FireFox 52 - how to disable location services (setting geo.enabled to False doesn't work)

Hello all,

I've recently installed FF 52 (my older FF 16 didn't have the demanded
encryption levels anymore, and I saw no way to upgrade them), and have a
number problems with it. One of them is that it is a blabbermouth, and I
can't seem to be able to completely shut it up.

First point in case: "location.services.mozilla.com"

Even though I've set, in "about:config" the entry "geo.enabled" to False and
I cannot find any other settings for it, it *still* (tries to) connect to
it.

Second point in case: "tiles.services.mozilla.com"

For this I've even set all the "browser.newtabpage" booleans to false, and
it *still* (tries to) connect to it.

Does anyone have any idea how to kill this "connect to the mothership"
behaviour ?


The preferred solution would be that *all* the "phoning home" is set to Off,
after which I than can switch the settings/services I think I need to On
(and wil get informed that that will need some "phoning home").

A super-duper cherry-on-top would be that these settings (somehow) become
default, so that when I create another profile I do not need to go thru the
whole dance again.

Regards,
Rudy Wieser

P.s.
FF 52 seems to be a slow beast in comparision to FF 16. Both in starting up
(even after the first time of the day) as well as in "inspect element".
Anything I can do about either ?

R.Wieser wrote:
Hello all,

I've recently installed FF 52 (my older FF 16 didn't have the demanded
encryption levels anymore, and I saw no way to upgrade them), and have a
number problems with it. One of them is that it is a blabbermouth, and I
can't seem to be able to completely shut it up.

First point in case: "location.services.mozilla.com"

Even though I've set, in "about:config" the entry "geo.enabled" to False and
I cannot find any other settings for it, it *still* (tries to) connect to
it.

Second point in case: "tiles.services.mozilla.com"

For this I've even set all the "browser.newtabpage" booleans to false, and
it *still* (tries to) connect to it.

Does anyone have any idea how to kill this "connect to the mothership"
behaviour ?


The preferred solution would be that *all* the "phoning home" is set to Off,
after which I than can switch the settings/services I think I need to On
(and wil get informed that that will need some "phoning home").

A super-duper cherry-on-top would be that these settings (somehow) become
default, so that when I create another profile I do not need to go thru the
whole dance again.

Regards,
Rudy Wieser

P.s.
FF 52 seems to be a slow beast in comparision to FF 16. Both in starting up
(even after the first time of the day) as well as in "inspect element".
Anything I can do about either ?



There appears to be a build time option.

https://bugzilla.redhat.com/show_bug.cgi?id=1063739

"Fedora Firefox RPM are build using --disable-necko-wifi ,
which means geolocation is disabled."

There's a single mozconfig file in the build directory
which you can set as required. For example, it can
support debug or release builds. The debug build
would have .pdb symbol files for debugging purposes
in Windows. It's assumed the build recipe still uses
the compiler and linker from a Visual Studio install.
(VS makes .pdb and works with WinDBG, while gcc tool
flow makes files that work with gdb, never the twain
shall meet).

https://developer.mozilla.org/en-US/..._Build_Options

And using the disable-necko-wifi then shows up in about: URI

"And, having to use it explicitely means it will show up in about:buildconfig"

https://bugzilla.mozilla.org/show_bug.cgi?id=653902

"necko wifi uses the system services to identify the MAC address
of nearby wifi access points. This is used to provide opt-in
geolocation services to the web."

Building used to be easy, in that the downloadable tarball
was pretty well ready to go. The last time, yeah, I got
a tarball, but then I had to clone a mercurial (HG) tree
of some sort, because stuff was missing from the tarball
and it was never going to work.

