simplest domain.run use fails without nextTick if surrounded in try/catch

[bmeck]

From the REPL or other try/catch situation:

var domain = require('domain');
var d=domain.create();
d.on('error', function () {console.log('OK')})
d.run(function () {throw e})

Domain.run fails to use the domain and emit the error event unless the thrown error is on a different tick, should this be documented or changed?

[othiym23]

+1 for fixing this. Module code can throw on startup, even if the functionality is async, and those errors should be emitted on the domain.

[indutny]

+1 from me too. @isaacs ?

[cxreg]

I'm unable to reproduce this issue, in v0.8 it catches the error and prints 'OK'

What version are you seeing the problem on, and what is the result?

[bmeck]

@cxreg 0.8.16

Appears to only affect the repl???

git@(unnamed branch) bradleymeck node>./node
> .load dom.js
> var domain = require('domain');
undefined
> var d=domain.create();
undefined
> d.on('error', function () {console.log('OK')})
{ domain: null,
  _events: { error: [Function] },
  _maxListeners: 10,
  members: [] }
> d.run(function () {throw e})
ReferenceError: e is not defined
    at repl:1:27
    at Domain.bind.b (domain.js:201:18)
    at Domain.run (domain.js:141:23)
    at repl:1:4
    at REPLServer.self.eval (repl.js:109:21)
    at Interface.<anonymous> (repl.js:248:12)
    at Interface.EventEmitter.emit (events.js:96:17)
    at Interface._onLine (readline.js:200:10)
    at Interface._line (readline.js:518:8)
    at Interface._ttyWrite (readline.js:783:20)
> 
(^C again to quit)
> 
git@(unnamed branch) bradleymeck node>./node -v
v0.8.16
git@(unnamed branch) bradleymeck node>./node dom.js
OK

[cxreg]

Yes, this is an issue with the repl code using try/catch, which prevents the domain's uncaughtException handler from picking up the error

This patch appears to make it work as expected: https://gist.github.com/4525700

[cxreg]

It's worth mentioning that a similar concern has come up in node_redis, and it would be nice if there was a common idiom for mixing try/catch with domains, as it places a high mental burden on developers who want to support both.

[dougwilson]

It was also noted in express recently: expressjs/express#1467

[al6x]

It was also noted in express recently: expressjs/express#1467

The "bug" (not actually a bug) in express is different - the problem there - inner try/catch block in expressjs, domains itself are fine.

[dougwilson]

@alexeyPetrushin yes, I know, which is the same thing going on in the REPL. My comment was more for @cxreg on why there should be a common idiom for implementatiors.

[al6x]

@dougwilson got it, yea, seems like concept of domain sometimes conflicting with already existing solution.

[jasnell]

Appears to have been resolved.