mon 19 Apr 2021 11:50:21 AM CST) ==> Checking runtime dependencies... ==> Checking buildtime dependencies... ==> Retrieving sources... -> Found v8.pc -> Found d8 ==> Validating source files with sha512sums... v8.pc ... Passed d8 ... Passed ==> Removing existing $srcdir/ directory... ==> Extracting sources... ==> Starting prepare()... mktemp: failed to create file via template ‘/opt/depot_tools/.cipd_client.XXXXXXX’: Permission denied curl: (23) Failure writing output to destination /opt/depot_tools/bootstrap_python3: line 37: bootstrap-3.8.0.chromium.8_bin/python3/bin/python3: No such file or directory cat: /opt/depot_tools/python3_bin_reldir.txt: No such file or directory /opt/depot_tools/vpython3: line 52: /opt/depot_tools/.cipd_bin/vpython3: No such file or directory ==> ERROR: A failure occurred in prepare(). Aborting... error making: v8

sarpik commented on 2020-11-20 12:28 (UTC) (edited on 2020-11-20 12:28 (UTC) by sarpik)

The --mflags "--nocheck" did not work and the check() step failed because of some tests.

To fix it, edit the PKGBUILD and replace the contents of check() { ... } with just check() { echo "hi" } (if using yay, pass in --editmenu and select to edit the PKGBUILD once asked)

relrel commented on 2020-10-30 11:56 (UTC) (edited on 2021-02-01 20:07 (UTC) by relrel)

Recommendations
  1. Before installing this package, install depot-tools-git and follow its instructions.

  2. To speed-up compilation, build on tmpfs and disable tests:

BUILDDIR=/tmp/makepkg makepkg --nocheck

Or using yay:

BUILDDIR=/tmp/makepkg yay -S --mflags --nocheck v8

marcs commented on 2020-10-24 10:24 (UTC)

I had some weird building problems with the last few versions that didn't allowed me to update this package. I tried some tests on a standalone environment but I couldn't fix them. Since someone filed for an ownership claim and got approved by AUR I will stop supporting the package and let someone else try. Good luck and have fun.

marcs commented on 2020-07-25 15:42 (UTC)

Some notes: * Latest version is based on the version on branch-heads from the V8 repositories, I stick to stable, but sometimes I could use beta but I never use dev (check this: https://omahaproxy.appspot.com/). * This is a "generic" configuration, v8 is highly configurable, if you need v8 for some specific integration please check the Google documentation. * Prefer yay/aurman instead of yaourt/pacaur. * Gclient has to download the whole V8 repo, it takes some time to complete (depending on your internet connection). * This package requires around 3/4 gigs of space so be sure to have a large enough /tmp partition if you use yaourt, also be aware that aurman/yay/pacaur will save this package in ~/.cache/(aurman|yay|pacaur)/v8, so be sure to have a /home partition with enough space. * The build and testing process takes some time, I recommend to use at least a 4 cores with 8 threads cpu, better an 8 cores 16 threads cpu. * If you have at least 32GB of RAM you can use a ram partition to speed up the build process (yaourt uses /tmp by default).

somakd commented on 2020-07-19 17:00 (UTC) (edited on 2020-07-19 17:00 (UTC) by somakd)

Cannot resolve "python2-wrapt" and "python2-pylint" dependencies.

gj545rndmu commented on 2019-09-24 11:14 (UTC) (edited on 2019-09-24 11:15 (UTC) by gj545rndmu)

There is a package that uses a standard linker path now: https://aur.archlinux.org/packages/v8-r/

Hence, it can be found by other applications such as the R package V8. If you have other packages that rely on v8 at standard linker paths you might want to try the v8-r package as well.

Big thanks to JanMarvin!

izahn commented on 2019-05-31 16:03 (UTC)

https://gist.github.com/JanMarvin/61ee142b21c394f5c67a55df4f960adc by @JanMarvin allowed me to install the R V8 package. It would be great to incorporate the required changes to the main PKGBUILD here.

gj545rndmu commented on 2019-04-15 09:06 (UTC)

@marcs as you can see in other comments below this v8 package here doesn't work with V8 for R. See issue here: https://github.com/jeroen/V8/issues/58

  • /usr/lib/v8 is not a standard linker paths.
  • Even when trying to use your PKGBUILD with standard linker paths: /usr/lib/libv8.so: undefined symbol: u_strToLower_63

One solution would be to have another package on AUR that we would call v8-R or such. But as JanMarvin pointed out the best solution would be to make it work with this package, right?

What do you think?

marcs commented on 2019-04-10 10:31 (UTC) (edited on 2019-04-10 10:46 (UTC) by marcs)

hey @dflemstr,

I haven't checked that yaourt was dead, good to know thanks. This package is now compatible with aurman and yay.

