This commit adds support to run the linter without running any other
targets. This commit also makes the lint task a little more quite and
more inline with the output on other operating systems.
Below is an example of running (with a lint error to show that it is not
so quite that errors are hidden):
C:\Users\danbev\working\node>vcbuild.bat lint
running lint-cpp 'src\*.c src\*.cc src\*.h test\addons\*.cc
test\addons\*.h test\cctest\*.cc test\cctest\*.h test\gc\binding.cc
tools\icu\*.cc tools\icu\*.h'
src\env.h:24: Should have a space between // and comment
[whitespace/comments] [4]
Total errors found: 1
"C:\Python27\python.exe" tools/check-imports.py
running lint-js
The help message now looks like:
C:\Users\danbev\working\node>vcbuild.bat /?
vcbuild.bat [debug/release] [msi] [test/test-ci/test-all/test-uv/
test-inspector/test-internet/test-pummel/test-simple/test-message/
test-async-hooks/test-v8/test-v8-intl/test-v8-benchmarks/
test-v8-all] [clean] [noprojgen] [small-icu/full-icu/without-intl]
[nobuild] [sign] [x86/x64] [vs2015/vs2017] [download-all]
[enable-vtune] [lint/lint-ci] [no-NODE-OPTIONS]
[link-module path-to-module]
Examples:
vcbuild.bat : builds release build
vcbuild.bat debug : builds debug build
vcbuild.bat release msi : builds release build and MSI installer package
vcbuild.bat test : builds debug build and runs tests
vcbuild.bat build-release : builds the release distribution as used by nodejs.org
vcbuild.bat enable-vtune : builds nodejs with Intel VTune profiling support to profile JavaScript
vcbuild.bat link-module my_module.js : bundles my_module as built-in module
vcbuild.bat lint : runs the C++ and JavaScript linter
PR-URL: https://github.com/nodejs/node/pull/16176
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Gireesh Punathil <gpunathi@in.ibm.com>
Currently the help message when using --with-lttng looks like this:
$ ./configure --help | grep 'with-lttng'
--with-lttng build with Lttng (Only available to Linux)
This commit makes the help message consistent with the error message
that is raised if --with-lttng is used on a non-Linux operating
system:
$ ./configure --help | grep 'with-lttng'
--with-lttng build with Lttng (Only available on Linux)
PR-URL: https://github.com/nodejs/node/pull/16101
Reviewed-By: Lance Ball <lball@redhat.com>
Reviewed-By: Gibson Fahnestock <gibfahn@gmail.com>
Currently the variables set in vcbuild.bat are mostly global and
escape/leak out into the calling process. For example, running
vcbuild.bat test and then echoing the config variable gives:
vcbuild.bat test
...
echo %config%
Release
After this change the same command give:
vcbuild.bat test
...
echo %config%
%config%
PR-URL: https://github.com/nodejs/node/pull/15754
Reviewed-By: James M Snell <jasnell@gmail.com>
The headers node_trace_writer.h and node_trace_buffer.h are not used in
agent.h but are more of an implementation detail of agent.cc.
This commit suggests moving the inclusion of these headers to agent.cc.
PR-URL: https://github.com/nodejs/node/pull/15598
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Currently if the buffer size exceeds the maximum int size the following
error will be thrown:
dgram.js:180
throw new errors.Error('ERR_SOCKET_BUFFER_SIZE', e);
^
Error [ERR_SOCKET_BUFFER_SIZE]: Could not get or set buffer size: Error:
Unknown system error 22: Unknown system error 22, uv_recv_buffer_size
at bufferSize (dgram.js:180:11)
It looks like perhaps UV_EINVAL was intended to give the following error
message instead:
dgram.js:180
throw new errors.Error('ERR_SOCKET_BUFFER_SIZE', e);
^
Error [ERR_SOCKET_BUFFER_SIZE]: Could not get or set buffer size: Error:
EINVAL: invalid argument, uv_recv_buffer_size
at bufferSize (dgram.js:180:11)
This commit changes EINVAL to use UV_EINVAL.
PR-URL: https://github.com/nodejs/node/pull/15444
Reviewed-By: Tobias Nießen <tniessen@tnie.de>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Currently the following compiler warning is generated:
1 warning generated.
../src/udp_wrap.cc:238:12: warning: comparison of integers of different
signs: 'int' and 'uint32_t' (aka 'unsigned int') [-Wsign-compare]
if (size != args[0].As<Uint32>()->Value()) {
~~~~ ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 warning generated.
