lib/net: Convert to using internal/errors #14782

pmatzavin · 2017-08-12T05:16:21Z

Covert lib/net.js over to using lib/internal/errors.js

Ref: #11273

I have not addressed the cases that use errnoException(),
for reasons described in GH-12926

Replace thrown errors in lib/net.js
with errors from lib/internal/errors
Update tests according to the above modification

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)

lib/net, doc

Trott · 2017-08-12T06:46:16Z

doc/api/errors.md

@@ -1004,6 +1004,22 @@ Used when a callback is called more then once.
 can either be fulfilled or rejected but not both at the same time. The latter
 would be possible by calling a callback more then once.

+<a id="ERR_NET_INVALID_LISTEN_ARG"></a>
+### ERR_NET_INVALID_LISTEN_ARG


Any reason not to use the existing ERR_INVALID_ARG_TYPE instead of creating this new error?

There is an ongoing debate on whether error codes should be very specific (like this case), or more general (like what @Trott is suggesting)
For now the trend is towards a clear separation between bad arguments or options, and run time errors. But minimal separation between different argument errors.
@pmatzavin IMHO, the description you've given here should be in the net.Server.listen API documentation, and ERR_INVALID_ARG_TYPE is good enough.

@Trott @refack The reason that I used this new error was that the listen arg had many valid combinations of types/schema. I was not sure about the way that I handled it. I see your point and I agree with you. I will replaced it, with ERR_INVALID_ARG_TYPE.

@Trott @refack I am having trouble using the 'ERR_INVALID_ARG_TYPE' for the
Server.prototype.listen() in the lib/net.js.
I am describing the problem below :

The Server.prototype.listen() implementation
has some early return statements for the following cases:

options = options._handle || options.handle || options;

options instanceof TCP

typeof options.fd === 'number' && options.fd >= 0

(args.length === 0 || typeof args[0] === 'function' || (typeof options.port === 'undefined' && 'port' in options)) || typeof options.port === 'number' || typeof options.port === 'string'

options.path && isPipeName(options.path)

At the end, if none of these early return statements runs, then the error is thrown.

So I am having trouble using the
E('ERR_INVALID_ARG_TYPE', invalidArgType) error.
As I cannot compute a suitable error message using the
invalidArgType(name, expected, actual) (file: lib/internal/errors.js) callback,
for the case that I described above.

I have not yet found a similar case in the source code.
That's why I used the new error ERR_NET_INVALID_LISTEN_ARG in the first place.
Do you have any suggestions on how should I handle this case or on how should I use the ERR_INVALID_ARG_TYPE error for this case?

Should I refactor the invalidArgType method in the lib/internal/errors.js
to return a more generic error message (like: invalid type)
if the expected argument is not specified?

If I'm understanding correctly, the error isn't (necessarily?) an invalid argument type, but an invalid argument.

If we can make ERR_INVALID_OPT_VALUE work instead, that would be great. Will that result in a similar difficulty? If so, then I'm OK with a new error. If we've been trying for more general errors, E_INVALID_OPTION_OBJECT or something like that might be preferable so it can be re-used elsewhere. But there's a valid argument for specific codes too, so either way is fine by me.

@Trott I used the ERR_INVALID_OPT_VALUE error instead of creating a new error.
I amended the change and pushed -f to the current pull request.

Trott

Hi @pmatzavin! Welcome and thanks for the PR! I left a question/suggestion for a change.

pmatzavin · 2017-08-15T05:59:40Z

Hi @Trott! Thanks for the help and the suggestion. I am traveling right now without my laptop. I will update my PR by this Saturday.

discussion happened, dismissing review

Trott · 2017-08-19T17:33:00Z

@nodejs/ctc This is semver-major, needs some CTC reviews.

pmatzavin · 2017-08-20T09:51:31Z

updated

refack

LGTM
(just some nits)

refack · 2017-08-20T13:34:43Z

test/parallel/test-net-connect-options-port.js

-    const prefix = '"port" option should be a number or string: ';
-    const cleaned = opt.replace(/[\\^$.*+?()[\]{}|=!<>:-]/g, '\\$&');
-    return new RegExp(`^TypeError: ${prefix}${cleaned}$`);
+  function portTypeError() {