I was not aware of the pletora of aur helpers nowadays, I personally don't use them, I just clone a package and build it with makepkg on a specific account designed to do that.

dflemstr commented on 2019-04-08 16:03 (UTC)

I would like to remind about this patch (which might be stale now...): https://gist.github.com/dflemstr/2c06e81da38420d29e702c5bd3601afc

Building does not work in a lot of cases now when prepare(), build() and check() are run in different phases, which is generally the case in AUR helpers such as yay, pacaur etc.

Note that both yaourt and pacaur are deprecated and yaourt has had a lot of security issues in the past (see https://itsfoss.com/best-aur-helpers/ for more info) so it might not make sense to recommend using those tools any more.

marcs commented on 2019-03-25 15:53 (UTC) (edited on 2019-03-26 15:29 (UTC) by marcs)

@znmeb you can limit the number of threads from your makepkg.conf file.

For the rest of the building errors you can change the version using one listed on the V8 branches from:

https://chromium.googlesource.com/v8/v8.git/+log/branch-heads/7.4

just change the pkgver variable and you are good to go.

This package is not suited for community IMO, the PKGBUILD is heavily documented to let it know the user that this package needs some configuration (sometimes), the default parameters is what i generally use and it should generally works.

I'm gonna update the PKGBUILD with the missing deps, thanks everyone.

jathroppaPH commented on 2019-03-25 07:58 (UTC)

I am installing this package but it cannot complete the installation.

In file included from ../../src/inspector/v8-inspector-impl.cc:41:
../../src/inspector/v8-debugger-agent-impl.h:124:44: internal compiler error: Segmentation fault
   bool enabled() const { return m_enabled; }
                                            ^
Please submit a full bug report,
with preprocessed source if appropriate.
See  for instructions.
[230/1604] CXX obj/src/inspector/inspector/v8-profiler-agent-impl.o
[231/1604] CXX obj/src/inspector/inspector/v8-inspector-session-impl.o
[232/1604] CXX obj/src/inspector/inspector/value-mirror.o
[233/1604] CXX obj/src/inspector/inspector/v8-runtime-agent-impl.o
[234/1604] ACTION //:run_torque(//build/toolchain/linux:x64)
[235/1604] CXX obj/third_party/googletest/gtest/gtest-matchers.o
[236/1604] CXX obj/third_party/googletest/gtest/gtest-printers.o
[237/1604] CXX obj/third_party/googletest/gtest/gtest.o
[238/1604] SOLINK ./libv8_libplatform.so
ninja: build stopped: subcommand failed.
==> ERROR: A failure occurred in build().
Aborting...

I used the PKGBUILD from here: https://gist.github.com/JanMarvin/61ee142b21c394f5c67a55df4f960adc

This is the full installation report in my system: https://gist.github.com/jethroppaPH/6a40453cc262d32e99bdb3437accbefb

JanMarvin commented on 2019-03-15 14:39 (UTC)

@zbraniecki you need to install the icudtl.dat file. Can be done adding this line to the PKGBUILD:

install -Dm755 $OUTFLD/icudtl.dat ${pkgdir}/usr/lib/v8/icudtl.dat

zbraniecki commented on 2019-03-11 05:15 (UTC)

I tried to call new Intl.PluralRules(); and it crashed with:

d8
V8 version 7.2.502.25
d8> new Intl.PluralRules


#
# Fatal error in , line 0
# Failed to create ICU PluralRules, are ICU data files missing?
#
#
#
#FailureMessage Object: 0x7ffeb91884f0
==== C stack trace ===============================

    /usr/lib/libv8_libbase.so(v8::base::debug::StackTrace::StackTrace()+0x16) [0x7f66d516c046]
    /usr/lib/libv8_libplatform.so(+0x8fab) [0x7f66d5142fab]
    /usr/lib/libv8_libbase.so(V8_Fatal(char const*, int, char const*, ...)+0x170) [0x7f66d5163430]
    /usr/lib/libv8.so(+0x6984ee) [0x7f66d580e4ee]
    /usr/lib/libv8.so(+0x23aea2) [0x7f66d53b0ea2]
    /usr/lib/libv8.so(+0x249eb8) [0x7f66d53bfeb8]
    /usr/lib/libv8.so(+0xa5916b) [0x7f66d5bcf16b]
Received signal 4 ILL_ILLOPN 7f66d515c5ff
[1]    20807 illegal hardware instruction (core dumped)  d8

I do have libicu installed. Is there a way to link it?

znmeb commented on 2019-03-07 08:45 (UTC) (edited on 2019-03-07 18:34 (UTC) by znmeb)

I'm testing it now - is there any way to limit the number of cores it uses? I have eight cores and it fired off 11 jobs and pushed my workstation so hard it shut down on a thermal overload!

