Add napi_fatal_error API #13971

kfarnung · 2017-06-28T21:19:14Z

Checklist

make -j4 test (UNIX), or vcbuild test (Windows) passes
tests and/or benchmarks are included
documentation is changed or added
commit message follows commit guidelines

Affected core subsystem(s)

n-api

kfarnung · 2017-06-28T21:19:47Z

/cc @jasongin @mhdawson

cjihrig · 2017-06-28T21:21:13Z

doc/api/n-api.md

+- `[in] location`: The location at which the error occurred.
+- `[in] message`: The message associated with the error.
+
+Call does not return, the process will be terminated.


Call -> The function call

cjihrig · 2017-06-28T21:28:45Z

src/node_api.cc

@@ -772,6 +772,11 @@ napi_status napi_get_last_error_info(napi_env env,
  return napi_ok;
 }

+NAPI_NO_RETURN void napi_fatal_error(const char* location,
+                                     const char* message) {
+  node::FatalError(location, message);


Doesn't this require node internals?

It does, but since it was already included in the file I assumed it wasn't a big deal. Since this is the v8-specific implementation of N-API, this can still be considered internal (I think).

Sorry, I forgot this hadn't landed. #13892

Ahh, ok, I hadn't seen that. It seems like it would be best then to adapt an implementation into node_api.cc that mimics the node::FatalError behavior?

Yes, that is what I had expected we would need to do.

Based on discussion today in the WG meeting, we'll be reverting back to the original change and removing the duplicated code.

cjihrig · 2017-06-28T21:29:30Z

test/addons-napi/test_fatal/binding.gyp

+  "targets": [
+    {
+      "target_name": "test_fatal",
+      "sources": [ "test_fatal.cc" ]


Any reason not to use just C?

cjihrig · 2017-06-28T21:31:11Z

test/addons-napi/test_fatal/test_fatal.cc

+#include <node_api.h>
+#include "../common.h"
+
+#if defined _WIN32


Do you need windows.h and/or unistd.h?

Nope, I meant to try removing this and forgot.

kfarnung · 2017-06-28T23:36:47Z

src/node_api.cc

+  fflush(stderr);
+  ABORT();
+  // to suppress compiler warning
+  ABORT();


Whoops, I'll clean this up.

kfarnung · 2017-06-29T00:20:18Z

@cjihrig @jasongin I grabbed the relevant portions of the node::FatalError method directly into node_api.cc, but it feels a little dirty to have to copy the PrintErrorString method. I don't see any other instances of print in node_api.cc, so I wasn't sure how else to do it.

kfarnung · 2017-06-29T01:13:19Z

Resolves #13927

jasongin · 2017-06-29T03:45:33Z

test/addons-napi/test_fatal/test.js

+const test_fatal = require(`./build/${common.buildType}/test_fatal`);
+
+// Exception thrown from async completion callback.
+// (Tested in a spawned process because the exception is fatal.)


I recognize these comments, they were copy-pasted from an async callback test I wrote. Please update them. :)

jasongin · 2017-06-29T03:56:25Z

src/node_api.cc

+
+  // Don't include the null character in the output
+  CHECK_GT(n, 0);
+  WriteConsoleW(stderr_handle, wbuf.data(), n - 1, nullptr, nullptr);


Does anyone know the reason for this special case of calling WriteConsoleW() for a windows console? node::Assert() is very similar to node::FatalError(), but it just does a regular fprintf(stderr, ...).

Oh, I think it's required for proper display of UTF-8 encoded error messages? I guess assertion messages are more likely to be limited to plain ASCII. I suppose we do want UTF-8 capability here.

TimothyGu · 2017-06-29T06:09:09Z

doc/api/n-api.md

+```
+
+- `[in] location`: The location at which the error occurred.
+- `[in] message`: The message associated with the error.


Make it clear that location is NULL-able, while message isn't. Speaking of that, I'd actually prefer the function signature to be (message, location) as it seems customary for NULL-able parameters to go last.

Speaking of this, maybe it would be helpful if message can be a printf format string instead? That'd invalidate my comment above but would provide callers more flexibility.

I think it would lead to mistakes when people are migrating code to N-API if the parameter order is different from node::FatalError().

bnoordhuis

Copy/pasting large chunks of code from src/node.cc is not acceptable. See also #13892 (comment).

mhdawson · 2017-06-29T19:23:43Z

Based on discussion today in N-API meeting we are going to look at having a version of the node internals file as part of the wrapper so that we can avoid having to have copies in the core Node.js codebase.

kfarnung · 2017-07-05T20:21:08Z

@bnoordhuis Can you take another look? I've eliminated all of the code duplication in node source files.

CI: https://ci.nodejs.org/job/node-test-pull-request/8996

bnoordhuis

LGTM with a style nit.

bnoordhuis · 2017-07-05T20:44:04Z

test/addons-napi/test_fatal/test.js

+}
+
+const p = child_process.spawnSync(
+  process.execPath, [ '--napi-modules', __filename, 'child' ]);


4 space indent here and below (line continuation.)

bnoordhuis · 2017-07-05T20:51:04Z

test/addons-napi/test_fatal/test_fatal.c

+  };
+
+  NAPI_CALL_RETURN_VOID(env, napi_define_properties(
+    env, exports, sizeof(properties) / sizeof(*properties), properties));


Tangential: macros like NAPI_ASSERT_BASE and NAPI_CALL_BASE should have their bodies wrapped in do { } while (0) in order for them to be safe to use in if/else statements. Right now they can create bugs that will be hard to figure out. Something for another PR.