*******

In the example here, you can run a localhost and fake
your own location. The South Pole is popular this time
of year.

https://stackoverflow.com/questions/...tting#24937564

*******

You can shop here for a tarball, and at least
have source to read. And don't expect "geo.enabled"
to pop up if you run grep/findstring either. Again,
that would be too convenient. We can't have you
reading the source, now can we. 52.9.0esr is the
last ESR I see in the list.

http://releases.mozilla.org/pub/firefox/releases/

http://releases.mozilla.org/pub/fire...9.0esr/source/

File firefox-52.9.0esr.bundle 988M 25-Jun-2018 08:55
File firefox-52.9.0esr.source.tar.xz 204M 25-Jun-2018 08:55

And if you get the idea into your head, of "fixing" all
the line endings so Windows text editors work properly,
don't do that to the tree you will be building from. There
is *one* source file, that will throw a compile time error,
and no amount of fiddling with it, would tame the bug.
It was probably a bug in the compiler, but I couldn't
figure out how a line ending change (on hundreds of
thousands of other files), wasn't doing something
similar. I had a script that attempted to change the
line endings on around 50000 files or so, and just one file
of those select few, became unhappy.

Paul

Paul,

There appears to be a build time option.

https://bugzilla.redhat.com/show_bug.cgi?id=1063739

"Fedora Firefox RPM are build using --disable-necko-wifi ,
which means geolocation is disabled."

Geolocation is just one of the problems, and not even the most obnoxious one
either (they *promissed* they would first ask me before providing it to the
website that asks for it :-) ). I also mentioned
"tiles.services.mozilla.com", and I thought I got rid of
"shavar.services.mozilla.com", but it popped up again.

And although I think I can use that sourcefile (I can use it to try to
figure out how an SSL connection verifies the certificate
(SSL_AuthCertificateHook) - something I've been unable to find), I'm not at
all sure if I want to build a full environment just to recompile FF (with
one feature feature less). I'm not even sure if I can still find a
compiler which would work on XPsp3 and be able to handle that source ...

In the example here, you can run a localhost and fake
your own location.

Yup, I found a similar article too. And although somewhat funny, it would
do me no good: I'm running my browser with JS disabled - which means a
website would not be able to execute the request anyway. :-)

You can shop here for a tarball, and at least
have source to read

And that can never hurt. Especially not when I still have something to
figure out. So thanks.

By the way, what is that "firefox-52.9.0esr.bundle" supposed to be ? A
bundle of something, but it dosn't say what (and I do not really want to
download a gig only to throw it away 5 seconds later ...)

And if you get the idea into your head, of "fixing" all
the line endings so Windows text editors work properly,

... which is something I normally do. Way-back-when (in DOS times) I wrote a
small program for it I'm stil using (there might be better/easier programs
for it, but that one is mine :-) )

Regards,
Rudy Wieser

P.s.
Currently I'm blocking the domains another way, so they do not get out. But
it irks me to no end that, as far as I'm concerned, FF misbehaves this way.

R.Wieser wrote:
Hello all,

I've recently installed FF 52 (my older FF 16 didn't have the demanded
encryption levels anymore, and I saw no way to upgrade them), and have a
number problems with it. One of them is that it is a blabbermouth, and I
can't seem to be able to completely shut it up.

First point in case: "location.services.mozilla.com"

Even though I've set, in "about:config" the entry "geo.enabled" to False and
I cannot find any other settings for it, it *still* (tries to) connect to
it.

Second point in case: "tiles.services.mozilla.com"

For this I've even set all the "browser.newtabpage" booleans to false, and
it *still* (tries to) connect to it.

Does anyone have any idea how to kill this "connect to the mothership"
behaviour ?


The preferred solution would be that *all* the "phoning home" is set to Off,
after which I than can switch the settings/services I think I need to On
(and wil get informed that that will need some "phoning home").

A super-duper cherry-on-top would be that these settings (somehow) become
default, so that when I create another profile I do not need to go thru the
whole dance again.

Regards,
Rudy Wieser

P.s.
FF 52 seems to be a slow beast in comparision to FF 16. Both in starting up
(even after the first time of the day) as well as in "inspect element".
Anything I can do about either ?



OK, how about this. Hard to believe this is the right control.

https://www.thewindowsclub.com/disab...ation-browsers

"Click on Settings and press the Privacy tab. Here under Tracking,
check the Tell websites I do not want to be tracked check-box."