Better yet - get enough votes to move the package into "community"


It's crashing in the tests:


Actual call chain length: 8
setAsyncCallStackDepth(maxDepth): 0
reported: 0
Command: /home/znmeb/Projects/v8/src/v8/out.gn/Release/inspector-test /home/znmeb/Projects/v8/src/v8/test/inspector/protocol-test.js /home/znmeb/Projects/v8/src/v8/test/inspector/debugger/set-async-call-stack-depth.js --random-seed=1823301583 --nohard-abort
exit code: -11
--- CRASHED ---
>>> Too many failures, exiting...
=== inspector/debugger/set-blackbox-patterns ===                          
Tests blackboxing by patterns
Pattern parser error: Uncaught SyntaxError: Invalid regular expression: /(foo([)/: Unterminated character class
Paused in
(...):1
Paused in
(...):1
Paused in
qwe:3
baz:3
(...):1
Paused in
bar:3
foo:3
qwe:3
baz:3
(...):1
Paused in
qwe:4
baz:3
(...):1
Paused in
qwe:4
baz:3
(...):1
Paused in
(...):1
Command: /home/znmeb/Projects/v8/src/v8/out.gn/Release/inspector-test /home/znmeb/Projects/v8/src/v8/test/inspector/protocol-test.js /home/znmeb/Projects/v8/src/v8/test/inspector/debugger/set-blackbox-patterns.js --random-seed=1823301583 --nohard-abort
exit code: -11
--- CRASHED ---
[04:05|%  67|+ 10255|- 100]: Done                                    
>>> 10355 tests ran
==> ERROR: A failure occurred in check().
    Aborting...

I can run it again and post a full logfile if that's necessary.

JanMarvin commented on 2019-03-04 21:01 (UTC)

Thank you for your confirmation @ramesesjd. My PKGBUILD disables the use of this libc++ (using use_custom_libcxx=false). I experimented with patchelf to fix the build, but in the end, why bother with a lib that is not required. Later I found that two of these flags (buildtools and libcxx) are used for the Arch chromium build as well, so maybe they could be adapted for the AUR v8 package as well? Not sure about the external_startup_data flag, the glib one is most likely not required.

ramesesjd commented on 2019-03-02 05:28 (UTC)

JanMarvin, thank you for the PKGBUILD. It works for me on Arch: V8 installs just fine and I can install geojsonio in R without complaints about libc++.so being too short.

JanMarvin commented on 2019-03-01 20:36 (UTC)

The following PKGBUILD should work for R-Users. Builds with linux_use_bundled_binutils=false, v8_use_external_startup_data=false, use_glib=false, use_custom_libcxx=false. Not sure about the glib-flag, but building takes ages and I'll figure this out with the next release of v8.

https://gist.github.com/JanMarvin/61ee142b21c394f5c67a55df4f960adc

ramesesjd commented on 2019-02-28 15:13 (UTC)

I too can confirm the same issue that Melierax is having. I am on Arch, fully updated. I've tried using the PKGBUILD povided by a user on from the rstats geojsonio discussion page: https://github.com/ropensci/geojsonio/issues/143#issuecomment-462378725

geojsonio is an R package for which I need to install the V8 R package, which needs v8 installed on the system.

Even using the above PKGBUIL I get the same error: libc++.so: file too short. The R V8 developer believes it is a bug in arch v8 library: https://github.com/jeroen/V8/issues/52

jolars commented on 2019-02-28 11:45 (UTC)

I can confirm the libc++ file issue that Melierax is having. I am on Antergos.

Melierax commented on 2019-02-12 11:06 (UTC)

Hi,

I am trying to install V8 rstats package on a machine running Manjaro Linux. All is updated. I get the following error message (either installing from CRAN or github): /lib/libc++.so: file too short. I have tried to re-install libc++, then compiled it from source but it doesn't solve the issue. The dev thinks it is a bug in the V8 library.

Any idea ?

Thanks

znmeb commented on 2019-02-05 18:53 (UTC)

Build is failing:

