Should we make content console messages print to stdout when browser.dom.window.dump.enabled is true?

Brian Grinstead bgrinstead at
Wed Aug 8 18:32:26 UTC 2018

Based on the discussion here and on the bug, I do think there’s value in being able to separately control chrome and content logs going to stdout. I’m leaning towards either:

1) Keep the current behavior for chrome callers and then add a second pref (false by default everywhere) to allow content logs to go out to stdout.
2) Stop overloading `browser.dom.window.dump.enabled`, and instead add two new prefs like `` and `devtools.console.stdout.content`. The chrome variation would be true by default in local builds and automation, I guess, and the content variation would be false by default everywhere.


> On Aug 6, 2018, at 11:21 AM, Eric Rahm <erahm at> wrote:
> Note the console message service also has 'consoleservice.logcat' [1] for android that dumps to logcat at the appropriate log levels when enabled (I believe it's enabled by default for nightly), we had a similar implementation for b2g as well [2] that covered the mapping of console functions to log levels. Our gecko logging code prefixes log levels [3] and supports logging only certain levels [4] to keep verbosity down, this might be something worth reusing.
> In general, please keep this pref opt-in and disabled by default in automation to avoid excessive log spam.
> -e
> [1]
> [2]
> [3]
> [4]
> On Fri, Aug 3, 2018 at 7:40 AM, Brian Grinstead <bgrinstead at> wrote:
> Huh - must have missed that during error console removal. I don’t think the pref name fits close enough with this proposed behavior to reuse it, so I’ll file a bug to remove it.
> Brian
> > On Aug 2, 2018, at 7:25 PM, Mike Hommey <mh at> wrote:
> > 
> >> On Fri, Aug 03, 2018 at 12:15:50PM +1000, Mark Hammond wrote:
> >>> On 3/08/2018 5:16 am, Brian Grinstead wrote:
> >>> In Bug 1439686 we started printing chrome console messages to stdout. At that time we restricted it to chrome callers to avoid extra noise in stdout during local builds - especially since filtering out content messages is a much-requested feature for the Browser Console (Bug 1260877).
> >>> 
> >>> Interestingly, a page can already get text out to stdout if browser.dom.window.dump.enabled is true using `window.dump`. But the output isn't great when dealing with objects, and it requires the page to change its code.
> >>> 
> >>> We got a request to enable this for console.log in Bug 1480544 - I'm curious if people have opinions on if/how this should be exposed. Should the output be controlled based on the browser.dom.window.dump.enabled pref? Should we have a second pref controlling itf? Something else?
> >> 
> >> This is tricky. I use dump() more than I'd care to admit - it works well in
> >> content code because you almost never see dump() used in real sites (does
> >> Chrome even support it?), so there's not a huge amount of "spam" in the
> >> terminal. Having dump() output from *both* chrome and content can be very
> >> helpful.
> >> 
> >> If we also wrote console.log() from content to the terminal, I fear we would
> >> then end up in a situation where content spams the terminal, with the only
> >> mitigation being to flip browser.dom.window.dump.enabled to false - but that
> >> means we also lose dump() output, throwing the baby out with the bathwater.
> >> 
> >> So my vote would be:
> >> 
> >> * Leave dump alone - it's basically a firefox-specific easter-egg and no one
> >> seems to mind that content *could* spam the terminal because it almost never
> >> does. (I know this wasn't proposed, but it's worth clarifying imo)
> >> 
> >> * A new pref that would allow content console.* output to be written to this
> >> terminal. I'd be fine with this defaulting to true wherever
> >> browser.dom.window.dump.enabled does.
> >> 
> >> Having "dump" in the name of this pref wouldn't be a good option, because
> >> IMO, it should *not* affect the behaviour of dump()
> > 
> > There's javascript.options.showInConsole, too, but I'm not sure it does
> > anything these days, despite being defined in all.js, firefox.js and
> > other places, it seems it was made a no-op by the error console removal
> > in bug 1278368.
> > 
> > Mike
> > _______________________________________________
> > firefox-dev mailing list
> > firefox-dev at
> >
> _______________________________________________
> firefox-dev mailing list
> firefox-dev at

More information about the firefox-dev mailing list