This commit changes the check to see that the Uint32 value does not
exceed the max int size instead of first casting and then comparing.
PR-URL: https://github.com/nodejs/node/pull/15402
Reviewed-By: Franziska Hinkelmann <franziska.hinkelmann@gmail.com>
The following deprecation warning is displayed when compiling:
../src/module_wrap.cc:187:18: warning: 'Instantiate' is deprecated
[-Wdeprecated-declarations]
bool ok = mod->Instantiate(ctx, ModuleWrap::ResolveCallback);
^
../deps/v8/include/v8.h:1158:22: note: 'Instantiate' has been explicitly
marked deprecated here
bool Instantiate(Local<Context> context,
^
This commit changes this function call to use InstantiateModule instead
which returns a Maybe<bool>.
PR-URL: https://github.com/nodejs/node/pull/15423
Reviewed-By: Michaël Zasso <targos@protonmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Franziska Hinkelmann <franziska.hinkelmann@gmail.com>
Currently when building using --debug-http2 there are a number of
compilation errors like the following:
In file included from ../src/node_http2.cc:3:
In file included from ../src/node_http2.h:6:
../src/node_http2_core-inl.h:43:32: error: too few arguments to function
call, single argument 'type'
was not specified
handle->TypeName(), len, message);
This commit adds the type argument to the calls.
PR-URL: https://github.com/nodejs/node/pull/15197
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Currently the options available for HTTP2 are not displayed when running
configure --help.
This commit enables the following HTTP2 section to be included in the
help output:
HTTP2:
Flags that allows you to control HTTP2 features in Node.js
--debug-http2 build with http2 debug statements on (default is
false)
--debug-nghttp2 build nghttp2 with DEBUGBUILD (default is false)
PR-URL: https://github.com/nodejs/node/pull/15198
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Evan Lucas <evanlucas@me.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Currently, there are a few compiler warnings generated from node_perf.cc
regarding unused return values:
../src/node_perf.cc:58:3: warning: ignoring return value of function
declared with warn_unused_result attribute [-Wunused-result]
env->performance_entry_callback()->Call(context,
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~
../src/node_perf.cc:293:5: warning: ignoring return value of function
declared with warn_unused_result attribute [-Wunused-result]
obj->Set(context, idx, args[idx]);
^~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~
../src/node_perf.cc:373:3: warning: ignoring return value of function
declared with warn_unused_result attribute [-Wunused-result]
target->DefineOwnProperty(context,
^~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~
../src/node_perf.cc:378:3: warning: ignoring return value of function
declared with warn_unused_result attribute [-Wunused-result]
target->DefineOwnProperty(context,
^~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~
This commit add checks/casts to avoid the warnings.
PR-URL: https://github.com/nodejs/node/pull/15112
Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com>
Reviewed-By: Tobias Nießen <tniessen@tnie.de>
Reviewed-By: Michaël Zasso <targos@protonmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: James M Snell <jasnell@gmail.com>
The motivation for this commit is to pick up early on missing checks for
crypto support (when Node is built --without-ssl).
There are currently usages of common.hasCrypto which are not just for
detecting if crypto support is available and then skip the test in
question. For these case we still want to have a lint error generated
which can then be disabled using an ESLint comment.
PR-URL: https://github.com/nodejs/node/pull/13813
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Teddy Katz <teddy.katz@gmail.com>
Currently when building with --enable-static and running the test target
the following error will be reported:
Building addon
/node/test/addons/01_function_arguments/
env: ./node: No such file or directory
make[1]: *** [test/addons/.buildstamp] Error 1
Note that this is with a clean build where no prior node executable was
built.
This commit suggests only running the cctest target when --enable-static
is specified.
PR-URL: https://github.com/nodejs/node/pull/14892
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: Lance Ball <lball@redhat.com>
Currently the files specified in libraries in node.gyp `cctest` target are
getting a '.lib' extension on windows when generated with ninja.
This commit adds a check to see if a file has a '.obj' extension and in
that case no '.lib' extension will be added.
Also, the LIBS specified in the 'libraries' section are not
being included in the --start-group --end-group section which
means that these libraries will not be searched causing issue
with linkers where the order matters.
PR-URL: https://github.com/nodejs/node/pull/12484
Fixes: https://github.com/nodejs/node/issues/12448
Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: Gibson Fahnestock <gibfahn@gmail.com>
Currently when building with --enabled-static the cctest target will
include libraries to be linked regardless. This commit adds a condition
to only add the libraries when dynamically linking.