You can replace this with:

const portTypeError = common.expectsError({ code: 'ERR_INVALID_ARG_TYPE', type: TypeError }, 5); and the same in L46

@refack I have read the CONTRIBUTING.md but I am not sure which of the following should I use to make the requested changes:

amend the commit and push --force-with-lease

git commit --fixup

simply add a new commit

IMHO simply add a commit, since it makes re-reviews easier.
A common practice is to prefix the new commit's message with [fixup] ... so it's clear that it should be squashed when the PR is landed.

refack · 2017-08-20T13:42:28Z

test/parallel/test-net-localerror.js

-  assert.throws(() => {
-    net.connect(opts);
-  }, msg);
+function connect(opts, code, type) {


This can now be:

const connect = (opts, code, type) => common.expectsError( () => net.connect(opts), { code, type } );

refack · 2017-08-20T13:46:49Z

test/parallel/test-net-server-options.js

 const assert = require('assert');
+const common = require('../common');


Actually require('../common'); should be first. It does some wiring, and maybe we'll want it to monkeypatch assert in the future.

refack · 2017-08-20T13:56:11Z

Hello @pmatzavin and welcome. Thank you very much for the contribution 🥇
If you haven't already, it's recommended you take a look at CONTRIBUTING.md guide (especialy the part about "discuss and update"). Customarily PRs are kept open for at least 48 hour so that anyone interested gets a chance to comment or review.
About Error migrations, since they are semver-major they require 2 CTC members' approval (which can take longer than usual PR approval). A consequence of that is that this PR might become conflicted with master, so you might be asked to rebase it late.
Anyway thank you for the contribution, and good luck 🎉

P.S. If you have any question you can also feel free to contact me directly (@ mention, email, IM, or IRC)

jasnell · 2017-08-20T16:50:26Z

lib/net.js

-    throw new TypeError('options must be an object');
+    throw new errors.TypeError('ERR_INVALID_ARG_TYPE',
+                               'options',
+                               ['object'],


s/[object']/`Object'

When there is a single value, it doesn't need to be in an array

rvagg · 2017-08-20T23:55:50Z

SGTM, but I defer to @jasnell on the integration with the full new error messaging approach, when he gives a +1 then consider mine in here too

refack

Small non blocking nit

refack · 2017-08-21T02:58:20Z

lib/net.js

@@ -1485,7 +1497,9 @@ Server.prototype.listen = function() {
    return this;
  }

-  throw new Error('Invalid listen argument: ' + util.inspect(options));
+  throw new errors.TypeError('ERR_INVALID_OPT_VALUE',
+                             'listen options',


IMHO this should be just options so it will be rendered as:

The value "{ gaga: 4 }" is invalid for option "options"

which seems a tiny bit better then
https://github.com/nodejs/node/blob/77b02d2d8345c407dc0e0121420d823b2ce6fb18/test/parallel/test-net-server-listen-options.js#L63

refack

💯

refack · 2017-08-21T11:50:54Z

@jasnell PTAL

jasnell · 2017-08-23T18:52:57Z

lib/net.js

  if (type === 'PIPE') return new Pipe();
  if (type === 'TCP') return new TCP();
-  throw new TypeError('Unsupported fd type: ' + type);
+  throw new errors.Error('ERR_INVALID_HANDLE_TYPE');


I'd prefer the message to include the handle type.

@pmatzavin look at

node/lib/internal/errors.js

Line 200 in cbd3708

E('ERR_INVALID_FD', '"fd" must be a positive integer: %s');

You can add just after it:

E('ERR_INVALID_FD_TYPE', 'Unsupported fd type: %s');

jasnell · 2017-08-23T18:55:46Z

test/parallel/test-net-localerror.js

+const connect = (opts, code, type) => common.expectsError(
+  () => net.connect(opts),
+  { code, type }
+);


this doesn't quite match our typical code format .... it should be something like

const connect = (opts, code, type) => { common.expectsError(...); };

That's my mistake.
@pmatzavin for arrow functions that span multiple lines, in node's code it's prefered to use the {} syntax for extra clarity of the scope (even if it is just one expression like here).

jasnell

A few nits but generally LGTM

pmatzavin · 2017-08-25T05:27:04Z

@refack There was no test for the fd error thrown in the lib/net.js Socket() constructor so I added a test in a new file test/parallel/test-net-socket-fd-error.js PTAL. Sorry for the late update.

maclover7 · 2017-08-25T05:46:42Z

lib/net.js

    }
    if (!isLegalPort(port)) {
-      throw new RangeError('"port" option should be >= 0 and < 65536: ' + port);
+      throw new errors.RangeError('ERR_SOCKET_BAD_PORT');


Should the port be passed as a parameter to the throw here, and also on line 1475?

