Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Fedora's Firefox package to use the Mozilla Location Service
346 views
Skip to first unread message
Hanno Schlichting
Sep 3, 2014, 3:41:28 PM9/3/14
to Moz-Dev-Geolocation
Hi.
I’m happy to report that the Fedora project is going to use the Mozilla Location Service (MLS) as the default geolocation provider for their Firefox Desktop packages.
Most desktop users use installers and builds provided directly by Mozilla, which can also be used under Linux operating systems. But some users prefer to use packages and builds from a distribution specific package management or installation tool.
Mozilla has a contract with Google to provide geolocation services for our desktop users, but this is limited to the builds and installers we produce ourselves. If a third party provides a separate build or package of Firefox Desktop, they have to reach their own deal for geolocation services with Google, like for example the Ubuntu project has done. The Fedora project for various reasons has decided to not buy a commercial contract for geolocation services and thus has been left without working HTML5 geolocation API’s for some time.
Today we as Mozilla run our own geolocation project with the Mozilla Location Service (MLS). The service is still provided as-is and not mature, but it does provide a basic service offering. For users of Fedora’s Firefox package using MLS will certainly provide a better experience than not having geolocation at all.
In addition users can also use our apps (https://location.services.mozilla.com/apps) to contribute to MLS and ensure that we have location data for their neighborhoods, improving the quality of service they get.
If you are curious and want to test MLS in your own browser, you can go to the settings in about:config and adjust the geo.wifi.uri setting. If you are using a Firefox build provided by Mozilla, you can use the following value:
https://location.services.mozilla.com/v1/geolocate?key=%GOOGLE_API_KEY%
If you are using a build provided by a third party, you can use the value:
https://location.services.mozilla.com/v1/geolocate?key=test
but might run into rate limitations on the service side, to protect the service from abuse.
We do not have any plans to switch the default geolocation provider in Mozilla’s own builds. But using MLS is very much an option for any third party packaging Firefox Desktop. Please come talk to me or reach out on this mailing list.
Best,
Hanno
P.S. Our legal team has volunteered to update our privacy policies and explanations, and we’ll shortly extend the pages at https://www.mozilla.org/en-US/firefox/geolocation/ to cover the special Fedora/MLS use-case and prevent Fedora from having to set up their own privacy policy pages.
I’m happy to report that the Fedora project is going to use the Mozilla Location Service (MLS) as the default geolocation provider for their Firefox Desktop packages.
Most desktop users use installers and builds provided directly by Mozilla, which can also be used under Linux operating systems. But some users prefer to use packages and builds from a distribution specific package management or installation tool.
Mozilla has a contract with Google to provide geolocation services for our desktop users, but this is limited to the builds and installers we produce ourselves. If a third party provides a separate build or package of Firefox Desktop, they have to reach their own deal for geolocation services with Google, like for example the Ubuntu project has done. The Fedora project for various reasons has decided to not buy a commercial contract for geolocation services and thus has been left without working HTML5 geolocation API’s for some time.
Today we as Mozilla run our own geolocation project with the Mozilla Location Service (MLS). The service is still provided as-is and not mature, but it does provide a basic service offering. For users of Fedora’s Firefox package using MLS will certainly provide a better experience than not having geolocation at all.
In addition users can also use our apps (https://location.services.mozilla.com/apps) to contribute to MLS and ensure that we have location data for their neighborhoods, improving the quality of service they get.
If you are curious and want to test MLS in your own browser, you can go to the settings in about:config and adjust the geo.wifi.uri setting. If you are using a Firefox build provided by Mozilla, you can use the following value:
https://location.services.mozilla.com/v1/geolocate?key=%GOOGLE_API_KEY%
If you are using a build provided by a third party, you can use the value:
https://location.services.mozilla.com/v1/geolocate?key=test
but might run into rate limitations on the service side, to protect the service from abuse.
