errors: make code property mutable #15694

Trott · 2017-09-30T00:10:06Z

Userland code can break if it depends on a mutable code property for
errors. Allow users to change the code property but do not propagate
changes to the error name.

Additionally, make message and name consistent with Error object
(non-enumerable). Test that console.log() and .toString() calls on
internal Error objects with mutated properties have analogous results
with the standard ECMAScript Error objects.

Fixes: #15658

Checklist

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

Affected core subsystem(s)

errors

Trott · 2017-09-30T00:10:41Z

ping @jasnell @jdalton @evanlucas

Trott · 2017-09-30T00:11:04Z

Is this semver-major?

Trott · 2017-09-30T00:11:35Z

Also I couldn't locate user-facing documentation explaining that code was immutable, but maybe I didn't look in the right place. Did I miss it? @jasnell

jdalton · 2017-09-30T00:19:11Z

Here's my take on it:

function createClass(Super) {
  return class NodeError extends Super {
    constructor(key, ...args) {
      super(getMessage(key, args))
      this[codeSym] = key
    }

    get code() {
      return this[codeSym]
    }

    set code(value) {
      setProperty(this, "code", { value })
    }

    get name() {
      return super.name + " [" + this[codeSym] + "]"
    }

    set name(value) {
      setProperty(this, "name", { value })
    }
  }
}

The setProperty helper codifies to:

Object.defineProperty(this, "code", {
  configurable: true,
  enumerable: true,
  value,
  writable: true
})

This way they're still getters until they are overwritten.

Trott · 2017-09-30T04:35:27Z

This way they're still getters until they are overwritten.

Naïve question perhaps, but what's the benefit of keeping it as a getter?

jdalton · 2017-09-30T04:58:36Z

Naïve question perhaps, but what's the benefit of keeping it as a getter?

Keeps the existing dynamic behavior. I don't have the background on why the getter was used in the first place though.

Trott · 2017-09-30T05:02:59Z

CI: https://ci.nodejs.org/job/node-test-pull-request/10352/

jdalton · 2017-09-30T05:10:58Z

Shouldn't name be made mutable too?

Trott · 2017-09-30T06:02:16Z

Shouldn't name be made mutable too?

I suppose so. And probably add a test to confirm that when name and message have been modified, the results of e.toString() and console.log(e) are analogous to modifying those properties on an Error object and calling e.toString() and console.log(e) in the browser (assuming the console.log() behavior is reasonably consistent across browsers). Specifically, on Chrome at least, I see that toString() uses the new over-ridden values, but console.log() uses the original values.

Trott · 2017-09-30T06:52:27Z

CI: https://ci.nodejs.org/job/node-test-pull-request/10353/

targos · 2017-09-30T07:01:24Z

it looks like we can get rid of the symbol now

jdalton · 2017-09-30T08:09:27Z

lib/internal/errors.js

+        enumerable: false,
+        value: `${super.name} [${this[kCode]}]`,
+        writable: true
+      });
    }


If you moved the name assignment to the prototype it would align with builtin errors
(no-own name prop on error objects).

~~nix that~~

Object.defineProperty(NodeError.prototype, 'name', { configurable: true, enumerable: false, value: `${Base.prototype.name} [${this[kCode]}]`, writable: true }); return NodeError

Ah, this is why the getter and symbols where used!

Following up on my previous comment.
It looks like getters and a symbol was used to mimic built-in errors which have non-enumerable inherited properties (no-own properties).

So the tweaks needed to the existing errors would be adding setters for code and name to assign a value with enumerable, configurable, and writable of true. To avoid the current chattier console.log the kCode symbol assignment could be done with Object.defineProperty and made enumerable of false.

@Trott If you're up for it I could send a PR to your PR, since I'm kind of walking through the code as is and I've never had a commit merged to Node. 😋

It looks like getters and a symbol was used to mimic built-in errors which have non-enumerable inherited properties (no-own properties).

Yep. That's exactly why

@Trott If you're up for it I could send a PR to your PR, since I'm kind of walking through the code as is and I've never had a commit merged to Node.