I think that would be nice 👍

The ERR_SOCKET_BAD_PORT definition in lib/internal/errors.js
does not handle a port argument. Should I refactor it to something like this?

E('ERR_SOCKET_BAD_PORT', (port) => { const portMsg = port ? `: ${port}` : ''; return `Port${portMsg} should be > 0 and < 65536`; });

I commited the change to pass the port value in the 'ERR_SOCKET_BAD_PORT' call in lib/net.js. This will not be included in the err msg until we modify the E('ERR_SOCKET_BAD_PORT', 'Port should be > 0 and < 65536'); in lib/internal/errors.js(see my above comment/question)

BridgeAR

Thanks a lot for your contribution. It looks pretty good overall. A few Error classes seemed to be mistaken though.

Besides that it would be nice to use common.expectsError(throwingFunction, options) instead of assert.throws(throwingFunction, common.expectsError(options)) but this should not block the PR.

BridgeAR · 2017-08-25T05:50:14Z

lib/net.js

@@ -1018,21 +1020,25 @@ function lookupAndConnect(self, options) {
  var localPort = options.localPort;

  if (localAddress && !cares.isIP(localAddress)) {
-    throw new TypeError('"localAddress" option must be a valid IP: ' +
-                        localAddress);
+    throw new errors.Error('ERR_INVALID_IP_ADDRESS', localAddress);


Why was this changed to a Error? It was a TypeError before?

I actually think not using TypeError is appropriate here. I've never felt that TypeError was a good fit.

my bad. I corrected this, now the error types in lib/net.js have the correct value.

@jasnell i have already committed (some minutes ago) a change that preserve the error types as they were in lib/net.js. So this is now a TypeError as before. Should I re-change it?

BridgeAR · 2017-08-25T05:51:37Z

test/parallel/test-net-server-listen-options.js

+  const assertPort = () => {
+    return common.expectsError({
+      code: 'ERR_SOCKET_BAD_PORT',
+      type: Error


This was a RangeError before?

Good point, this one should be a RangeError

my bad. I corrected this, now the error types in lib/net.js have the correct value.

BridgeAR · 2017-08-25T05:55:48Z

test/parallel/test-net-localerror.js

-  localPort: 'foobar',
-}, /^TypeError: "localPort" option should be a number: foobar$/);
+  localAddress: 'foobar',
+}, 'ERR_INVALID_IP_ADDRESS', Error);


This was a TypeError before?

my bad. I corrected this, now the error types in lib/net.js have the correct value.

BridgeAR · 2017-08-29T06:27:26Z

lib/net.js

    }
    if (!isLegalPort(port)) {
-      throw new RangeError('"port" option should be >= 0 and < 65536: ' + port);
+      throw new errors.RangeError('ERR_SOCKET_BAD_PORT', port);


The error type must also be able to handle the provided port.

I have made a change to include the port in the err msg if the port arg is passed to the error constructor.

BridgeAR · 2017-08-31T17:06:07Z

lib/internal/errors.js

+    portMsg = `:${port}`;
+  }
+  return `Port${portMsg} should be > 0 and < 65536`;
+});


This is more complicated then it has to be. I would rather have

E('ERR_SOCKET_BAD_PORT', 'Port should be > 0 and < 65536. Received %s.');

fixed (commit: b336aa1)

Covert lib/net.js over to using lib/internal/errors.js Ref: nodejs#11273 I have not addressed the cases that use errnoException(), for reasons described in nodejsGH-12926 - Replace thrown errors in lib/net.js with errors from lib/internal/errors. The ERR_INVALID_OPT_VALUE error have been used in the Server.prototype.listen() method after a discussion in Ref: nodejs#14782 - Update tests according to the above modifications