==> Sources are ready.
==> Making package: v8 7.1.302.29-1 (Tue 05 Feb 2019 10:50:40 AM PST)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> WARNING: Using existing $srcdir/ tree
==> Starting build()...
  -> Adding icu missing folders
  -> Building, this will take a while...
ninja: Entering directory `out.gn/Release'
[1/1666] ACTION //:js2c(//build/toolchain/linux:x64)
FAILED: gen/libraries.cc gen/libraries.bin
python ../../tools/js2c.py gen/libraries.cc CORE ../../src/js/macros.py ../../src/messages.h ../../src/js/prologue.js ../../src/js/array.js ../../src/js/typedarray.js ../../src/js/intl.js --startup_blob gen/libraries.bin
Traceback (most recent call last):
  File "../../tools/js2c.py", line 592, in <module>
    main()
  File "../../tools/js2c.py", line 588, in main
    options.js)
  File "../../tools/js2c.py", line 546, in JS2C
    prepared_sources = PrepareSources(sources, native_type, emit_js)
  File "../../tools/js2c.py", line 395, in PrepareSources
    assert len(macro_files) in [0, 1]
TypeError: object of type 'filter' has no len()
[2/1666] ACTION //:js2c_experimental_extras(//build/toolchain/linux:x64)
FAILED: gen/experimental-extras-libraries.cc gen/libraries_experimental_extras.bin
python ../../tools/js2c.py gen/experimental-extras-libraries.cc EXPERIMENTAL_EXTRAS ../../test/cctest/test-experimental-extra.js --startup_blob gen/libraries_experimental_extras.bin
Traceback (most recent call last):
  File "../../tools/js2c.py", line 592, in <module>
    main()
  File "../../tools/js2c.py", line 588, in main
    options.js)
  File "../../tools/js2c.py", line 546, in JS2C
    prepared_sources = PrepareSources(sources, native_type, emit_js)
  File "../../tools/js2c.py", line 395, in PrepareSources
    assert len(macro_files) in [0, 1]
TypeError: object of type 'filter' has no len()
[3/1666] ACTION //:d8_js2c(//build/toolchain/linux:x64)
FAILED: gen/d8-js.cc
python ../../tools/js2c.py gen/d8-js.cc D8 ../../src/d8.js ../../src/js/macros.py
Traceback (most recent call last):
  File "../../tools/js2c.py", line 592, in <module>
    main()
  File "../../tools/js2c.py", line 588, in main
    options.js)
  File "../../tools/js2c.py", line 546, in JS2C
    prepared_sources = PrepareSources(sources, native_type, emit_js)
  File "../../tools/js2c.py", line 395, in PrepareSources
    assert len(macro_files) in [0, 1]
TypeError: object of type 'filter' has no len()
[4/1666] ACTION //:js2c_extras(//build/toolchain/linux:x64)
FAILED: gen/extras-libraries.cc gen/libraries_extras.bin
python ../../tools/js2c.py gen/extras-libraries.cc EXTRAS ../../test/cctest/test-extra.js --startup_blob gen/libraries_extras.bin
Traceback (most recent call last):
  File "../../tools/js2c.py", line 592, in <module>
    main()
  File "../../tools/js2c.py", line 588, in main
    options.js)
  File "../../tools/js2c.py", line 546, in JS2C
    prepared_sources = PrepareSources(sources, native_type, emit_js)
  File "../../tools/js2c.py", line 395, in PrepareSources
    assert len(macro_files) in [0, 1]
TypeError: object of type 'filter' has no len()
[10/1666] CXX obj/buildtools/third_party/libc++/libc++/locale.o
ninja: build stopped: subcommand failed.
==> ERROR: A failure occurred in build().
    Aborting...
Error making: v8

Tazmain commented on 2018-10-04 13:54 (UTC)

Build fails for me, {standard input}:226988: Error: unknown .loc sub-directive view' {standard input}:226988: Error: junk at end of line, first unrecognized character is-' {standard input}:226990: Error: unknown .loc sub-directive view' {standard input}:226990: Error: unknown pseudo-op:.lvu50714' ninja: build stopped: subcommand failed. ==> ERROR: A failure occurred in build(). Aborting...

dflemstr commented on 2018-09-17 14:29 (UTC)

@marcs did you have time to have a look at this? I haven't seen the new patches being applied yet.

marcs commented on 2018-09-08 17:40 (UTC)

@dflemstr

That's good news, I'll update the package next week.

Thanks

dflemstr commented on 2018-08-31 16:01 (UTC)

Correction: the new version of v8 I proposed does not have the libicu issue.

dflemstr commented on 2018-08-31 15:47 (UTC)

I'd like to see these patches applied: https://gist.github.com/dflemstr/2c06e81da38420d29e702c5bd3601afc

Should fix the build for many users, but the libicu issue might still exist.

thanx commented on 2018-08-30 13:21 (UTC)

Hi Thank you for the package, I am getting following error:

/usr/lib/v8/d8: symbol lookup error: /usr/lib/libv8.so: undefined symbol: _ZN6icu_618ByteSink5FlushEv

It seems like the v8 code has its own icu libraries used during compiling but these are not installed resulting in missing symbols.

Thank you.

ram commented on 2018-07-09 16:35 (UTC)

I'd like to avoid compiling and get this as a pre-compiled package.

thx for the work, guys ;-)

marcs commented on 2018-06-27 16:51 (UTC)

@oneup Thanks for the info, I hope people will find it useful. This package is more of general layout than all-in-one solution due to the complexity of V8.

oneup commented on 2018-06-17 00:02 (UTC)

I'm having a hard time linking against libv8.so. It seems to be due to the fact the libc++.so is located at /usr/lib/v8/libc++.so. I assume that's to not conflict with an existing libc++.so? Unfortunately that makes it difficult both to link against and to dynamically load. Even if you set up pkg-config so that the linker is able to find /usr/lib/v8/libc++.so, you'll still be unable to run the executable because the loader won't know where to find it.

I think adding an RPATH of ${ORIGIN}/v8 to libv8.so would fix the problem, but good luck figuring out how to jam it in to the v8 build process.