We do not have any plans to switch the default geolocation provider in Mozilla’s own builds. But using MLS is very much an option for any third party packaging Firefox Desktop. Please come talk to me or reach out on this mailing list.
Best,
Hanno
P.S. Our legal team has volunteered to update our privacy policies and explanations, and we’ll shortly extend the pages at https://www.mozilla.org/en-US/firefox/geolocation/ to cover the special Fedora/MLS use-case and prevent Fedora from having to set up their own privacy policy pages.
Zeeshan Ali (Khattak)
Sep 4, 2014, 8:07:54 AM9/4/14
to Hanno Schlichting, Moz-Dev-Geolocation
On Wed, Sep 3, 2014 at 8:41 PM, Hanno Schlichting
<hschli...@mozilla.com> wrote:
> Hi.
Hi Hanno,
> I’m happy to report that the Fedora project is going to use the Mozilla Location Service (MLS) as the default geolocation provider for their Firefox Desktop packages.
While I'm very happy about all MLS developments, rely on it a lot in
geoclue and help with collection of data as much as I can, I'm not
exactly happy with this development in particular. IMHO this will
undermine my efforts to push for geoclue to be *the* geolocation
framework every app should be using on Linux desktop (especially
Fedora, since I work for RH). It will make it harder for me to
convince about the importance of firefox needing to use geoclue and
therefore getting any contributor to make that happen [1].
This is also not so good for firefox actually since it can't use
modems and thus you are limited to geoip and wifi-geolocation only.
Just my two cents.
--
Regards,
Zeeshan Ali (Khattak)
________________________________________
Befriend GNOME: http://www.gnome.org/friends/
[1] https://bugzilla.mozilla.org/show_bug.cgi?id=485472
<hschli...@mozilla.com> wrote:
> Hi.
Hi Hanno,
> I’m happy to report that the Fedora project is going to use the Mozilla Location Service (MLS) as the default geolocation provider for their Firefox Desktop packages.
geoclue and help with collection of data as much as I can, I'm not
exactly happy with this development in particular. IMHO this will
undermine my efforts to push for geoclue to be *the* geolocation
framework every app should be using on Linux desktop (especially
Fedora, since I work for RH). It will make it harder for me to
convince about the importance of firefox needing to use geoclue and
therefore getting any contributor to make that happen [1].
This is also not so good for firefox actually since it can't use
modems and thus you are limited to geoip and wifi-geolocation only.
Just my two cents.
--
Regards,
Zeeshan Ali (Khattak)
________________________________________
Befriend GNOME: http://www.gnome.org/friends/
[1] https://bugzilla.mozilla.org/show_bug.cgi?id=485472
Hanno Schlichting
Sep 4, 2014, 8:39:58 AM9/4/14
to Zeeshan Ali (Khattak), Moz-Dev-Geolocation
Hi.
First off, I merged your two leaderboard entries ;)
On 04.09.2014, at 14:07 , Zeeshan Ali (Khattak) <zees...@gnome.org> wrote:
> While I'm very happy about all MLS developments, rely on it a lot in
> geoclue and help with collection of data as much as I can, I'm not
> exactly happy with this development in particular. IMHO this will
> undermine my efforts to push for geoclue to be *the* geolocation
> framework every app should be using on Linux desktop (especially
> Fedora, since I work for RH). It will make it harder for me to
> convince about the importance of firefox needing to use geoclue and
> therefore getting any contributor to make that happen [1].
>
> This is also not so good for firefox actually since it can't use
> modems and thus you are limited to geoip and wifi-geolocation only.
I think it makes sense for Fedora to use MLS via the existing code base. All that needs to be done to get working geolocation is a build config change and adding the api key. The code to talk to MLS / Google is already there in Firefox, so this is very simple non-coding change.
In my opinion switching to geoclue still makes sense for a different set of reasons. As you mentioned, geoclue can actually talk to more sources like GPS and cell modems. It can also act as an app-agnostic caching layer, so if a native maps app already got your location, maybe Firefox doesn’t have to call out to the network at all anymore, as geoclue already has a recent and good location fix.