This commit re-adds the previously reverted commit (be63c26e8c
"build: don't add libraries when --enable-static"). The reason that
commit failed was a mistake on my part where I moved the the list
into a variable (thinking that it was a simple change that would not
affect anything) and not re-running CI. But the list itself contains
variables used to enable differences in operating systems/compilers
where the object file locations vary.
PR-URL: https://github.com/nodejs/node/pull/14912
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Refael Ackermann <refack@gmail.com>
This commit removes the following comment and attempts to fix the issue
as the the requirement of g++ version is now 4.9.4 or newer:
// XXX(bnoordhuis) Work around a g++ 4.9.2 template type inferrer bug
// when the expression is written as info.Data().As<v8::External>().
PR-URL: https://github.com/nodejs/node/pull/14732
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Tobias Nießen <tniessen@tnie.de>
While stepping though SetupPromises I noticed that the environments
Isolate is used but not when creating the string "_setupPromises".
Is there a reason for using args.GetIsolate() instead of using the
environments isolate? I see that GetIsolate() is an inline call, but
could there be situations where it returns a different Isolate?
If not perhaps using the local isolate variable would be a litte
clearer.
PR-URL: https://github.com/nodejs/node/pull/14768
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Timothy Gu <timothygu99@gmail.com>
Reviewed-By: Tobias Nießen <tniessen@tnie.de>
Reviewed-By: Alexey Orlenko <eaglexrlnk@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
test-tls-session-cache currently fails if built --without-ssl:
internal/util.js:82
throw new errors.Error('ERR_NO_CRYPTO');
^
Error [ERR_NO_CRYPTO]: Node.js is not compiled with OpenSSL crypto
support
at Object.assertCrypto (internal/util.js:82:11)
at tls.js:26:14
at NativeModule.compile (bootstrap_node.js:586:7)
at Function.NativeModule.require (bootstrap_node.js:531:18)
at Function.Module._load (module.js:449:25)
at Module.require (module.js:517:17)
at require (internal/module.js:11:18)
at Object.<anonymous>
(/node/test/parallel/test-tls-session-cache.js:26:13)
at Module._compile (module.js:573:30)
at Object.Module._extensions..js (module.js:584:10)
The test has a crypto check but it come after the require of the tls
module.
This commit moves the crypto check to come before the require of tls and
allows the test to pass.
PR-URL: https://github.com/nodejs/node/pull/14708
Reviewed-By: Rich Trott <rtrott@gmail.com>
Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
Reviewed-By: David Cai <davidcai1993@yahoo.com>
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Tobias Nießen <tniessen@tnie.de>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
The sanity_check AtExit callback needs to come last to verify that the
other callbacks have been completed. This was not noticed before as this
code was not been executed.
PR-URL: https://github.com/nodejs/node/pull/14048
Reviewed-By: Michaël Zasso <targos@protonmail.com>
When working on commit 0d95a0b60a
("test: remove undef NDEBUG from at-exit addons test) I searched for
usages of undef NDEBUG but did not include the doc directory (but I did
include the test directory) and missed this one.
Commit 1f02569f8db9cb0101807df4982534738f0161b2 ("tools: fix
tools/addon-verify.js") enables the code in the "AtExit" section to be
included in the test/addons diretory and this code will again be
tested.
PR-URL: https://github.com/nodejs/node/pull/14048
Reviewed-By: Michaël Zasso <targos@protonmail.com>
The current implementation of addon-verify.js is including the code
for the "Function arguments" section in test/addons/01_callbacks and
there is no directory generated or the "Function arguments section".
This continues and leads to the last section, "AtExit", code to be
excluded. There is an test/addons/07_atexit_hooks but it contains code
from the "Passing wrapped objects around" section.
This commit modifies addon-verify to associate headers with code and
then iterates over the set and generates the files as a separate step.
PR-URL: https://github.com/nodejs/node/pull/14048
Reviewed-By: Michaël Zasso <targos@protonmail.com>
When building --without-ssl and running the tests some of the http2 test
fail with the following error message:
internal/util.js:82
throw new errors.Error('ERR_NO_CRYPTO');
^
Error [ERR_NO_CRYPTO]: Node.js is not compiled with OpenSSL crypto
support
at Object.assertCrypto (internal/util.js:82:11)
at internal/http2/core.js:5:26
at NativeModule.compile (bootstrap_node.js:586:7)
at NativeModule.require (bootstrap_node.js:531:18)
at http2.js:17:5
at NativeModule.compile (bootstrap_node.js:586:7)
at Function.NativeModule.require (bootstrap_node.js:531:18)
at Function.Module._load (module.js:449:25)
at Module.require (module.js:517:17)
at require (internal/module.js:11:18)
This commit adds hasCrypto checks and skips the tests if there is no
crypto support.