@jdalton Sure, you can do that or if you'd rather just take it over, you can open your own PR and I can close this one. Either way is fine by me.

refack · 2017-10-01T20:12:11Z

@jdalton are you suggesting to make both .name and .code overridable? and in that case it will be possible to completely hide the preassigned code.

makeNodeErr((e) => {
  e.name = 'g';
  e.code = 'f';
  e.message = 'h';
  e.toString() === 'g: h';
});

How about keeping e[kCode] in some way part of the toString() output?
Or add a .nodeCode() accessor?

refack · 2017-10-01T20:17:21Z

it looks like we can get rid of the symbol now

As mentioned in #15694 (comment), pristine NodeErrors should have .name and .code as prototype getters so that instances of the class will only have .stack as an own property. In that case assigning the code to a symbol field still makes sense.

jdalton · 2017-10-01T21:10:16Z

@refack

are you suggesting to make both .name and .code overridable? and in that case it will be possible to completely hide the preassigned code.

Yep. Error objects are created with certain info then thrown to user-land where folks can do what they want with them.

How about keeping e[kCode] in some way part of the toString() output?
Or add a .nodeCode() accessor?

I'm not a fan of obstructing user-land here. My take is that Node creates an error that walks like an error and talks like an error (no own-enumerable properties, standard toString behavior, etc.) and then throws it out. After that it's no longer Node's concern. Folks can catch it, augment it, decorate it, or shallow it. The tweaks to makeNodeError are minimal and the kCode symbol will still exist on the error object so folks can restore the original getters/behavior of a modified error if desired.

Trott · 2017-10-01T23:21:15Z

CI: https://ci.nodejs.org/job/node-test-pull-request/10367/

Trott · 2017-10-02T05:07:06Z

@jdalton I have to make one significant change in addition to a few unimportant linting changes.

The significant change is to define the name property with enumerable: false to be consistent with standard ES Error objects. The test fails otherwise (which is why CI is red red red). I'm assuming that I should do that, but just in case that's totally wrong in your view or anyone else's, I'm mentioning it here...

I'm leaving code as enumerable.

Trott · 2017-10-02T05:12:15Z

CI: https://ci.nodejs.org/job/node-test-pull-request/10375/

jdalton · 2017-10-02T05:32:43Z

@Trott

The significant change is to define the name property with enumerable: false to be consistent with standard ES Error objects.

That's handled by the getter being defined in the class definition. It's non-enumerable by default.
When you set an own property however it will become enumerable:

var e = new Error("x")
console.log(e.name) // "Error"
console.log(Object.keys(e)) // []
e.name = "foo"
console.log(Object.keys(e)) // ["name"]

targos · 2017-10-02T07:59:09Z

+1 to make the name property enumerable after the setter is used.

jdalton · 2017-10-02T13:56:30Z

@Trott As mentioned in my previous comment, making the name property non-enumerable when assigned, myError.name = "foo", is the wrong behavior (doesn't match built-in). If you need to restore the original getter for unit tests you can simple delete the own property:

console.log(myError.name); // "foo"
delete myError.name
console.log(myError.name) // "Error [ERR_TLS_HANDSHAKE_TIMEOUT]"

That removes the own property and goes back to the prototypal property.
For reference I've done this in my own tests.

refack · 2017-10-02T15:16:12Z

I'm not a fan of obstructing user-land here. My take is that Node creates an error that walks like an error and talks like an error (no own-enumerable properties, standard toString behavior, etc.) and then throws it

My only concern is 3rd party unintentional clobbering. I'd rather we make it a tiny bit hard for libraries to completely clobber the new behaviour. So can we think of a way to get to the code even if .code and .name are overridden?

jasnell · 2017-10-02T15:40:04Z

lib/internal/errors.js

+    set name(value) {
+      defineProperty(this, 'name', {
+        configurable: true,
+        enumerable: false,


Likely should be true after setting.

jdalton · 2017-10-04T00:24:17Z

It kinda looks like you're in a bad testing state somehow @Trott. Could the test be running with showHidden turned on by accident? That would explain why you're seeing braces for untouched errors which have no enumerable properties but have the kCode as a non-enumerable property.