I opened PR #14095 to fix this.

kfarnung · 2017-07-07T00:00:45Z

FreeBSD rerun: https://ci.nodejs.org/job/node-test-commit-freebsd/10188/

mhdawson

LGTM

kfarnung · 2017-07-07T22:18:09Z

@mhdawson @jasongin I believe this is ready to land whenever someone is ready to do it.

kfarnung · 2017-07-13T18:37:58Z

Rebased and new CI: https://ci.nodejs.org/job/node-test-pull-request/9113/

mhdawson · 2017-07-13T19:28:08Z

Failed in CI on PPCBE is known issue.

mhdawson · 2017-07-13T19:32:49Z

Failure on Alpine is known issue.

mhdawson · 2017-07-13T19:34:28Z

Linux fips issue was a machine issue.

mhdawson · 2017-07-13T22:47:47Z

Ok CI completed, landing.

PR-URL: #13971 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Jason Ginchereau <[email protected]> Reviewed-By: Ben Noordhuis <[email protected]> Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Michael Dawson <[email protected]>

mhdawson · 2017-07-13T23:07:11Z

landed as 73078d6

PR-URL: #13971 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Jason Ginchereau <[email protected]> Reviewed-By: Ben Noordhuis <[email protected]> Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Michael Dawson <[email protected]>

PR-URL: nodejs#13971 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Jason Ginchereau <[email protected]> Reviewed-By: Ben Noordhuis <[email protected]> Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Michael Dawson <[email protected]>

Backport-PR-URL: #19447 PR-URL: #13971 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Jason Ginchereau <[email protected]> Reviewed-By: Ben Noordhuis <[email protected]> Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Michael Dawson <[email protected]>

nodejs-github-bot added c++ Issues and PRs that require attention from people who are familiar with C++. node-api Issues and PRs related to the Node-API. labels Jun 28, 2017

cjihrig reviewed Jun 28, 2017

View reviewed changes

kfarnung force-pushed the fatal_error branch 2 times, most recently from 434945d to 70e0919 Compare June 28, 2017 22:14

kfarnung commented Jun 28, 2017

View reviewed changes

src/node_api.cc Outdated

fflush(stderr);

ABORT();

// to suppress compiler warning

ABORT();

Copy link

Contributor Author

kfarnung Jun 28, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Whoops, I'll clean this up.

kfarnung force-pushed the fatal_error branch 2 times, most recently from fc27c86 to 44bcf8e Compare June 29, 2017 00:05

jasnell approved these changes Jun 29, 2017

View reviewed changes

jasongin reviewed Jun 29, 2017

View reviewed changes

jasongin approved these changes Jun 29, 2017

View reviewed changes

kfarnung force-pushed the fatal_error branch from d15975c to abcf5b2 Compare June 29, 2017 06:05

TimothyGu reviewed Jun 29, 2017

View reviewed changes

bnoordhuis requested changes Jun 29, 2017

View reviewed changes

bnoordhuis mentioned this pull request Jun 29, 2017

N-API: Remove reference to node internals #13892

Closed

2 tasks

kfarnung force-pushed the fatal_error branch 2 times, most recently from e8ac6ae to bf61c1f Compare June 29, 2017 22:08

kfarnung mentioned this pull request Jun 30, 2017

Add support for Error::Fatal nodejs/node-addon-api#70

Closed

kfarnung force-pushed the fatal_error branch 2 times, most recently from 945c457 to ee3aa34 Compare July 5, 2017 20:18

bnoordhuis approved these changes Jul 5, 2017

View reviewed changes

cjihrig approved these changes Jul 5, 2017

View reviewed changes

kfarnung force-pushed the fatal_error branch 2 times, most recently from 4068780 to 6a32422 Compare July 6, 2017 22:53

mhdawson approved these changes Jul 7, 2017

View reviewed changes

n-api: add napi_fatal_error API

9f16c16

kfarnung force-pushed the fatal_error branch from 6a32422 to 9f16c16 Compare July 13, 2017 18:36

mhdawson closed this Jul 13, 2017

kfarnung deleted the fatal_error branch July 13, 2017 23:10

cjihrig mentioned this pull request Jul 14, 2017

N-API: add function for fatal error #13927

Closed

addaleax mentioned this pull request Jul 18, 2017

v8.2.0 proposal #13744

Merged

MylesBorins added the dont-land-on-v6.x label Aug 16, 2017

digitalinfinity mentioned this pull request Aug 31, 2017

Can't use node::FatalError() nodejs/abi-stable-node#232

Closed

MylesBorins mentioned this pull request Apr 16, 2018

v6.14.2 proposal #19996

Merged

Add napi_fatal_error API #13971

Add napi_fatal_error API #13971

Conversation

kfarnung commented Jun 28, 2017

Checklist

Affected core subsystem(s)

kfarnung commented Jun 28, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kfarnung commented Jun 29, 2017

kfarnung commented Jun 29, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

bnoordhuis left a comment

Choose a reason for hiding this comment

mhdawson commented Jun 29, 2017

kfarnung commented Jul 5, 2017

bnoordhuis left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kfarnung commented Jul 7, 2017

mhdawson left a comment

Choose a reason for hiding this comment

kfarnung commented Jul 7, 2017

kfarnung commented Jul 13, 2017

mhdawson commented Jul 13, 2017

mhdawson commented Jul 13, 2017

mhdawson commented Jul 13, 2017

mhdawson commented Jul 13, 2017

mhdawson commented Jul 13, 2017