Good luck - beats building from source.

*******

With regard to speeding up your new browser,
you could test with a new video card :-/
Disabling hardware acceleration, just makes the
CPU work harder, so you should not be able to
beat a speed problem by doing that.

That's the problem with compositing at 60FPS and
relying on the video card. The user better have
a shiny fresh new video card. The chances of
us getting a buzzword compliant video card with
the right hardware interface on it now, are just
about zero. NVidia no longer even provides 32 bit
drivers, let alone new WinXP drivers. NVidia now
is 64-bit.

*******

"In Firefox, type about:support in the URL bar, and check under
"Graphics", it should say "GPU Accelerated Windows: 1/1", i.e.
one of one browser windows is using HW acceleration. If it says
"0/1", then you probably have compatibility problems (GPU driver
issues that would lead to security holes or crashes, mostly),
and it won't matter whether you enable acceleration or not."

Dunno which version of Firefox that applies to.

HTH,
Paul

Paul,

OK, how about this. Hard to believe this is the right control.

Than thats two of us. When I go to that setting it says "Firefox will send
a signal that you don't want to be tracked whenever Tracking Protection is
on". The whole point is that I want *less* blabbering, not more. :-)

With regard to speeding up your new browser,
you could test with a new video card :-/

Lolz. Thats not going to happen on old hardware running XP I'm afraid.
I'm not even sure if a "new" videocard for my hardware does still exist.

"In Firefox, type about:support in the URL bar, and check under
"Graphics", it should say "GPU Accelerated Windows: 1/1",

I can't find such a line, but I noticed something which could explain some
of the slowness (though I have little problems with its overall performance)

"unavailable by default: Direct2D requires Direct3D 11 compositing"

Never knew that XPsp3 supported D3D 11 ...

Dumb question maybe, but do you have any idea if I can maybe downgrade the
browser a bit but still have the newer encryptions supported ? I sought
for a feature list once (in regard to supported encryptions), but have never
been able to find one.

Regards (and thank you)
Rudy Wieser

R.Wieser wrote:
Paul,

OK, how about this. Hard to believe this is the right control.

Than thats two of us. When I go to that setting it says "Firefox will send
a signal that you don't want to be tracked whenever Tracking Protection is
on". The whole point is that I want *less* blabbering, not more. :-)

With regard to speeding up your new browser,
you could test with a new video card :-/

Lolz. Thats not going to happen on old hardware running XP I'm afraid.
I'm not even sure if a "new" videocard for my hardware does still exist.

"In Firefox, type about:support in the URL bar, and check under
"Graphics", it should say "GPU Accelerated Windows: 1/1",

I can't find such a line, but I noticed something which could explain some
of the slowness (though I have little problems with its overall performance)

"unavailable by default: Direct2D requires Direct3D 11 compositing"

Never knew that XPsp3 supported D3D 11 ...

Dumb question maybe, but do you have any idea if I can maybe downgrade the
browser a bit but still have the newer encryptions supported ? I sought
for a feature list once (in regard to supported encryptions), but have never
been able to find one.

Regards (and thank you)
Rudy Wieser

Direct2D has existed for a good while.

Compositing requires 128MB of RAM, on a typical platform.
My Macintosh is the first hardware here, with a compositing
layer buttered on top. And the same minimum requirement
might still be enough. The platform there had a 128MB minimum
too. It was based on an estimate of how many panes
might reasonably be opened at once, then sorted
in Z buffer order and presented in a final representation.

The question would be, whether actual D3D 11 constructs
are #included or it's just the bollox of the support
files. Like, maybe the version of Visual Studio, when
called upon to offer graphics primitives, dredged that
up. I think there might even be an option in Visual
Studio to crank that down.

Compositing in the CPU is a bad idea, as you can
imagine. It certainly is in Linux VMs, where compiz
is running. It sucks all the life out of the VM to
do that.

Um, good luck with that project... I guess :-)

Browsers and old OSes - the war never ends.

I often dreamed of moving the crypto suite
from a modern FF to an older one, but
what are the odds that is modular. Maybe
I can find a Vegas bookie to give me odds.