Trott · 2017-10-04T00:28:41Z

I'm not seeing that with this PR.

@jdalton I'm not seeing it anymore either even after I roll back the last commit I just added. I must have gotten confused about what binary corresponded to what code changes and been running a binary against an earlier version of this PR. Sorry about the unproductive detour.

Trott · 2017-10-04T00:29:57Z

@jdalton I'll split out the console.log()/util.inspect() changes into a separate PR so this one can land without an extended debate as to whether such things are breaking changes or not.

For internal errors, make `code` and `name` settable while keeping them non-own properties by default. Fixes: nodejs#15658

Trott · 2017-10-04T00:47:45Z

CI: https://ci.nodejs.org/job/node-test-pull-request/10398/

Trott · 2017-10-04T02:42:21Z

CI looks good. (OS X failure is infra issues and unrelated.)

Userland code can break if it depends on a mutable `code` property for errors. Allow users to change the `code` property but do not propagate changes to the error `name`. Additionally, make `message` and `name` consistent with `Error` object (non-enumerable). Test that `console.log()` and `.toString()` calls on internal `Error` objects with mutated properties have analogous results with the standard ECMAScript `Error` objects. PR-URL: nodejs#15694 Fixes: nodejs#15658 Reviewed-By: Evan Lucas <evanlucas@me.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>

For internal errors, make `code` and `name` settable while keeping them non-own properties by default. PR-URL: nodejs#15694 Fixes: nodejs#15658 Reviewed-By: Evan Lucas <evanlucas@me.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>

Trott · 2017-10-04T23:06:17Z

Landed in 6e172be and 42a2a9b . Thanks, @jdalton! 🎉

MylesBorins · 2017-10-07T21:41:13Z

This is not landing cleanly on 8.x, could you please send a backport.

Trott · 2017-10-08T00:25:21Z

Added the dont-land labels for v6.x and v4.x because the new error system does not exist in those release lines. Will backport for v8.x.

Trott · 2017-10-08T01:11:04Z

@MylesBorins Backport in #16078

Userland code can break if it depends on a mutable `code` property for errors. Allow users to change the `code` property but do not propagate changes to the error `name`. Additionally, make `message` and `name` consistent with `Error` object (non-enumerable). Test that `console.log()` and `.toString()` calls on internal `Error` objects with mutated properties have analogous results with the standard ECMAScript `Error` objects. PR-URL: nodejs/node#15694 Fixes: nodejs/node#15658 Reviewed-By: Evan Lucas <evanlucas@me.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>

For internal errors, make `code` and `name` settable while keeping them non-own properties by default. PR-URL: nodejs/node#15694 Fixes: nodejs/node#15658 Reviewed-By: Evan Lucas <evanlucas@me.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>

Userland code can break if it depends on a mutable `code` property for errors. Allow users to change the `code` property but do not propagate changes to the error `name`. Additionally, make `message` and `name` consistent with `Error` object (non-enumerable). Test that `console.log()` and `.toString()` calls on internal `Error` objects with mutated properties have analogous results with the standard ECMAScript `Error` objects. PR-URL: nodejs#15694 Fixes: nodejs#15658 Reviewed-By: Evan Lucas <evanlucas@me.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>

For internal errors, make `code` and `name` settable while keeping them non-own properties by default. PR-URL: nodejs#15694 Fixes: nodejs#15658 Reviewed-By: Evan Lucas <evanlucas@me.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>

Userland code can break if it depends on a mutable `code` property for errors. Allow users to change the `code` property but do not propagate changes to the error `name`. Additionally, make `message` and `name` consistent with `Error` object (non-enumerable). Test that `console.log()` and `.toString()` calls on internal `Error` objects with mutated properties have analogous results with the standard ECMAScript `Error` objects. Backport-PR-URL: #16078 PR-URL: #15694 Fixes: #15658 Reviewed-By: Evan Lucas <evanlucas@me.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>

For internal errors, make `code` and `name` settable while keeping them non-own properties by default. Backport-PR-URL: #16078 PR-URL: #15694 Fixes: #15658 Reviewed-By: Evan Lucas <evanlucas@me.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com>

nodejs-github-bot added the errors Issues and PRs related to JavaScript errors originated in Node.js core. label Sep 30, 2017

