Firefox stil dies with aprs.fi

OH1KH Saku

unread,

Dec 18, 2016, 2:33:37 PM12/18/16

to aprs.fi

Sad to tell that FF50.0.2 did not solve the problem mentioned here earlier. It just looked like, after starting new version at first time.
After that I did not use "kitchen computer" for aprs.fi for many days.

With ham shack computer I use mainly Midori because it is light and works well with all needed pages including aprs.fi (excluding Facebook, but who cares :) . So that's why I did not notice earlier that problem still exists.

I did some more testing now and know something more about this:

When aprs.fi starts so that cursor is "white hand" everywhere, even when moved on targets, FF will die within few seconds.

When aprs.fi starts so that "white hand" changes to "black hand" when placed over a target FF will stay on working.

And how to get into "working situation"?
Remove "temporary files" and all cookies addressed to aprs.fi before enter to aprs.fi web page.
Both must be removed, it does not help to remove just files, or just cookies.

And as mentioned before aprs.fi seems to be the only web page that kills FF so far. All other pages are working normally.
This computer is Lenovo x201 laptop (quad core) with Fedora23/64bit/SSD Hd. So it should have power enough to handle all aprs.fi scripts without any pain.
And also with this computer and Midori aprs.fi works normally.

So something happened when FF 50.0 came and still exists at FF 50.0.2 that does not fit with aprs.fi.

--

73 Saku
OH1KH

Matthew Pitts

unread,

Dec 19, 2016, 2:08:27 AM12/19/16

to apr...@googlegroups.com

U am running Firefox 52 (the developer edition) on Windows 10 and cannot reproduce this. Is there anyone out there with a similar configuration that can?

--
You received this message because you are subscribed to the Google Groups "aprs.fi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to aprsfi+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Heikki Hannikainen

unread,

Dec 19, 2016, 2:54:07 AM12/19/16

to aprs.fi

Hi,

I just checked the Analytics for this. 17.9% of aprs.fi visits come from
users of Firefox. The bounce rate (amount of users which figure out the
site is not working, or otherwise leave the site immediately) is not
higher for Firefox than it is for Chrome (the most popular browser), and
most importantly, there is no significant change recently in bounce rate
for Firefox users. Bounce rate for Firefox 50.0 is not significantly
higher than for previous versions.

There are a lot of Firefox users there, and it is working for me on 50.0,
and others are reporting it works for them. Based on this, I can only
assume this is something that is quite specific to your configuration or
environment, not a generic Firefox issue. If it would totally not work on
a lot of 50.0 installs, there would be quite a few people asking about it.

Do you have some special things in use on your setup? Extensions/plugins,
can you try disabling/removing all of them? VPNs?

Can you check the Developer Tools error console, do you get any red errors
there when it fails?

- Hessu

OH1KH Saku

unread,

Dec 19, 2016, 10:46:50 AM12/19/16

to aprs.fi

Do you have some special things in use on your setup? 
Extensions/plugins, can you try disabling/removing all of them? VPNs?

I did remove all extensions I have not used for long time and disabled 
those I want still keep. Restarted FF and still have same issue.
Disabling Adblock looked first like it did solve the problem, but it 
just took longer to appear.

Can you check the Developer Tools error console, do you get any red 
errors there when it fails?

There is nothing to see as console goes to bit heaven with FF and at the 
moment I see "white hand" does not change any more to black one there is 
no (new) lines appearing to console.

This testing was made with ham shack HP Compaq mini. But it is 
soooooooooo sloooooow...with FF.

Then went to kitchen and did same extension blocking with same results. 
No glue.

The only working solution is to remove temporary files and cookies 
before entering to aprs.fi.
It works *always*. But the effect stays only as long as FF is open. No 
matter do I shut it down while aprs.fi is open or do I first move to 
another web page before closing.

I could understand this if it just would be one PC's problem.
But being the same with 2 computers although they do have same version 
OS (but it is not cloned!) makes me wonder what really is wrong.

As I mentioned at first mail (6-Dec-2016) I had also problem to enter 
groups.google.com (* only with ham shack PC) when FF was saying that java script must be enabled 
(even it was!), could perhaps lead to distro's Java problem.
But that has not happened now and there has not been any java updates 
lately.

So that leaves aprs.fi the only mysterious web page of those I normally 
visit.

Well. I can live with this (like with other unresolved bugs elsewhere).