PR-URL: https://github.com/nodejs/node/pull/14657
Reviewed-By: Rich Trott <rtrott@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Tobias Nießen <tniessen@tnie.de>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Currently the following warning is issued when buildning:
Building addon
/work/nodejs/node/test/addons-napi/test_general/
CC(target) Debug/obj.target/test_general/test_general.o
../test_general.c:116:14: warning: variable 'result' is used
uninitialized whenever 'if' condition is false
[-Wsometimes-uninitialized]
} else if (argument_type == napi_null) {
^~~~~~~~~~~~~~~~~~~~~~~~~~
../test_general.c:119:10: note: uninitialized use occurs here
return result;
^~~~~~
../test_general.c:116:10: note: remove the 'if' if its condition is
always true
} else if (argument_type == napi_null) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../test_general.c:101:20: note: initialize the variable 'result' to
silence this warning
napi_value result;
^
= NULL
This commit simply initializes result to NULL to avoid this warning.
PR-URL: https://github.com/nodejs/node/pull/14104
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>
Currently this test fail when configured --without-inspector or
--without-ssl as it is expected to fail but the skipIfInspectorDisabled
check will exit as if the test was sucessful.
This commit checks if inspector support is available and fails the test
allowing the test to be skipped.
PR-URL: https://github.com/nodejs/node/pull/14074
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Refael Ackermann <refack@gmail.com>
The at-exit addons test uses asserts like the other addons tests,
but at-exit is the only one that undefines NDEBUG to make sure
that asserts are enabled. This commit removes the undef for
consistency.
PR-URL: https://github.com/nodejs/node/pull/13998
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Currently, node_crypto_bio and node_crypto_clienthello are not in the
crypto namespace but simply in the node namespace. Not sure if this was
intentional or not, but I think it would make sense to move them to be
consistent.
PR-URL: https://github.com/nodejs/node/pull/13957
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: James M Snell <jasnell@gmail.com>
The motivation for this commit is that we are building Node with
--shared-openssl and in our case the system OpenSSL version
supports FIPS.
The tests in test-crypto-fips that toggle fips mode on/off using the
config file option might succeed and return 1 instead of an error
being thrown from OpenSSL (which is what happens for a default build
but the error is not processed/displayed in any way at the moment):
openssl config failed: error:060B10A7:digital envelope
routines:ALG_MODULE_INIT:fips mode not supported
Note that this only concerns the test that use the configuration file
option which is different from when calling the fips setter as
the handling of the configuration file is handled by OpenSSL, so it
is not possible for us to try to call the fips setter as that would
throw an error ("Error: Cannot set FIPS mode in a non-FIPS build.").
The suggestion is to skips these tests when --shared-openssl is used.
PR-URL: https://github.com/nodejs/node/pull/13786
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Currently when configuring the project using --openssl-fips a gyp
include file name config_fips.gypi will be created. If the project is
later configured but without the --openssl-fips flag an error will
occur. For example:
$ ./configure --openssl-fips=bogus
$ ./configure && make -j8
...
/node/deps/openssl/fips/fipsld:
line 8: /bin/fipsld: No such file or directory
Error 127
This commit suggests removing the generate config_fips.gypi when the
--openssl-fips flag is not give on the command line.
PR-URL: https://github.com/nodejs/node/pull/13837
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
We are currenly builing Node with --shared-zlib which happens to be
version 1.2.8. The test for zlib.createDeflateRaw is expected to fail
but does not when using version 1.2.8.
As far as I can tell the fix referred to in the comments was
introduced in version 1.2.9:
- Reject a window size of 256 bytes if not using the zlib wrapper
This commit suggests adding a check for the version and skipping this
assert if the version is less than 1.2.9.
Refs: http://zlib.net/ChangeLog.txt
PR-URL: https://github.com/nodejs/node/pull/13697
Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
Currently, there are a few test-tls-wrap-econnreset test that fail when
Node is configured --without-ssl:
Error: Node.js is not compiled with openssl crypto support
This commit adds crypto checks and skips these tests if no crypto
support unavailable.