- refactor the portTypeError and portRangeError assertions, by passing the expected call count to common.expectsError in test-net-connect-options-port.js (Each syncFailToConnect() call will call 4 or 2 times the doConnect() method, wich returns an array of size 6, so the call counts are 96 and 168 for portTypeError and portRangeError respectively) - refactor the connect() method in test-net-localerror.js, by removing the reduntant assert and use the common.expectsError() assertion - move require('../common') to the top in test-net-server-options.js

- refactor the valid types argument of ERR_INVALID_ARG_TYPE errors, to be a string instead of an array when the number of valid types equals 1, in lib/net.js

replace "listen options" with "options"

- Create a new error 'ERR_INVALID_FD_TYPE' in lib/internal/errors.js, following the corresponding discussion at Ref: nodejs#14782. This error should be thrown when a file descriptor type is not valid. - add test for the new err in test-internal-errors.js - update the error doc in errors.md - use the new err in createHandle() in lib/net.js - Add brackets {} in arrow function in test-net-localerror.js to match the typical source code's format.

- add missing test for the 'ERR_INVALID_FD_TYPE' err that is thrown from the net.Socket constructor - correct the err type from Error to TypeError

- correct 'ERR_INVALID_IP_ADDRESS' type to TypeError err in net.js - correct 'ERR_SOCKET_BAD_PORT' type to RangeError in net.js - correct 'ERR_INVALID_OPT_VALUE' type to Error in net.js - pass the port value to the the 'ERR_SOCKET_BAD_PORT' err in net.js

- Modify the ERR_SOCKET_BAD_PORT definition in internal/errors.js to include the specified port in the err msg, if the port arg is passed to the err constructor. - Add tests for ERR_SOCKET_BAD_PORT msg in test-internal-errors.js

- simplify the ERR_SOCKET_BAD_PORT error definition in errors.js by replacing the cb arg with the string: 'Port should be > 0 and < 65536. Received %s.' - update the ERR_SOCKET_BAD_PORT test in test-internal-errors.js - update the ERR_SOCKET_BAD_PORT constructor call in lib/dgram.js and lib/dns to include the actual port as the second arg - update the already existing ERR_SOCKET_BAD_PORT message assertion in test-dns.js, to include the actual port.

jasnell · 2017-09-25T22:21:25Z

I went ahead and rebased the branch.

jasnell · 2017-09-25T22:22:01Z

CI: https://ci.nodejs.org/job/node-test-pull-request/10263/

jasnell · 2017-09-25T22:26:53Z

lib/net.js

@@ -1065,7 +1071,10 @@ function lookupAndConnect(self, options) {
  }

  if (options.lookup && typeof options.lookup !== 'function')
-    throw new TypeError('"lookup" option should be a function');
+    throw new errors.TypeError('ERR_INVALID_ARG_TYPE',


This should likely be ERR_INVALID_OPTION_VALUE

There are thrown errors for all the following 4 values in lib/net lookupAndConnect function :

function lookupAndConnect(self, options) { // ... options.port options.localAddress options.localPort options.lookup // ...

So I guess that for consistency, all the errors thrown for invalid values of the above vars should have a code of ERR_INVALID_OPTION_VALUE

I checked how we use mainly use this right now in the code base and it seems like ERR_INVALID_ARG_TYPE is the right type here. We might improve those types in general later on further but I think this should not block the PR from being merged.

jasnell · 2017-09-25T22:34:21Z

This is failing in CI: https://ci.nodejs.org/job/node-test-commit-linuxone/8827/nodes=rhel72-s390x/

pmatzavin · 2017-09-26T12:52:14Z

Sry for being unresponsive. I was unavailable due to some hlth problem. I will look into the issues.

jasnell · 2017-09-26T14:03:43Z

@pmatzavin ... no worries. Hope you're feeling better.