Paul

Paul,

Direct2D has existed for a good while.

I know. But as it specifically mentions D3D 11 ....

The question would be, whether actual D3D 11 constructs
are #included or it's just the bollox of the support
files.

I don't think I've got the expertise to work that out (other than installing
D3D versions in a row and see which one will cause the message to
disappear - which is something I do not really spend my time on).

Browsers and old OSes - the war never ends.

I often dreamed of moving the crypto suite
from a modern FF to an older one, but
what are the odds that is modular.

I already did so to see if I could move the FF 52 parts onto FF 16. I
intercepted the function calls to the different DLLs in both versions, and
tried to match them up. I gave up on it when I saw too many functions
disappearing and structures changing. :-(


But what I tried to ask was if you knew if there maybe is an earlier version
of FF which supports encryptions just like FF 52 does, but doesn't have that
"phone home" stuff yet.

Regards,
Rudy Wieser

R.Wieser wrote:
But what I tried to ask was if you knew if there maybe is an earlier version
of FF which supports encryptions just like FF 52 does, but doesn't have that
"phone home" stuff yet.

Regards,
Rudy Wieser

That's just not going to happen.

Who knows whether anything in the latest crypto package,
has dependencies elsewhere in the executable. Even if
it is tie-ins to some logging subsystem or something,
it might still have been made non-modular enough to
prevent easy swappage.

While some students did a project where they were
supposed to draw a high level block diagram by
examining the code, there are so many files in
the package, what single individual has ever looked
at each and every file ? The bundle might have 600000 files,
the tarball about 400000 files. Many files are test
cases to run on the results of compilation, so they're
not all actual source files.

Even the text editor that was used to edit the
files isn't all that consistent. I ran the Linux
"file" command through the tree, and it coughed up
all sorts of line-ending combinations. It's a mess
in there.

I even tried single-stepping in Windbg in there one
day, and gave up after two hours. I was in the
print code. I single stepped through three routines
with different names, which seemed to be doing the
same things, but with a slightly different subset
of print preferences. It was almost like V1 of some
print routine was called first, then a V2 version,
then a V3 version, sequentially. Like the author of
that mess, couldn't be bothered to clean up and
remove the code that was no longer relevant. Now
imagine if the crypto person started doing that
to a module.

Maybe it's dead-easy to back port the crypto,
but my exposure to the code so far doesn't
build confidence.

Paul

R.Wieser wrote:
Paul,

OK, how about this. Hard to believe this is the right control.

Than thats two of us. When I go to that setting it says "Firefox will send
a signal that you don't want to be tracked whenever Tracking Protection is
on". The whole point is that I want *less* blabbering, not more. :-)

With regard to speeding up your new browser,
you could test with a new video card :-/

Lolz. Thats not going to happen on old hardware running XP I'm afraid.
I'm not even sure if a "new" videocard for my hardware does still exist.

"In Firefox, type about:support in the URL bar, and check under
"Graphics", it should say "GPU Accelerated Windows: 1/1",

I can't find such a line, but I noticed something which could explain some
of the slowness (though I have little problems with its overall performance)

"unavailable by default: Direct2D requires Direct3D 11 compositing"

Never knew that XPsp3 supported D3D 11 ...

Dumb question maybe, but do you have any idea if I can maybe downgrade the
browser a bit but still have the newer encryptions supported ? I sought
for a feature list once (in regard to supported encryptions), but have never
been able to find one.

Regards (and thank you)
Rudy Wieser

There are supposed to be controls in about:config
that affect D3D usage.

The set of three is just about video playback, and
the example is probably irrelevant for the majority
of web content.