PR-URL: https://github.com/nodejs/node/pull/13691
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Gibson Fahnestock <gibfahn@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
There are a number of void casts of clear_error_on_return which is a
usage of the RAII idiom. The ClearErrorOnReturn struct only has a
destructor and no constructor which I believe was an issue in GCC
prior to version 4.8.0, which lead to a unused variable warning.
I'm wondering if these cast could be removed since GCC 4.8.5 or newer
is required now. An alternative solution would be to add an empty
constructor which should work allowing the compiler to detect that a
variable is used only for its side-effects.
Not sure if this was the sole reason for having these casts but wanted
to bring it up just in case.
Refs: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=10416
PR-URL: https://github.com/nodejs/node/pull/13669
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Original commit message:
This commit adds a getter for the private is_verbose_ member.
The use case for this comes from Node.js where the ability to avoid
calling FatalException if the TryCatch is verbose would be nice
to have.
BUG=
Review-Url: https://codereview.chromium.org/2840803002
Cr-Commit-Position: refs/heads/master@{#45018}
PR-URL: https://github.com/nodejs/node/pull/12826
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
When configuring node --without-ssl or --without-inspector these test
will fail. The underlying issue will be:
Inspector support is not available with this Node.js build
/work/nodejs/node/out/Release/node: bad option: --inspect=0
This commit adds checks to see if inspector support is enabled and if
not skips these tests.
PR-URL: https://github.com/nodejs/node/pull/13324
Reviewed-By: Gibson Fahnestock <gibfahn@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Eugene Ostroukhov <eostroukhov@chromium.org>
Reviewed-By: Sam Roberts <vieuxtech@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Currently this test will fail with the following error message when
configured --without-ssl:
Error: Node.js is not compiled with openssl crypto support
This commit checks for crypto and skips this tests if such support
is not available.
PR-URL: https://github.com/nodejs/node/pull/13253
Reviewed-By: Rich Trott <rtrott@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
When upgrading OpenSSL, Step 6 in upgrading guide explains the steps
that need to be taken if asm files need updating. This might not
always be the case and something that needs to be checked from
release to release.
This commit adds an example of using github to manually compare two tags
to see if any changes were made to asm files.
PR-URL: https://github.com/nodejs/node/pull/13234
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
VerifyCallback returns 1 in two locations but CHECK_CERT_REVOKED in a
third return statment. This commit suggests that CHECK_OK is used
instead of 1. CHECK_OK is also used as the return value in
CheckWhitelistedServerCert so it seems to be consitent change to make.
PR-URL: https://github.com/nodejs/node/pull/13241
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Currently the InspectorStarted function is guarded by the else clause of
the NODE_USE_V8_PLATFORM macro. If node is configured --without-ssl then
NODE_USE_V8_PLATFORM will be 1 but the nested HAVE_INSPECTOR macro
will not be 0 which will lead to that there will be no InspectorStarted
function defined.
If building --without-inspector or --without-ssl the following
compilation error will occur:
../src/node.cc:4470:57: error: no member named 'InspectorStarted' in
'node::(anonymous struct at ../src/node.cc:241:8)'
if (debug_options.inspector_enabled() &&
!v8_platform.InspectorStarted(&env))
~~~~~~~~~~~ ^
../src/node.cc:4470:57: error: no member named 'InspectorStarted' in
'node::(anonymous struct at ../src/node.cc:241:8)'
if (debug_options.inspector_enabled() &&
!v8_platform.InspectorStarted(&env))
~~~~~~~~~~~ ^
1 error generated.
This commit adds a separate if preprocessor directive to catch the case
when either --without-ssl/--without-inspector and --without-v8-platform
combinations are used to configure node.
PR-URL: https://github.com/nodejs/node/pull/13167
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Refael Ackermann <refack@gmail.com>
Currently the following compiler warning is displayed:
../src/inspector_agent.cc:218:5: warning: ignoring return value of
function declared with warn_unused_result attribute [-Wunused-result]
callback->Call(env_->context(), receiver, 1, &argument);
^~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 warning generated.
This commit does a static cast of the result as there are tests that
fail if we try to do something like ToLocalChecked.
PR-URL: https://github.com/nodejs/node/pull/13188
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
root_cert_store is defined as extern in node_crypto.h but only used in
node_crypto.cc. It is then set using SSL_CTX_set_cert_store. The only
usages of SSL_CTX_get_cert_store are in node_crypto.cc which would all
be accessing the same X509_STORE through the root_cert_store pointer as
far as I can tell. Am I missing something here?