And if we’d ever get any sort of offline database, geoclue would be a much better layer to implement it than making this specific to each app.
In that light geoclue makes sense to get better quality and improvements. That’s maybe a harder sell than just going from broken to basic functionality, but there’s still value to be had here.
Hanno
First off, I merged your two leaderboard entries ;)
On 04.09.2014, at 14:07 , Zeeshan Ali (Khattak) <zees...@gnome.org> wrote:
> While I'm very happy about all MLS developments, rely on it a lot in
> geoclue and help with collection of data as much as I can, I'm not
> exactly happy with this development in particular. IMHO this will
> undermine my efforts to push for geoclue to be *the* geolocation
> framework every app should be using on Linux desktop (especially
> Fedora, since I work for RH). It will make it harder for me to
> convince about the importance of firefox needing to use geoclue and
> therefore getting any contributor to make that happen [1].
>
> This is also not so good for firefox actually since it can't use
> modems and thus you are limited to geoip and wifi-geolocation only.
In my opinion switching to geoclue still makes sense for a different set of reasons. As you mentioned, geoclue can actually talk to more sources like GPS and cell modems. It can also act as an app-agnostic caching layer, so if a native maps app already got your location, maybe Firefox doesn’t have to call out to the network at all anymore, as geoclue already has a recent and good location fix.
And if we’d ever get any sort of offline database, geoclue would be a much better layer to implement it than making this specific to each app.
In that light geoclue makes sense to get better quality and improvements. That’s maybe a harder sell than just going from broken to basic functionality, but there’s still value to be had here.
Hanno
Chris Peterson
Sep 4, 2014, 12:47:54 PM9/4/14
to mozilla-dev...@lists.mozilla.org
On 9/4/14 5:39 AM, Hanno Schlichting wrote:
> I think it makes sense for Fedora to use MLS via the existing code base. All that needs to be done to get working geolocation is a build config change and adding the api key. The code to talk to MLS / Google is already there in Firefox, so this is very simple non-coding change.
This is the tricky part. Gecko's location code will need to be
> I think it makes sense for Fedora to use MLS via the existing code base. All that needs to be done to get working geolocation is a build config change and adding the api key. The code to talk to MLS / Google is already there in Firefox, so this is very simple non-coding change.
refactored slightly to use different providers on different systems.
Geoclue is the right solution for platforms where it is available.
Zeeshan: what became of the Geoclue patch in bug 485472? dougt r-'d the
proposed patch (in 2009), but AFAICT only for minor issues that can be
corrected, not for any fundamental disagreement about using Geoclue.
chris
Zeeshan Ali (Khattak)
Sep 4, 2014, 12:57:28 PM9/4/14
to Hanno Schlichting, Moz-Dev-Geolocation
On Thu, Sep 4, 2014 at 1:39 PM, Hanno Schlichting
> On 04.09.2014, at 14:07 , Zeeshan Ali (Khattak) <zees...@gnome.org> wrote:
>> While I'm very happy about all MLS developments, rely on it a lot in
>> geoclue and help with collection of data as much as I can, I'm not
>> exactly happy with this development in particular. IMHO this will
>> undermine my efforts to push for geoclue to be *the* geolocation
>> framework every app should be using on Linux desktop (especially
>> Fedora, since I work for RH). It will make it harder for me to
>> convince about the importance of firefox needing to use geoclue and
>> therefore getting any contributor to make that happen [1].
>>
>> This is also not so good for firefox actually since it can't use
>> modems and thus you are limited to geoip and wifi-geolocation only.
>
> In my opinion switching to geoclue still makes sense for a different set of reasons. As you mentioned, geoclue can actually talk to more sources like GPS and cell modems. It can also act as an app-agnostic caching layer, so if a native maps app already got your location, maybe Firefox doesn’t have to call out to the network at all anymore, as geoclue already has a recent and good location fix.