pmatzavin · 2017-09-27T18:20:55Z

I am getting different results when running the tests locally. The CI failing test(parallel/test-net-socket-fd-error) passes in my local env. I haven't found yet, why this can be inconsistent.
I double-checked that I have the correct source-code (after the rebase)

removed the test/parallel/test-net-socket-fd-error.js this test was added in the current PR but is not consistent. As the used fd mock value does not consistenly resolve to a fd type not equal to 'PIPE' and not equal to 'TCP'

pmatzavin · 2017-09-27T19:43:38Z

I removed the failing test / file:
test/parallel/test-net-socket-fd-error.js

This test was added by me in the current PR
but is not consistent. As the used
fd mock value does not consistently
resolve to a fd type not equal to 'PIPE'
and not equal to 'TCP'. The test was trying to assert the err in:

/* lib/net.js */
function createHandle(fd) {
  const type = TTYWrap.guessHandleType(fd);
  if (type === 'PIPE') return new Pipe();
  if (type === 'TCP') return new TCP();
  throw new errors.TypeError('ERR_INVALID_FD_TYPE', type);
}

by calling

new net.Socket({
 fd: 'invalid'
});

but i could not find a fd value/number that will fit for this test.

BridgeAR · 2017-09-27T22:24:55Z

@pmatzavin I think only one review has to be addressed and it should be good to go. I will start a CI as soon as that is done.

BridgeAR · 2017-10-02T00:38:37Z

CI https://ci.nodejs.org/job/node-test-pull-request/10369/

BridgeAR · 2017-10-02T02:51:43Z

Running the tests again https://ci.nodejs.org/job/node-test-pull-request/10374/

joyeecheung · 2017-10-14T15:02:09Z

CI before landing : https://ci.nodejs.org/job/node-test-pull-request/10717/

joyeecheung · 2017-10-15T03:55:31Z

Landed in 7f55349

Note that there is still #15589 that we need to fix after this (See #14782 (review) regarding using ERR_INVALID_OPT_VALUE v.s. ERR_INVALID_ARG_TYPE v.s. a new error for the overloaded server.listen()). IMO We should not release this PR until that gets resolved.

Covert lib/net.js over to using lib/internal/errors.js - Replace thrown errors in lib/net.js with errors from lib/internal/errors. The ERR_INVALID_OPT_VALUE error have been used in the Server.prototype.listen() method - Update tests according to the above modifications PR-URL: #14782 Refs: #11273 Reviewed-By: Refael Ackermann <refack@gmail.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Ruben Bridgewater <ruben@bridgewater.de> Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com> Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>

Covert lib/net.js over to using lib/internal/errors.js - Replace thrown errors in lib/net.js with errors from lib/internal/errors. The ERR_INVALID_OPT_VALUE error have been used in the Server.prototype.listen() method - Update tests according to the above modifications PR-URL: nodejs/node#14782 Refs: nodejs/node#11273 Reviewed-By: Refael Ackermann <refack@gmail.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Ruben Bridgewater <ruben@bridgewater.de> Reviewed-By: Michael Dawson <michael_dawson@ca.ibm.com> Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>

nodejs-github-bot added errors Issues and PRs related to JavaScript errors originated in Node.js core. net Issues and PRs related to the net subsystem. labels Aug 12, 2017

pmatzavin mentioned this pull request Aug 12, 2017

Tracking Issue: Migrate errors to internal/errors.js #11273

Closed

80 tasks

Trott requested changes Aug 12, 2017

View reviewed changes

Trott previously requested changes Aug 12, 2017

View reviewed changes

refack added the semver-major PRs that contain breaking changes and should be released in the next major version. label Aug 12, 2017

pmatzavin force-pushed the lib_net_convert_err branch from df9e7cc to 2e0914e Compare August 20, 2017 09:39

refack approved these changes Aug 20, 2017

View reviewed changes

jasnell reviewed Aug 20, 2017

View reviewed changes

refack approved these changes Aug 21, 2017

View reviewed changes

jasnell reviewed Aug 23, 2017

View reviewed changes

jasnell approved these changes Aug 23, 2017

View reviewed changes

jasnell approved these changes Aug 24, 2017

View reviewed changes

maclover7 reviewed Aug 25, 2017

View reviewed changes

BridgeAR requested changes Aug 25, 2017

View reviewed changes

BridgeAR requested changes Aug 29, 2017

View reviewed changes

BridgeAR reviewed Aug 31, 2017

View reviewed changes

matzavinos added 9 commits September 25, 2017 15:19

[fixup] lib/net: refac ERR_INVALID_ARG_TYPE err

16d8519

- refactor the valid types argument of ERR_INVALID_ARG_TYPE errors, to be a string instead of an array when the number of valid types equals 1, in lib/net.js