This commit suggests removing it from the header and making it static
in node_crypto.cc.
PR-URL: https://github.com/nodejs/node/pull/13194
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Sam Roberts <vieuxtech@gmail.com>
If node is configured --without-inspector/--without-ssl this test will
fail with the following error:
Path: inspector/test-bindings
inspector.js:8
throw new Error('Inspector is not available');
^
Error: Inspector is not available
at inspector.js:8:9
at NativeModule.compile (bootstrap_node.js:549:7)
This commit skips this test if the inspector is disabled.
PR-URL: https://github.com/nodejs/node/pull/13186
Reviewed-By: Rich Trott <rtrott@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>
Reviewed-By: Eugene Ostroukhov <eostroukhov@google.com>
Reviewed-By: Richard Lau <riclau@uk.ibm.com>
This commit attempts to fix the following TODO:
// TODO(bnoordhuis) Should perhaps also check whether
getauxval(AT_SECURE) is non-zero on Linux.
This can be manually tested at the moment using the following steps:
$ setcap cap_net_raw+ep out/Release/node
$ NODE_PENDING_DEPRECATION="1" out/Release/node -p
"process.binding('config').pendingDeprecation"
true
$ useradd test
$ su test
$ NODE_PENDING_DEPRECATION="1" out/Release/node -p
"process.binding('config').pendingDeprecation"
undefined
PR-URL: https://github.com/nodejs/node/pull/12548
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
This commit attempts to fix an issue when building on windows using the
following command line options:
.\vcbuild.bat dll debug x64 vc2015
This will result in the following options passed to configure:
configure --debug --shared --dest-cpu=x64 --tag=
This commit excludes the dependency to openssl if node is configured
with --shared.
Also, FP_API to the categories to export in mkssldef when generating
the module definition (openssl.def) allowing the build to compile and
link successfully.
Fixes: https://github.com/nodejs/node/issues/12952
PR-URL: https://github.com/nodejs/node/pull/13078
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Refael Ackermann <refack@gmail.com>
Currently the following compiler warning is displayed when building:
../test/cctest/test_inspector_socket_server.cc:142:8: warning:
'ServerDone' overrides a member function but is not marked 'override'
[-Winconsistent-missing-override]
void ServerDone() {
^
../src/inspector_socket_server.h:30:16: note: overridden virtual
function is here
virtual void ServerDone() = 0;
^
This commit marks ServerDone with override to get rid of the warning.
PR-URL: https://github.com/nodejs/node/pull/13166
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Refael Ackermann <refack@gmail.com>
I came across this template class but I don't understand why it is
there. It is not used in the template specialization following it.
I just wanted to bring it up just in case this is something that
has been overlooked.
PR-URL: https://github.com/nodejs/node/pull/12993
Reviewed-By: Fedor Indutny <fedor@indutny.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Currently it is possible to call TLSWrap() without using new and the
following error message:
FATAL ERROR: v8::Object::SetAlignedPointerInInternalField() Internal
field out of bounds
1: node::Abort()
[/Users/danielbevenius/work/nodejs/node/out/Debug/node]
2: node::OnFatalError(char const*, char const*)
[/Users/danielbevenius/work/nodejs/node/out/Debug/node]
3: v8::Utils::ReportApiFailure(char const*, char const*)
[/Users/danielbevenius/work/nodejs/node/out/Debug/node]
4: v8::Utils::ApiCheck(bool, char const*, char const*)
[/Users/danielbevenius/work/nodejs/node/out/Debug/node]
5: v8::InternalFieldOK(v8::internal::Handle<v8::internal::JSReceiver>,
int, char const*)
[/Users/danielbevenius/work/nodejs/node/out/Debug/node]
6: v8::Object::SetAlignedPointerInInternalField(int, void*)
[/Users/danielbevenius/work/nodejs/node/out/Debug/node]
7: node::TLSWrap::Initialize(v8::Local<v8::Object>,
v8::Local<v8::Value>,
v8::Local<v8::Context>)::$_0::operator()(v8::FunctionCallbackInfo<v8::Value>
const&) const [/Users/danielbevenius/work/nodejs/node/out/Debug/node]
This commit adds a IsConstructCall check which will produce the
following error message:
/Users/danielbevenius/work/nodejs/node/out/Debug/node[2212]:
../src/tls_wrap.cc:936:auto node::TLSWrap::Initialize(Local<v8::Object>,
Local<v8::Value>, Local<v8::Context>)::(anonymous
class)::operator()(const FunctionCallbackInfo<v8::Value> &) const:
Assertion `args.IsConstructCall()' failed.