For other people running into this issue, you can fix it for your own project by adding -Wl,-rpath,/usr/lib/v8 to your link command. But ideally the libv8.so library itself should know where to find libc++.so without other projects telling the linker/loader where to find it.

marcs commented on 2018-05-28 14:20 (UTC)

@MallardDuck Thanks for the tip, I used your idea. It should build fine now, also package updated to the latest 6.8 version.

MallardDuck commented on 2018-05-17 02:22 (UTC)

Hey,

@ryanvade, @marcs, @znmeb, and anyone else with the building issues, I have a solution!

The root issue is that these scripts expect and depend on Python2 to build. Hence looking for the Python2 style urllib2. So in order to get this building right I simply did a local override; my local user's path has a ~/.bin and then I linked Python2 into that folder.

Once I did that things started working correctly for compiling. I'm gonna unlink it once the build is over, but this trick should work for others too. It's sorta documented in the Wiki: https://wiki.archlinux.org/index.php/python#Python_2 Start from: "Finally, you may not be able to control the script calls"

Hope that helps.

znmeb commented on 2018-04-07 08:11 (UTC)

Build is crashing:

__ running 'download_from_google_storage --no_resume --platform=linux* --no_auth --bucket chromium-clang-format -s v8/buildtools/linux64/clang-format.sha1' in '/d/TMPDIR/yaourt-tmp-znmeb/aur-v8/src' Failed to fetch file gs://chromium-clang-format/5349d1954e17f6ccafb6e6663b0f13cdb2bb33c8 for v8/buildtools/linux64/clang-format. [Err: Traceback (most recent call last): File "/d/TMPDIR/yaourt-tmp-znmeb/aur-v8/src/depot_tools/gsutil.py", line 20, in <module> import urllib2 ModuleNotFoundError: No module named 'urllib2'</module>

marcs commented on 2018-01-08 15:22 (UTC)

@ryanvade I noticed, the problem seems to be with gclient, before the building process, could be related on some ongoing changes on the v8 repository.

ryanvade commented on 2018-01-03 04:27 (UTC)

Its currently failing to build: Failed to fetch file gs://chromium-clang-format/5349d1954e17f6ccafb6e6663b0f13cdb2bb33c8 for v8/buildtools/linux64/clang-format. [Err: [E2018-01-02T22:20:08.898896-06:00 1169 0 annotate.go:242] original error: package "infra/python/wheels/cffi/linux-amd64_cp32_abi3" is not registered

rackelhahn69 commented on 2017-12-13 15:23 (UTC) (edited on 2017-12-13 15:50 (UTC) by rackelhahn69)

==> Building and installing package
==> Making package: v8 6.4.388.6-1 (Wed 13 Dec 16:17:42 CET 2017)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> Retrieving sources...
  -> Cloning depot_tools git repo...
Cloning into bare repository '/tmp/yaourt-tmp-blablabla/aur-v8/depot_tools'...
remote: Sending approximately 29.67 MiB ...
remote: Counting objects: 25739, done
remote: Finding sources: 100% (25739/25739)
remote: Total 50586 (delta 20798), reused 50586 (delta 20798)
Receiving objects: 100% (50586/50586), 44.23 MiB | 3.47 MiB/s, done.
Resolving deltas: 100% (20798/20798), done.
  -> Found v8.pc
  -> Found v8_libbase.pc
  -> Found v8_libplatform.pc
  -> Found d8
  -> Found ctest.patch
==> Validating source files with sha256sums...
    depot_tools ... Skipped
    v8.pc ... Passed
    v8_libbase.pc ... Passed
    v8_libplatform.pc ... Passed
    d8 ... Passed
    ctest.patch ... Passed
==> Extracting sources...
  -> Creating working copy of depot_tools git repo...
Cloning into 'depot_tools'...
done.
==> Starting prepare()...
  -> Creating Python Virtual Environment
Traceback (most recent call last):
  File "/usr/bin/virtualenv2", line 6, in <module>
    from pkg_resources import load_entry_point
ImportError: No module named pkg_resources

EDIT: solved! i needed to install python-pkgconfig and/or python-pip manually

SrTobi commented on 2017-12-02 22:10 (UTC)

It still does not install the headers for libplatform! For that add: install -d ${pkgdir}/usr/include/libplatform install -Dm644 include/libplatform/*.h ${pkgdir}/usr/include/libplatform

marcs commented on 2017-07-22 13:21 (UTC)

@chendaniely > would it make sense to have this depend/work on the gcc6 aur package? I'd rather not, this package is already on the heavy side, and gcc is time/resource consuming to compile from aur. Keep a downgrade of gcc if possible, get gcc6.3.1 from here: https://archive.archlinux.org/ > sjewo has made a PKGBUILD patch to work with gcc6 that might help. > It's based on v8-314, but could be a reference. The problem is a compiling error with gcc7 not a segfault with gcc6 so it is unrelated.

chendaniely commented on 2017-07-20 22:16 (UTC) (edited on 2017-07-21 14:40 (UTC) by chendaniely)

would it make sense to have this depend/work on the gcc6 aur package? edit: sjewo has made a PKGBUILD patch to work with gcc6 that might help. It's based on v8-314, but could be a reference. https://github.com/jeroen/V8/issues/32

marcs commented on 2017-06-21 11:19 (UTC) (edited on 2020-07-25 15:41 (UTC) by marcs)

Updates (Oct-2019):

  • Since 7.7.299.10-1 v8 uses standard lib paths as requested (apart what is needed to run d8).