Trott mentioned this pull request Sep 30, 2017

Readonly error.code has ecosystem impact. #15658

Closed

Trott force-pushed the code-writable branch from 9245993 to 7436ac7 Compare September 30, 2017 06:49

jdalton reviewed Sep 30, 2017

View reviewed changes

Trott force-pushed the code-writable branch from 4bc91ea to fa6e02b Compare October 2, 2017 05:04

Trott force-pushed the code-writable branch 2 times, most recently from da94494 to aa33f44 Compare October 2, 2017 05:10

evanlucas approved these changes Oct 2, 2017

View reviewed changes

jasnell reviewed Oct 2, 2017

View reviewed changes

Trott added the console Issues and PRs related to the console subsystem. label Oct 4, 2017

error: make code and name properties settable

9efeea3

For internal errors, make `code` and `name` settable while keeping them non-own properties by default. Fixes: nodejs#15658

Trott force-pushed the code-writable branch from c686953 to 9efeea3 Compare October 4, 2017 00:44

Trott mentioned this pull request Oct 4, 2017

util,errors: make display of errors consistent with browsers #15765

Closed

3 tasks

Trott closed this Oct 4, 2017

MylesBorins added the backport-requested-v8.x label Oct 7, 2017

Trott added dont-land-on-v4.x labels Oct 8, 2017

lance added backported-to-v8.x and removed backport-requested-v8.x labels Oct 28, 2017

Trott deleted the code-writable branch January 13, 2022 22:46

errors: make code property mutable #15694

errors: make code property mutable #15694

Conversation

Trott commented Sep 30, 2017 • edited Loading

Checklist

Affected core subsystem(s)

Trott commented Sep 30, 2017

Trott commented Sep 30, 2017

Trott commented Sep 30, 2017 • edited Loading

jdalton commented Sep 30, 2017 • edited Loading

Trott commented Sep 30, 2017 • edited Loading

jdalton commented Sep 30, 2017 • edited by Trott Loading

Trott commented Sep 30, 2017

jdalton commented Sep 30, 2017

Trott commented Sep 30, 2017

Trott commented Sep 30, 2017

targos commented Sep 30, 2017

jdalton Sep 30, 2017 • edited Loading

Choose a reason for hiding this comment

jdalton Oct 1, 2017 • edited Loading

Choose a reason for hiding this comment

jasnell Oct 1, 2017

Choose a reason for hiding this comment

Trott Oct 1, 2017

Choose a reason for hiding this comment

refack commented Oct 1, 2017

refack commented Oct 1, 2017

jdalton commented Oct 1, 2017 • edited Loading

Trott commented Oct 1, 2017

Trott commented Oct 2, 2017

Trott commented Oct 2, 2017

jdalton commented Oct 2, 2017 • edited Loading

targos commented Oct 2, 2017

jdalton commented Oct 2, 2017 • edited Loading

refack commented Oct 2, 2017

jasnell Oct 2, 2017

Choose a reason for hiding this comment

jdalton commented Oct 4, 2017 • edited Loading

Trott commented Oct 4, 2017

Trott commented Oct 4, 2017

Trott commented Oct 4, 2017

Trott commented Oct 4, 2017

Trott commented Oct 4, 2017

MylesBorins commented Oct 7, 2017

Trott commented Oct 8, 2017

Trott commented Oct 8, 2017

Trott commented Sep 30, 2017 •

edited

Loading

Trott commented Sep 30, 2017 •

edited

Loading

jdalton commented Sep 30, 2017 •

edited

Loading

Trott commented Sep 30, 2017 •

edited

Loading

jdalton commented Sep 30, 2017 •

edited by Trott

Loading

jdalton Sep 30, 2017 •

edited

Loading

jdalton Oct 1, 2017 •

edited

Loading

jdalton commented Oct 1, 2017 •

edited

Loading

jdalton commented Oct 2, 2017 •

edited

Loading

jdalton commented Oct 2, 2017 •

edited

Loading

jdalton commented Oct 4, 2017 •

edited

Loading