1: node::Abort()
[/Users/danielbevenius/work/nodejs/node/out/Debug/node]
2: node::Assert(char const* const (*) [4])
[/Users/danielbevenius/work/nodejs/node/out/Debug/node]
3: node::TLSWrap::Initialize(v8::Local<v8::Object>,
v8::Local<v8::Value>,
v8::Local<v8::Context>)::$_0::operator()(v8::FunctionCallbackInfo<v8::Value>
const&) const [/Users/danielbevenius/work/nodejs/node/out/Debug/node]
PR-URL: https://github.com/nodejs/node/pull/13097
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Currently the function CryptoPemCallback is used for two things:
1. As a passphrase callback.
2. To avoid the default OpenSSL passphrase routine.
The default OpenSSL passphase routine would apply if both
the callback and the passphrase are null pointers and the typical
behaviour is to prompt for the passphase which is not appropriate in
node.
This commit suggests that the PasswordCallback function only handle
passphrases, and that an additional function named NoPasswordCallback
used for the second case to avoid OpenSSL's passphase routine.
PR-URL: https://github.com/nodejs/node/pull/12827
Reviewed-By: Sam Roberts <vieuxtech@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Currently when building --without-ssl this test will report the
following error:
internal/util.js:82
throw new Error('Node.js is not compiled with openssl crypto
support');
This commit adds a check for crypto and skips this test if node was
built without ssl support.
PR-URL: https://github.com/nodejs/node/pull/13047
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: Rich Trott <rtrott@gmail.com>
Reviewed-By: David Cai <davidcai1993@yahoo.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Currently when building --without-ssl this test will report the
following error:
internal/util.js:82
throw new Error('Node.js is not compiled with openssl crypto
support');
This commit adds a check for crypto and skips this test if node was
built without ssl support.
PR-URL: https://github.com/nodejs/node/pull/13041
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Currently when running make targets the directory is printed on some
operating systems (Linux for example):
$ make lint
make[1]: Entering directory '/work/node'
Running JS linter...
./node tools/eslint/bin/eslint.js --cache --rulesdir=tools/eslint-rules
--ext=.js,.md \
benchmark doc lib test tools
make[1]: Leaving directory '/work/node'
make[1]: Entering directory '/work/node'
Running C++ linter...
On other operating systems the directory is not printed. This commit
suggests adding a flag to make this consistent for GNUMake by not
printing the directory.
PR-URL: https://github.com/nodejs/node/pull/13042
Reviewed-By: Gibson Fahnestock <gibfahn@gmail.com>
Original commit message:
This commit adds a getter for the private is_verbose_ member.
The use case for this comes from Node.js where the ability to avoid
calling FatalException if the TryCatch is verbose would be nice
to have.
BUG=
Review-Url: https://codereview.chromium.org/2840803002
Cr-Commit-Position: refs/heads/master@{#45018}
PR-URL: https://github.com/nodejs/node/pull/12826
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Currently this test will fail with the following error message when
configured --without-ssl:
Error: Node.js is not compiled with openssl crypto support
This commit checks for crypto and skips this tests if such support
is not available.
PR-URL: https://github.com/nodejs/node/pull/12987
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
When configured --without-ssl node_crypto.h will not be included but
async-wrap.h includes providers that are defined in node_crypto.h,
node_crypto.cc, and tls_wrap.cc:
AsyncWrap::PROVIDER_CONNECTION
AsyncWrap::PROVIDER_PBKDF2REQUEST
AsyncWrap::PROVIDER_RANDOMBYTESREQUEST
AsyncWrap::PROVIDER_TLSWRAP
These will be included as providers which will cause
test-async-wrap-getasyncid.js to fail.
This commit suggest adding a guard and exclude the providers that are
not available when configured --without-ssl
PR-URL: https://github.com/nodejs/node/pull/12967
Reviewed-By: Gibson Fahnestock <gibfahn@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Sam Roberts <vieuxtech@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Refael Ackermann <refack@gmail.com>
Currently the files specified in libraries in node.gyp `cctest` target are
getting a '.lib' extension on windows when generated with ninja.