>
> And if we’d ever get any sort of offline database, geoclue would be a much better layer to implement it than making this specific to each app.
>
> In that light geoclue makes sense to get better quality and improvements. That’s maybe a harder sell than just going from broken to basic functionality, but there’s still value to be had here.
Fair enough, I'm glad we agree though on the need for switching to geoclue.
> First off, I merged your two leaderboard entries ;)
Cool, thanks. :)
> On 04.09.2014, at 14:07 , Zeeshan Ali (Khattak) <zees...@gnome.org> wrote:
>> While I'm very happy about all MLS developments, rely on it a lot in
>> geoclue and help with collection of data as much as I can, I'm not
>> exactly happy with this development in particular. IMHO this will
>> undermine my efforts to push for geoclue to be *the* geolocation
>> framework every app should be using on Linux desktop (especially
>> Fedora, since I work for RH). It will make it harder for me to
>> convince about the importance of firefox needing to use geoclue and
>> therefore getting any contributor to make that happen [1].
>>
>> This is also not so good for firefox actually since it can't use
>> modems and thus you are limited to geoip and wifi-geolocation only.
>
> I think it makes sense for Fedora to use MLS via the existing code base. All that needs to be done to get working geolocation is a build config change and adding the api key. The code to talk to MLS / Google is already there in Firefox, so this is very simple non-coding change.
Ah ok, a low hanging fruit then.
> In my opinion switching to geoclue still makes sense for a different set of reasons. As you mentioned, geoclue can actually talk to more sources like GPS and cell modems. It can also act as an app-agnostic caching layer, so if a native maps app already got your location, maybe Firefox doesn’t have to call out to the network at all anymore, as geoclue already has a recent and good location fix.
>
> And if we’d ever get any sort of offline database, geoclue would be a much better layer to implement it than making this specific to each app.
>
> In that light geoclue makes sense to get better quality and improvements. That’s maybe a harder sell than just going from broken to basic functionality, but there’s still value to be had here.
Zeeshan Ali (Khattak)
Sep 4, 2014, 1:02:21 PM9/4/14
to Chris Peterson, mozilla-dev...@lists.mozilla.org
Hi Chris,
On Thu, Sep 4, 2014 at 5:47 PM, Chris Peterson <cpet...@mozilla.com> wrote:
> On 9/4/14 5:39 AM, Hanno Schlichting wrote:
>>
> Zeeshan: what became of the Geoclue patch in bug 485472? dougt r-'d the
> proposed patch (in 2009), but AFAICT only for minor issues that can be
> corrected, not for any fundamental disagreement about using Geoclue.
Any patches on that bug are obsolete for an year now since geoclue2 is
a different beast than geoclue1. Not only geoclue1 is totally
unmaintained for years, we ripped the whole repository (since geoclue2
was a complete re-write) so it doesn't make any sense now to use
geoclue1.
On Thu, Sep 4, 2014 at 5:47 PM, Chris Peterson <cpet...@mozilla.com> wrote:
> On 9/4/14 5:39 AM, Hanno Schlichting wrote:
>>
>> I think it makes sense for Fedora to use MLS via the existing code base.
>> All that needs to be done to get working geolocation is a build config
>> change and adding the api key. The code to talk to MLS / Google is already
>> there in Firefox, so this is very simple non-coding change.
>
>
>> All that needs to be done to get working geolocation is a build config
>> change and adding the api key. The code to talk to MLS / Google is already
>> there in Firefox, so this is very simple non-coding change.
>
>
> This is the tricky part. Gecko's location code will need to be refactored
> slightly to use different providers on different systems.
>
> Geoclue is the right solution for platforms where it is available.
Yeah, that makes sense.
> slightly to use different providers on different systems.
>
> Geoclue is the right solution for platforms where it is available.