Nice to be the only person in the universe that has same problem with FF 
on 2 different PCs .

Merry Xmas!
-- 
Saku
OH1KH

Heikki Hannikainen

unread,

Dec 19, 2016, 5:43:23 PM12/19/16

to aprs.fi

On Mon, 19 Dec 2016, OH1KH Saku wrote:

> As I mentioned at first mail (6-Dec-2016) I had also problem to enter
> groups.google.com (* only with ham shack PC) when FF was saying that java script must be enabled
> (even it was!), could perhaps lead to distro's Java problem.
> But that has not happened now and there has not been any java updates
> lately.

JavaScript is not Java, and Java is not JavaScript. JavaScript is a
scripting language that comes built in within the browser, and technically
has nothing to do with Java which you might install separately. You don't
need to even install Java to use JavaScript.

https://en.wikipedia.org/wiki/JavaScript#JavaScript_and_Java

> So that leaves aprs.fi the only mysterious web page of those I normally
> visit.
>
> Well. I can live with this (like with other unresolved bugs elsewhere).
> Nice to be the only person in the universe that has same problem with FF
> on 2 different PCs .

I googled around a bit ("fedora firefox 50 crash") and came upon this one:

https://bugzilla.redhat.com/show_bug.cgi?id=1396789

Can you try if those sites make it crash too:

"""
Steps to Reproduce:
1. Go to for example

https://www.tradingview.com/chart/?symbol=NYSE%3ASAN
http://www.investing.com/equities/tobii-ab-chart

and try to use the fullscreen technical chart.
"""

and then, as described in the comments, try to run firefox from a shell
window with:

$GDK_SYNCHRONIZE=1 firefox

and see if it still crashes.

In any case, if a browser crashes when you visit a web site, it's always a
browser bug. If a web site can kill a browser, the browser is at fault.

https://bugzilla.mozilla.org/show_bug.cgi?id=1271100#c45 hints that it
might be fixed in 50.1.0.

- Hessu

OH1KH Saku

unread,

Dec 20, 2016, 11:40:39 AM12/20/16

to aprs.fi

tiistai 20. joulukuuta 2016 0.43.23 UTC+2 Heikki Hannikainen kirjoitti:

JavaScript is not Java, and Java is not JavaScript. JavaScript is a

Thanks for pointing out my stupidity. Wasn't thinking what I was writing.

I googled around a bit ("fedora firefox 50 crash") and came upon this one:

You had better luck (=better words)!

http://www.investing.com/equities/tobii-ab-chart

At least this one works like aprs.fi. At first open everything is ok. When you close browser and open it again and go to that page FF dies.

I suspect that removing cookies and temporary files after first visit could help the second visit run through ok. But I didn't try it.

$GDK_SYNCHRONIZE=1 firefox

Did not try that yet. There was one link (@ #15)pointing to bug "ABORT: X_ShmAttach: BadAccess"

And suddenly another stupidity of mine cleared:

If I start firefox from command line, not from icon, it may leave something to console when it dies.

And yes it did! Same bug:

[5599] ###!!! ABORT: X_ShmAttach: BadAccess (attempt to access private resource denied); 16 requests ago: file /builddir/build/BUILD/firefox-50.0.2/firefox-50.0.2/toolkit/xre/nsX11ErrorHandler.cpp, line 147

[NPAPI 5854] ###!!! ABORT: Aborting on channel error.: file /builddir/build/BUILD/firefox-50.0.2/firefox-50.0.2/ipc/glue/MessageChannel.cpp, line 2056

Muistialueen ylitys (luotiin core-tiedosto)

But as those bugzilla people are going with Fedora 25 there may not be any hope for me as I'm still running F23.

I think I move to Debian as planned some times before. It has longer version lifetime.

In any case, if a browser crashes when you visit a web site, it's always a
browser bug. If a web site can kill a browser, the browser is at fault.

I never said it was web page's fault!

Suitable page just trigges the underlying bug. At this time it was aprs.fi (as I do not spent my time with those investor's pages that were on bug example :) ).

All I wanted to know was if someone else, using Linux (maybe even fedora and even v23) and firefox 50.0 / 50.0.2,

had noticed same kind of problem.

As it exists in bugzilla list it should happen also to other people using Fedora+FF and aprs.fi.

OK!

I think QSY CW now to play with my bug(s) :)

Thanks for comments!

--

Saku

OH1KH

Reply all

Reply to author

Forward