This commit adds a check to see if a file has a '.obj' extension and in
that case no '.lib' extension will be added.
Also, the LIBS specified in the 'libraries' section are not
being included in the --start-group --end-group section which
means that these libraries will not be searched causing issue
with linkers where the order matters.
PR-URL: https://github.com/nodejs/node/pull/12484
Fixes: https://github.com/nodejs/node/issues/12448
Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: Gibson Fahnestock <gibfahn@gmail.com>
When working on commit 6a09a69ec9
("build: enable cctest to use generated objects") I did not take into
account building with ninja:
$ ./configure
$ tools/gyp_node.py -f ninja
$ ninja -C out/Release
$ ln -fs out/Release/node node
When ninja generated the ninja build files, src files that are
relative to the src directory will be named with a dot instead of a
path separator, for example:
out/Release/obj/src/node/node.o
would instead become:
out/Release/obj/src/node.node.o
This commit adds an additional variable for the type of object separator
used for this case.
Currently the check for if ninja is being used is a normal if statement
as are the following os checks (win and aix). But the win and aix ones
should only be evaluated if the build is not generated by ninja.
This commit turns this logic into an if ninja else statement.
PR-URL: https://github.com/nodejs/node/pull/12484
Fixes: https://github.com/nodejs/node/issues/12448
Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: Gibson Fahnestock <gibfahn@gmail.com>
Currently when building --without-ssl or --without-inspector there will
be an error when trying to set up the console in bootstrap_node.js:
Can't determine the arch of: 'out/Release/node'
bootstrap_node.js:276
if (!globalConsole.hasOwnProperty(key))
^
TypeError: Cannot read property 'hasOwnProperty' of undefined
at installInspectorConsole (bootstrap_node.js:276:25)
at get (bootstrap_node.js:264:21)
at evalScript (bootstrap_node.js:395:30)
at startup (bootstrap_node.js:125:9)
at bootstrap_node.js:537:3
I think this issue was introduced in commit
3f48ab3042 ("inspector: do not add
'inspector' property").
This commit attempts to fix this.
PR-URL: https://github.com/nodejs/node/pull/12881
Reviewed-By: Eugene Ostroukhov <eostroukhov@google.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
Reviewed-By: Refael Ackermann <refack@gmail.com>
This commit adds lldbinit files from upstream V8 and also adds these so
that they get installed when `make install` is run.
Original commit message:
[tools] add lldbinit
The goal of this commit is to add the equivalent to gdbinit but
for lldb. I've tried to replicate the commands as close as possible
but I'm unsure about the jss command and hoping to get some feedback
on it in addition to the bta command which I'm not sure how/when
this could be used. This is probably just inexperience on my part.
The lldbinit file can be placed into a directory prefixed with dot
(.lldbinit) and the python script is currently expected to be in the
same directory. The path to the script can be changed manually if
needed as well.
NOTRY=true
Review-Url: https://codereview.chromium.org/2758373002
Cr-Commit-Position: refs/heads/master@{#44136}
PR-URL: https://github.com/nodejs/node/pull/12061
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
root_cert_vector currently has file scope and external linkage, but is
only used in the NewRootCertsStore function. If this is not required to
be externally linked perhaps it can be changed to be static and function
scoped instead.
PR-URL: https://github.com/nodejs/node/pull/12788
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Shigeki Ohtsu <ohtsu@ohtsu.org>
While reading through node_crypto.cc I think the code could perhaps
be be a made a little clearer if CryptPemCallback was renamed.
I admit that I'm very new to the code base and openssl but having a
name like PasswordCallback or something similar would have helped me
so I'm suggesting this change.
PR-URL: https://github.com/nodejs/node/pull/12787
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
When configured --without-ssl the inspect-brk option will not be
available and the process will exit with a exit value of 9 "Invalid
Argument/Bad option".
This commit adds a skipIfInspectorDisabled check since --without-ssl
implies that no inspector support is build as well.
PR-URL: https://github.com/nodejs/node/pull/12757
Reviewed-By: Refael Ackermann <refack@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Michaël Zasso <targos@protonmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
When test-cli-node-options is run it uses the --trace-events-enabled
option which generates a file named node_trace.1.log. This commit
changes the working directory to the test tmp directory to avoid this
file being created in the project root.
PR-URL: https://github.com/nodejs/node/pull/12660
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Michaël Zasso <targos@protonmail.com>
Reviewed-By: Rich Trott <rtrott@gmail.com>
Daniel Bevenius