  • For specific uses like R, I strongly recommend the package v8-r which has a custom configuration for R.

Some notes:

  • Latest version is based on the version on branch-heads from the V8 repositories, I stick to stable, but sometimes I could use beta but I never use dev (check this: https://omahaproxy.appspot.com/).

  • This is a "generic" configuration, v8 is highly configurable, if you need v8 for some specific integration please check the Google documentation.

  • Prefer yay/aurman instead of yaourt/pacaur.

  • Gclient has to download the whole V8 repo, it takes some time to complete (depending on your internet connection).

  • This package requires around 3/4 gigs of space so be sure to have a large enough /tmp partition if you use yaourt, also be aware that aurman/yay/pacaur will save this package in ~/.cache/(aurman|yay|pacaur)/v8, so be sure to have a /home partition with enough space.

  • The build and testing process takes some time, I recommend to use at least a 4 cores with 8 threads cpu, better an 8 cores 16 threads cpu.

  • If you have at least 32GB of RAM you can use a ram partition to speed up the build process (yaourt uses /tmp by default).

marcs commented on 2017-06-21 11:05 (UTC) (edited on 2017-06-25 10:46 (UTC) by marcs)

@OdinEidolon gclient has to download the whole repository takes quite some time depending on your internet connection. @chendaniely I really don't like pacaur, it doesn't display msg2 lines (in prepare) that in this package are important because some steps take long time to complete. You get that error because of pacaur, It's an easy fix but as I checked now the build is broken because of gcc7.

chendaniely commented on 2017-06-14 14:30 (UTC) (edited on 2017-06-14 14:35 (UTC) by chendaniely)

@OdinEidolon for me it seems like it (`Running: gclient sync --with_branch_heads `) hangs, but it just takes a (long) while for it go get past that step

chendaniely commented on 2017-06-14 14:16 (UTC)

I'm getting a syntax error in the python script: $ pacaur -S v8 :: Package v8 not found in repositories, trying AUR... :: resolving dependencies... :: looking for inter-conflicts... AUR Packages (1) v8-5.9.211.27-1 :: Proceed with installation? [Y/n] :: Retrieving package(s)... :: v8 build files are up-to-date -- skipping [sudo] password for dchen: :: Checking v8 integrity... ==> Making package: v8 5.9.211.27-1 (Wed Jun 14 10:10:19 EDT 2017) ==> Retrieving sources... -> Updating depot_tools git repo... Fetching origin -> Found v8.pc -> Found v8_libbase.pc -> Found v8_libplatform.pc -> Found d8 -> Found icu59.diff ==> Validating source files with sha256sums... depot_tools ... Skipped v8.pc ... Passed v8_libbase.pc ... Passed v8_libplatform.pc ... Passed d8 ... Passed icu59.diff ... Passed :: Building v8 package(s)... ==> Making package: v8 5.9.211.27-1 (Wed Jun 14 10:10:27 EDT 2017) ==> Checking runtime dependencies... ==> Checking buildtime dependencies... ==> WARNING: Using existing $srcdir/ tree ==> Removing existing $pkgdir/ directory... ==> Starting build()... -> Running GN... Traceback (most recent call last): File "../depot_tools/gn.py", line 15, in <module> import gclient_utils File "/home/dchen/.cache/pacaur/v8/src/depot_tools/gclient_utils.py", line 266 except OSError, e: ^ SyntaxError: invalid syntax ==> ERROR: A failure occurred in build(). Aborting... :: failed to build v8 package(s)

OdinEidolon commented on 2017-06-13 08:23 (UTC)

Installation seems to get stuck at: Running: gclient sync --with_branch_heads

marcs commented on 2017-05-19 08:03 (UTC) (edited on 2017-06-21 10:02 (UTC) by marcs)

@kageurufu v8-3.14 has conflicts with v8 v8-3.14-bin can be installed side-by-side, went overboard with package conflicts.

kageurufu commented on 2017-05-17 17:47 (UTC)

Why does this conflict with v8-3.14/-bin? v8-3.14-bin installs /usr/lib/libv8.so.3.14.5 specifically so it can be installed side-by-side with 5.x+

marcs commented on 2017-05-10 11:23 (UTC) (edited on 2017-05-13 13:54 (UTC) by marcs)

Testing the 5.9 update: * introduced minimal patch for icu 59 * one assert failing related to internationalization: https://groups.google.com/forum/#!topic/v8-users/tkN1B9RvnLc

marcs commented on 2016-11-01 12:02 (UTC)