[fixup] lib/net: update Server.listen() err msg

90eb1a5

replace "listen options" with "options"

[fixup] lib/net: ERR_INVALID_FD_TYPE add test

7373ee2

- add missing test for the 'ERR_INVALID_FD_TYPE' err that is thrown from the net.Socket constructor - correct the err type from Error to TypeError

[fixup] lib/net: ERR_SOCKET_BAD_PORT handle port

74d6b0b

- Modify the ERR_SOCKET_BAD_PORT definition in internal/errors.js to include the specified port in the err msg, if the port arg is passed to the err constructor. - Add tests for ERR_SOCKET_BAD_PORT msg in test-internal-errors.js

jasnell force-pushed the lib_net_convert_err branch from e1834d5 to 5a4e982 Compare September 25, 2017 22:21

jasnell reviewed Sep 25, 2017

View reviewed changes

[fixup] lib/net: remove failing test

30a1ca0

removed the test/parallel/test-net-socket-fd-error.js this test was added in the current PR but is not consistent. As the used fd mock value does not consistenly resolve to a fd type not equal to 'PIPE' and not equal to 'TCP'

joyeecheung mentioned this pull request Oct 15, 2017

seems to not account for when a review has been dismissed nodejs/node-review#14

Closed

joyeecheung closed this Oct 15, 2017

jasnell mentioned this pull request Oct 30, 2017

Proposal: 2017-10-31, Version 9.0.0 (Current) #15135

Merged

		const assert = require('assert');
		const common = require('../common');

lib/net: Convert to using internal/errors #14782

lib/net: Convert to using internal/errors #14782

Conversation

pmatzavin commented Aug 12, 2017 • edited Loading

Checklist

Affected core subsystem(s)

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

pmatzavin Aug 20, 2017 • edited Loading

Choose a reason for hiding this comment

Trott left a comment

Choose a reason for hiding this comment

pmatzavin commented Aug 15, 2017

Trott commented Aug 19, 2017

pmatzavin commented Aug 20, 2017

refack 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

Choose a reason for hiding this comment

Choose a reason for hiding this comment

refack commented Aug 20, 2017

Choose a reason for hiding this comment

rvagg commented Aug 20, 2017

refack left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

refack left a comment

Choose a reason for hiding this comment

refack commented Aug 21, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

refack Aug 23, 2017 • edited Loading

Choose a reason for hiding this comment

jasnell left a comment

Choose a reason for hiding this comment

pmatzavin commented Aug 25, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

pmatzavin Aug 25, 2017 • edited Loading

Choose a reason for hiding this comment

pmatzavin Aug 25, 2017 • edited Loading

Choose a reason for hiding this comment

BridgeAR 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

pmatzavin Aug 25, 2017 • edited Loading

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

BridgeAR Aug 31, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jasnell commented Sep 25, 2017

jasnell commented Sep 25, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jasnell commented Sep 25, 2017

pmatzavin commented Sep 26, 2017

jasnell commented Sep 26, 2017

pmatzavin commented Sep 27, 2017 • edited Loading

pmatzavin commented Sep 27, 2017 • edited Loading

BridgeAR commented Sep 27, 2017

BridgeAR commented Oct 2, 2017

BridgeAR commented Oct 2, 2017 • edited Loading

joyeecheung commented Oct 14, 2017

joyeecheung commented Oct 15, 2017 • edited Loading

pmatzavin commented Aug 12, 2017 •

edited

Loading

pmatzavin Aug 20, 2017 •

edited

Loading

refack Aug 23, 2017 •

edited

Loading

pmatzavin Aug 25, 2017 •

edited

Loading

pmatzavin Aug 25, 2017 •

edited

Loading

pmatzavin Aug 25, 2017 •

edited

Loading

BridgeAR Aug 31, 2017 •

edited

Loading

pmatzavin commented Sep 27, 2017 •

edited

Loading

pmatzavin commented Sep 27, 2017 •

edited

Loading

BridgeAR commented Oct 2, 2017 •

edited

Loading

joyeecheung commented Oct 15, 2017 •

edited

Loading