Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 8 July 2017 at 00:20, Anton Lundin  wrote:
> On 07 July, 2017 - Lubomir I. Ivanov wrote:
>
>> On 7 July 2017 at 23:47, Linus Torvalds  
>> wrote:
>> > On Fri, Jul 7, 2017 at 1:41 PM, Lubomir I. Ivanov  
>> > wrote:
>> >>
>> >> it appears that caching google maps tiles violates the google maps
>> >> (free / public) usage rules.
>> >
>> > I read that as "you can cache for performance" reasons. Bad networking
>> > is a *big* performance reason.
>> >
>> > You can't use the map data for anything else, or index it in any other
>> > way. But that's ok, we don't do that.
>> >
>>
>> ok, i guess i'm going to test this html5 caching feature to see if it works:
>> https://stackoverflow.com/a/13334284
>>
>> right now what happens is that the tiles are downloaded and cached,
>> but once the internet connection drops, zooming in and out does not
>> re-display the previously downloaded tiles.
>> also to my understanding this caching might be limited per session,
>> unless this is implemented:
>> http://doc.qt.io/qt-4.8/qwebsettings.html#setOfflineWebApplicationCachePath
>>
>> will try to test the above solutions during the weekend.
>
> Cache manifests (or the whole manifest thingie for that) is a nightmare
> to work with and debug, so I suggest using a serviceworker instead.
>

i realized that cache manifests do not support wildcards and the
manifest cannot possibly cover all the google maps tile URLs.

> A quick search found me https://github.com/boopathi/sw-demo-iss
>

in the above page it says:

"Pending:
[ ] Haven't figured a way to cache Google Maps API results using
Service Workers."

i'm not sure if the tile fetching falls under the above scope, but i
think it does.
so if the above guy hasn't been able to do it, maybe it's just way too tricky.

can't find a single clear example online in the lines of "OK, here is
how you cache google maps tiles using service workers".

but also, service workers do not work with file:/, qrc:/ paths, so
they seem to require either a localhost HTTP server to be running or
in case of a remote connection a HTTPS.
so to my understanding we need to make Subsurface create a HTTP server
on localhost (on a free port), load all resources like index.html,
service-worker.js, etc from Qt resources and then the server needs to
return the contents of these files when the QWebView opens
http://localhost:/index.html.

from that point on we go back to the thing the github person mentions:
"[ ] Haven't figured a way to cache Google Maps API results using
Service Workers."

> We could probably hook into WebKit / Webengine and cache the requests
> our self. I haven't looked at the apis yet, but one can do such things
> in the android webview and I guess the api's have some sort of feature
> parity.
>

yes, i have a feeling the fetching / caching has to be done on the C++ front.
https://stackoverflow.com/a/4588113
https://stackoverflow.com/questions/20616064/qwebview-caching

^ each QWebPage can have a custom QNetworkAccessManager which can
override requests or dispatch singnals when a QNetworkRequest has
finished, which is pretty much a can cover the "fetch" listeners (in
service-worker terminology). in the createRequest() override. so it
seems possible.

but another gotcha is that QNetworkAccesManager is non-portable to QWebEngine.
https://stackoverflow.com/a/38221276

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Anton Lundin]

On 07 July, 2017 - Lubomir I. Ivanov wrote:

> On 7 July 2017 at 23:47, Linus Torvalds  wrote:
> > On Fri, Jul 7, 2017 at 1:41 PM, Lubomir I. Ivanov  
> > wrote:
> >>
> >> it appears that caching google maps tiles violates the google maps
> >> (free / public) usage rules.
> >
> > I read that as "you can cache for performance" reasons. Bad networking
> > is a *big* performance reason.
> >
> > You can't use the map data for anything else, or index it in any other
> > way. But that's ok, we don't do that.
> >
> 
> ok, i guess i'm going to test this html5 caching feature to see if it works:
> https://stackoverflow.com/a/13334284
> 
> right now what happens is that the tiles are downloaded and cached,
> but once the internet connection drops, zooming in and out does not
> re-display the previously downloaded tiles.
> also to my understanding this caching might be limited per session,
> unless this is implemented:
> http://doc.qt.io/qt-4.8/qwebsettings.html#setOfflineWebApplicationCachePath
> 
> will try to test the above solutions during the weekend.