Applied dflemstr's patch and added wget as makedep. Thanks for the patch, feedback is welcome.

dflemstr commented on 2016-10-31 07:00 (UTC) (edited on 2016-10-31 07:00 (UTC) by dflemstr)

I have a bunch of patches to this library that do the following: - Use system ICU instead of bundled ICU - Enable more optimizations - Fix pkg-config files to actually be usable Full changes here: https://gist.github.com/dflemstr/01137016843183420a1c55395e6d1596

Cambi0n commented on 2016-10-14 19:07 (UTC)

wget is a dependency. It is needed by gclient sync in depot_tools.

void.dargon commented on 2016-09-11 21:01 (UTC) (edited on 2016-09-14 05:16 (UTC) by void.dargon)

is it just me or is the not being build libv8_platform.so? it seems to be missing.

marcs commented on 2016-09-09 08:40 (UTC)

Latest cherry-pick. Fixed missing libv8.so, Switched to host clang.

marcs commented on 2016-09-06 13:50 (UTC) (edited on 2016-09-06 13:51 (UTC) by marcs)

@dflemstr Yeah, the docs for GN and V8 are still lacking a bit. Now I'm using v8-5.2 but I'll have to switch to the latest V8 version very, very soon. So I'll have to work my way through GN. Also testing a build is kinda time consuming I would like to have a 36 cores Xeon beast ;)

dflemstr commented on 2016-09-04 16:15 (UTC)

Hmm... When I build this, it doesn't actually produce a libv8.so file any more, so this line fails: install -Dm755 out.gn/x64.release/libv8.so $pkgdir/usr/lib/v8/libv8.so Also, it would be nice to get static libraries being built again but I understand that that's tricky when using GN?

marcs commented on 2016-08-24 11:47 (UTC)

@MagiCarbon try with v8-5.2 has all the libs now.

MagiCarbon commented on 2016-08-24 04:06 (UTC)

sorry, i was wrong. i've asked the author of php-v8js, he said it does support the latest v8. sorry for all confusion. and the error below occurred because this package lost 'libv8.so', not the v8 version. configure: error: could not determine libv8 version

marcs commented on 2016-08-23 20:36 (UTC) (edited on 2016-08-23 20:38 (UTC) by marcs)

@r0qs Gyp is deprecated now for building v8, I switched to GN. So I guess compilation should work next time you try. @MagiCarbon (and others) Because seems that some people need v8 5.2 for phpjs (But Why ? http://i.imgur.com/TnQRX6v.gif ;) ) I've added a new package called v8-5.2 which still uses Gyp and it should generate all the libs necessary for phpjs.

r0qs commented on 2016-08-23 15:00 (UTC)

Hi, I'm getting this error during compile: ==> Starting build()... -> Running gyp... Updating projects from gyp files... -> Start building... ninja: Entering directory `out/Release' ninja: error: '../../third_party/WebKit/Source/platform/inspector_protocol/DispatcherBase.cpp', needed by 'obj/third_party/WebKit/Source/platform/inspector_protocol/inspector_protocol.DispatcherBase.o', missing and no known rule to make it ==> ERROR: A failure occurred in build(). Aborting... Anyone knows how to solve?

marcs commented on 2016-08-11 06:41 (UTC) (edited on 2016-08-11 06:41 (UTC) by marcs)

You can simply clone the package and set pkgver=5.2.361.49. If that version works with php-v8js you should upload a package with the name: v8-5.2.361.49

MagiCarbon commented on 2016-08-11 00:56 (UTC)

thanks for your working. LOL, php-v8js couldn't recognize this version. <quote> configure: error: could not determine libv8 version </quote> could you plz update the package to the latest chrome-pick(5.2.361.49)? it maybe a better choice.

marcs commented on 2016-08-10 12:37 (UTC) (edited on 2016-08-10 12:40 (UTC) by marcs)

In the new version the static library files (like libv8_libplatform.a) are installed in this path: /usr/share/v8/ Let me know if it is necessary to move them to /usr/lib for a package dependency or something like that.

marcs commented on 2016-08-10 09:17 (UTC)

I'll update the package to the latest cherry-pick (5.4.374.1), but first I need to solve some problems with the libraries.

MagiCarbon commented on 2016-08-10 02:04 (UTC) (edited on 2016-08-10 02:07 (UTC) by MagiCarbon)

@stonecold111 in https://github.com/phpv8/v8js/blob/master/README.Linux.md it says: <quote> # Install libv8_libplatform.a (V8 >= 5.2.51) echo -e "create /usr/lib/libv8_libplatform.a\naddlib > > out/native/obj.target/src/libv8_libplatform.a\nsave\nend" | sudo ar -M </quote> therefore, this package should be upgraded to the last release.

stonecold111 commented on 2016-07-06 14:36 (UTC)