> Zeeshan: what became of the Geoclue patch in bug 485472? dougt r-'d the
> proposed patch (in 2009), but AFAICT only for minor issues that can be
> corrected, not for any fundamental disagreement about using Geoclue.
a different beast than geoclue1. Not only geoclue1 is totally
unmaintained for years, we ripped the whole repository (since geoclue2
was a complete re-write) so it doesn't make any sense now to use
geoclue1.
Chris Peterson
Sep 4, 2014, 1:15:09 PM9/4/14
to mozilla-dev...@lists.mozilla.org
On 9/4/14 10:02 AM, Zeeshan Ali (Khattak) wrote:
>> >Zeeshan: what became of the Geoclue patch in bug 485472? dougt r-'d the
>> >proposed patch (in 2009), but AFAICT only for minor issues that can be
>> >corrected, not for any fundamental disagreement about using Geoclue.
> Any patches on that bug are obsolete for an year now since geoclue2 is
> a different beast than geoclue1. Not only geoclue1 is totally
> unmaintained for years, we ripped the whole repository (since geoclue2
> was a complete re-write) so it doesn't make any sense now to use
> geoclue1.
Based on your recommendation, I closed bug 485472 because geoclue 1 is
>> >Zeeshan: what became of the Geoclue patch in bug 485472? dougt r-'d the
>> >proposed patch (in 2009), but AFAICT only for minor issues that can be
>> >corrected, not for any fundamental disagreement about using Geoclue.
> Any patches on that bug are obsolete for an year now since geoclue2 is
> a different beast than geoclue1. Not only geoclue1 is totally
> unmaintained for years, we ripped the whole repository (since geoclue2
> was a complete re-write) so it doesn't make any sense now to use
> geoclue1.
obsolete. Work to integrate geoclue2 should happen in a new bug to avoid
confusion.
Please note that writing the code to use geoclue2 is not on our team's
roadmap, but patches are happily accepted. <:)
chris
Zeeshan Ali (Khattak)
Sep 5, 2014, 2:18:03 PM9/5/14
to Chris Peterson, mozilla-dev...@lists.mozilla.org
On Thu, Sep 4, 2014 at 6:15 PM, Chris Peterson <cpet...@mozilla.com> wrote:
> On 9/4/14 10:02 AM, Zeeshan Ali (Khattak) wrote:
>>>
> On 9/4/14 10:02 AM, Zeeshan Ali (Khattak) wrote:
>>>
>>> >Zeeshan: what became of the Geoclue patch in bug 485472? dougt r-'d the
>>> >proposed patch (in 2009), but AFAICT only for minor issues that can be
>>> >corrected, not for any fundamental disagreement about using Geoclue.
>>
>> Any patches on that bug are obsolete for an year now since geoclue2 is
>> a different beast than geoclue1. Not only geoclue1 is totally
>> unmaintained for years, we ripped the whole repository (since geoclue2
>> was a complete re-write) so it doesn't make any sense now to use
>> geoclue1.
>
>
>>> >proposed patch (in 2009), but AFAICT only for minor issues that can be
>>> >corrected, not for any fundamental disagreement about using Geoclue.
>>
>> Any patches on that bug are obsolete for an year now since geoclue2 is
>> a different beast than geoclue1. Not only geoclue1 is totally
>> unmaintained for years, we ripped the whole repository (since geoclue2
>> was a complete re-write) so it doesn't make any sense now to use
>> geoclue1.
>
>
> Based on your recommendation, I closed bug 485472 because geoclue 1 is
> obsolete. Work to integrate geoclue2 should happen in a new bug to avoid
> confusion.
Sure.
> obsolete. Work to integrate geoclue2 should happen in a new bug to avoid
> confusion.
> Please note that writing the code to use geoclue2 is not on our team's
> roadmap,
>but patches are happily accepted. <:)
but being a total noob to gecko code and having had zero experience
with C++ for last 12 years, I'd need a lot more time to do this then a
regular firefox hacker so don't know when I can look into this.
0 new messages