Cache manifests (or the whole manifest thingie for that) is a nightmare
to work with and debug, so I suggest using a serviceworker instead.

A quick search found me https://github.com/boopathi/sw-demo-iss



We could probably hook into WebKit / Webengine and cache the requests
our self. I haven't looked at the apis yet, but one can do such things
in the android webview and I guess the api's have some sort of feature
parity.


//Anton


-- 
Anton Lundin+46702-161604
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 7 July 2017 at 23:47, Linus Torvalds  wrote:
> On Fri, Jul 7, 2017 at 1:41 PM, Lubomir I. Ivanov  wrote:
>>
>> it appears that caching google maps tiles violates the google maps
>> (free / public) usage rules.
>
> I read that as "you can cache for performance" reasons. Bad networking
> is a *big* performance reason.
>
> You can't use the map data for anything else, or index it in any other
> way. But that's ok, we don't do that.
>

ok, i guess i'm going to test this html5 caching feature to see if it works:
https://stackoverflow.com/a/13334284

right now what happens is that the tiles are downloaded and cached,
but once the internet connection drops, zooming in and out does not
re-display the previously downloaded tiles.
also to my understanding this caching might be limited per session,
unless this is implemented:
http://doc.qt.io/qt-4.8/qwebsettings.html#setOfflineWebApplicationCachePath

will try to test the above solutions during the weekend.

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Linus Torvalds]

On Fri, Jul 7, 2017 at 1:41 PM, Lubomir I. Ivanov  wrote:
>
> it appears that caching google maps tiles violates the google maps
> (free / public) usage rules.

I read that as "you can cache for performance" reasons. Bad networking
is a *big* performance reason.

You can't use the map data for anything else, or index it in any other
way. But that's ok, we don't do that.

 Linus
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Dirk Hohndel]

> On Jul 7, 2017, at 1:41 PM, Lubomir I. Ivanov  wrote:
> 
> On 7 July 2017 at 20:37, Anton Lundin  wrote:
>> 
>> The only downside I see is that we loose the little offline support we
>> have. Our current Marble map caches the tiles, so you can, if they are
>> cached, view the map offline.
>> 
>> 
> 
> indeed, i just read this:
> https://stackoverflow.com/a/6196728
> 
> also this, from a marble developer:
> https://forum.kde.org/viewtopic.php?f=21=97184#p205314
> 
> it appears that caching google maps tiles violates the google maps
> (free / public) usage rules.

Back when we started I asked a contact at Google and the (paraphrased)
response was "we'd never consider an open source project with such minimal
use in violation".

/D

___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 7 July 2017 at 20:37, Anton Lundin  wrote:
>
> The only downside I see is that we loose the little offline support we
> have. Our current Marble map caches the tiles, so you can, if they are
> cached, view the map offline.
>
>

indeed, i just read this:
https://stackoverflow.com/a/6196728

also this, from a marble developer:
https://forum.kde.org/viewtopic.php?f=21=97184#p205314

it appears that caching google maps tiles violates the google maps
(free / public) usage rules.

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Linus Torvalds]

On Fri, Jul 7, 2017 at 10:37 AM, Anton Lundin  wrote:
>
> The only downside I see is that we loose the little offline support we
> have. Our current Marble map caches the tiles, so you can, if they are
> cached, view the map offline.

Oh, you're right - good call.

That offline caching is *really* convenient when you are in odd places
(and most diving spots count as "odd").

So losing caching is actually a fairly big deal. Not a deal-breaker,
but it's definitely noticeable.