I'm unable to install V8JS with this package. I'm missing a file called libv8_libplatform.a, which isn't in /usr/lib/v8/ to begin with: sudo pecl install v8js ...... checking for V8 version... 5.1.281.12 checking for libv8_libplatform.a... not found configure: error: Please provide libv8_libplatform.a next to the libv8.so, see README.md for details ERROR: `/tmp/pear/temp/v8js/configure --with-php-config=/usr/bin/php-config --with-v8js' failed Files in /usr/lib/v8: /usr/lib/v8/d8 /usr/lib/v8/libv8.so /usr/lib/v8/natives_blob.bin /usr/lib/v8/snapshot_blob.bin

marcs commented on 2016-06-29 14:47 (UTC)

Should build fine now: updated libtinfo5 to ncurses5-compat-libs.

pintergreg commented on 2016-06-21 18:19 (UTC)

There is no libtinfo5 anymore. libtinfo is at version 6, which does not provide /usr/lib/libtinfo.so.5, but /usr/lib/libtinfo.so.6 With libtinfo and a manually created libtinfo.so.5 symlink (and some PKGBUILD tweak) the package complies.

marcs commented on 2016-06-10 11:48 (UTC)

Updated the dep to libtinfo5

alex.henrie commented on 2016-05-30 23:18 (UTC)

I can't install this package anymore because vim-youcompleteme-git requires libtinfo5 and v8 requires libtinfo-5, but only one of the two libtinfo's can be installed at the same time.

hav3lock commented on 2016-05-24 22:31 (UTC)

libtinfo-5 will be now called libtinfo5, please update your package accordingly.

marcs commented on 2016-05-19 15:55 (UTC)

It is a connection problem, probably related to your connection or the repository was momentarily unavailable.

arturodz commented on 2016-05-16 01:43 (UTC)

When attempting to install I get this error: svn checkout --force http://gyp.googlecode.com/svn/trunk build/gyp \ --revision 1501 svn: E170013: Unable to connect to a repository at URL 'http://gyp.googlecode.com/svn/trunk' svn: E160013: '/svn/trunk' path not found Makefile:289: recipe for target 'dependencies' failed make: *** [dependencies] Error 1

marcs commented on 2016-04-23 13:21 (UTC)

Updated to latest cherry-pick

marcs commented on 2016-04-21 22:45 (UTC) (edited on 2016-04-23 13:21 (UTC) by marcs)

Ok, I saw the problem. Which package for ncurses are you referring to ? [EDIT] Found the package you were talking about, now the package builds fine thanks. In case of a missing gpg key error for installing ncurses5-compat-libs, import the key using this command: gpg --keyserver pgp.mit.edu --recv-keys F7E48EDB

WoefulDerelict commented on 2016-04-18 22:19 (UTC)

When attempting to build this package it complained that libtinfo.so.5 was missing. The version it pulled as a dependency was libtinfo which provides libtinfo.so.6 There is a package in the AUR to provide the symlink to a legacy ncurses or it is reported that symlinking to the new ncurses works fine.

marcs commented on 2016-03-09 11:11 (UTC)

Package builds fine now.

sleeping commented on 2016-03-08 15:36 (UTC)

marcs, Thanks for your response. That's too bad! Do you know how to solve it?

marcs commented on 2016-03-08 14:26 (UTC)

I know, my automated builds are failing for quite some time now. Probably related to glibc: pacman -Qo /usr/lib64/crti.o /usr/lib/crti.o is owned by glibc 2.23-1

sleeping commented on 2016-03-08 12:48 (UTC)

This might be broken now. https://bbs.archlinux.org/viewtopic.php?pid=1610654#p1610654

dietchrst commented on 2015-12-09 16:45 (UTC)

omangold, I had this same issue - but I was missing clang-tool-extra, like marcs said. I installed it, but when I tried to build again, it failed again. Came back after restarting the next chance I got and everything worked. Maybe a restart would help? I cant say for sure, but it was the only thing that I can think of that changed for me. Hope it helps!

omgold commented on 2015-12-09 11:46 (UTC)

Sorry about the long text. Checked, but clang-tools-extra was already installed for the try below, as it is in makedepends.

marcs commented on 2015-12-08 14:44 (UTC)

Hi omangold, Looks like you are missing the package: clang-tools-extra, please check. And please use pastebin for such a long build text, thanks.

omgold commented on 2015-12-08 11:22 (UTC) (edited on 2015-12-09 11:57 (UTC) by omgold)

I get the build error below. Any ideas? http://pastebin.com/5i4dDAjM

eugeneai commented on 2015-12-05 07:51 (UTC)

clang-tools-extra must be added to build dependencies.

marcs commented on 2015-12-02 01:26 (UTC) (edited on 2015-12-02 01:46 (UTC) by marcs)

I lost so much time in the building process that I didn't tested the package at all. Fixed now. Also added $* to the sh script so it can send arguments to the d8 executable.

hzy199411 commented on 2015-12-02 00:42 (UTC)

the last v8 need natives_blob.bin, or it will get "Segmentation fault (core dumped)" a fix is here: http://pastebin.com/EUjr5KhK