I basically picked these up, while looking for the
"right abbreviation" for them. I still cannot find
any substantive and modern description of the
dependencies. The "allow-d3d9-fallback" causes graphics
artifacts or even crashing, and should be selected with
care. You can always edit the setting later with a
text editor, if the browser won't stay running long
enough to switch it back.

layers.allow-d3d9-fallback : true

media.mediasource.webm.enabled : true
media.hardware-video-decoding.enabled : false
media.windows-media-foundation.allow-d3d11-dxva : false

Paul

In message , Paul
writes:
[]
"In Firefox, type about:support in the URL bar, and check under
"Graphics", it should say "GPU Accelerated Windows: 1/1", i.e.
one of one browser windows is using HW acceleration. If it says
"0/1", then you probably have compatibility problems (GPU driver
issues that would lead to security holes or crashes, mostly),
and it won't matter whether you enable acceleration or not."

Dunno which version of Firefox that applies to.

HTH,
Paul
It says "GPU Accelerated Windows 1/1 Direct3D 9" in "Version
27.0.1", for what it's worth (-:
--
J. P. Gilliver. UMRA: 1960/1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

live your dash. ... On your tombstone, there's the date you're born and the
date you die - and in between there's a dash. - a friend quoted by Dustin
Hoffman in Radio Times, 5-11 January 2013

J. P. Gilliver (John) wrote:
In message , Paul
writes:
[]
"In Firefox, type about:support in the URL bar, and check under
"Graphics", it should say "GPU Accelerated Windows: 1/1", i.e.
one of one browser windows is using HW acceleration. If it says
"0/1", then you probably have compatibility problems (GPU driver
issues that would lead to security holes or crashes, mostly),
and it won't matter whether you enable acceleration or not."

Dunno which version of Firefox that applies to.

HTH,
Paul
It says "GPU Accelerated Windows 1/1 Direct3D 9" in "Version
27.0.1", for what it's worth (-:

So yours is working a treat.

Paul

Paul wrote:
J. P. Gilliver (John) wrote:
In message , Paul
writes:
[]
"In Firefox, type about:support in the URL bar, and check under
"Graphics", it should say "GPU Accelerated Windows: 1/1", i.e.
one of one browser windows is using HW acceleration. If it says
"0/1", then you probably have compatibility problems (GPU driver
issues that would lead to security holes or crashes, mostly),
and it won't matter whether you enable acceleration or not."

Dunno which version of Firefox that applies to.

HTH,
Paul
It says "GPU Accelerated Windows 1/1 Direct3D 9" in "Version
27.0.1", for what it's worth (-:

So yours is working a treat.

Paul

More info here, using some keywords from your returned results.

https://wiki.mozilla.org/Blocklistin...aphics_Drivers

Paul

R.Wieser wrote:
Hello all,

I've recently installed FF 52

The .bundle file contains references to Mercurial Hg
and 7ZIP cannot take the file apart further. The
internal construction is reminiscent of tar.

You only need to download the .bundle, if following
a build recipe. Unpacking the .bundle, gives you
a Mercurial build tree with maybe clang or rust
or somesuch.

A machine with 32GB of RAM and a 64-bit OS, is
a good candidate for making a 32-bit version
of modern Firefox :-/

I had one build of Firefox, a while back, that
just barely fit in 3GB of working RAM during
the link stage. And you just know that's no longer
true. The linking stage has to be even bigger now.

Paul

Paul,

A machine with 32GB of RAM and a 64-bit OS, is
a good candidate for making a 32-bit version
of modern Firefox :-/

:-p That means that even if I wanted and would know how to I would not be
able to make it work. My beefiest machine has got just 4 GB of memory.
Yep, running XP.

Regards,
Rudy Wieser

On Tue, 31 Jul 2018 22:22:03 +0200, "R.Wieser"
wrote:

Paul,

A machine with 32GB of RAM and a 64-bit OS, is
a good candidate for making a 32-bit version
of modern Firefox :-/

:-p That means that even if I wanted and would know how to I would not be
able to make it work. My beefiest machine has got just 4 GB of memory.
Yep, running XP.



Try Palemoon Portable. Last build for XP is here. Still
handles most certificates/encryption. You can always delete the folder
if it's not what you want.



Install version is he



Runs fine on 1GB RAM. 4GB should be an overkill.
[]'s
--
Don't be evil - Google 2004
We have a new policy - Google 2012