Linus
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Anton Lundin]

On 07 July, 2017 - Lubomir I. Ivanov wrote:

> On 7 July 2017 at 16:54, Dirk Hohndel  wrote:
> >
> >> On Jul 7, 2017, at 6:38 AM, Lubomir I. Ivanov  wrote:
> >>
> >> On 7 July 2017 at 00:48, Lubomir I. Ivanov  wrote:
> >>> On 7 July 2017 at 00:26, Davide DB  wrote:
>  http://i.imgur.com/o8IP4Ho.png
> >>>
> >>> this is possible with the so called MarkerCluster libarary:
> >>> https://developers.google.com/maps/documentation/javascript/marker-clustering
> >>> https://github.com/googlemaps/v3-utility-library/tree/master/markerclusterer
> >>>
> >>> i haven't tried it, but Google promote it and it seems easy to use.
> >>>
> >>
> >> quickly sketched the cluster idea.
> >>
> >> here is a preview:
> >> https://www.dropbox.com/s/6gdygq68uba2pg8/cluster_clicks.png?dl=0
> >>
> >> each dive/marker now now has max. depth value. by clicking a cluster
> >> on the map at the bottom of the Qt UI it says "selected a cluster with
> >> N dives and average depth M".
> >> so basically stats are possible for a dive cluster.
> >>
> >> if anyone wants i can share the source.
> >
> > This looks neat. But it's still WebKit based, correct? So we replace Marble,
> > but we still need WebKit.
> >
> 
> yes, but it's not tightly coupled with the web page renderer.
> so it can be ported eventually to QWebEngine.
> it seems that  the only thing that will break is the JS <-> C++
> communication, as QWebEngine uses a different method TMK.

Nice work.

Using QWebEngine and the "real" maps js api would be a nice reduction of
our wonkey 3pp list.

The only downside I see is that we loose the little offline support we
have. Our current Marble map caches the tiles, so you can, if they are
cached, view the map offline.


//Anton


-- 
Anton Lundin+46702-161604
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Davide DB]

Willem take a look at this real implementatuon

http://www.diveboard.com/explore

davide@mobile

Il 07 lug 2017 5:32 PM, "Willem Ferguson" 
ha scritto:

> On 07/07/2017 17:08, Davide DB wrote:
>
>>
>> Very nice.
>> This could be a starting point for a new "map view" feature in an
>> upcoming Subsurface.
>> In my idea we should just display clusters: when user click on one
>> cluster the map zoom showing eventually other cluster until at a
>> certain zoom level he sees only dives.
>> In this Map view if I hoover on a dive spot I get a small box with
>> main dive details (just imagine what we have for a item in our mobile
>> dive list). The dive name it's a link/button which brings me to the
>> dive details. I mean that I will abandon the Map View and I swith to
>> our classic view.
>>
>> To avoid to forget this idea I copied it on Github:
>>
>> https://github.com/Subsurface-divelog/subsurface/issues/471
>>
>> Bye
>>
>> Lubomir,
>
> The cluster icons are so large and with so many colours that they actually
> seriously disrupt the map. Is there a way to remove the concentric, fading
> rings around each cluster and to make the the clusters the same colour?
> Although the present red flags on the Marble map cannot represent clusters,
> they are discreet enough not to disrupt the map (most of the time). I would
> envision a small, round, red disk that is large enough to contain any
> information that is relevant, noting more (probably no. dives in the
> cluster)? When zooming into a cluster, the individual dives eventually show
> up (perhaps with a custom symbol)??
>
> The other concern that I have is that on several of the sites with these
> maps, the update of the map was rather slow when one zooms or pans (using
> quad core CPU). But I am not aware how much of that is due to Internet
> access behind the scene and how much of it is due to slow updates in itself.
>
> Kind regards,
>
> willem
>
>
> ___
> subsurface mailing list
> subsurface@subsurface-divelog.org
> http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface
>
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Willem Ferguson]

On 07/07/2017 17:08, Davide DB wrote:


Very nice.
This could be a starting point for a new "map view" feature in an
upcoming Subsurface.
In my idea we should just display clusters: when user click on one
cluster the map zoom showing eventually other cluster until at a
certain zoom level he sees only dives.
In this Map view if I hoover on a dive spot I get a small box with
main dive details (just imagine what we have for a item in our mobile
dive list). The dive name it's a link/button which brings me to the
dive details. I mean that I will abandon the Map View and I swith to
our classic view.

To avoid to forget this idea I copied it on Github:

https://github.com/Subsurface-divelog/subsurface/issues/471

Bye


Lubomir,

The cluster icons are so large and with so many colours that they 
actually seriously disrupt the map. Is there a way to remove the 
concentric, fading rings around each cluster and to make the the 
clusters the same colour? Although the present red flags on the Marble 
map cannot represent clusters, they are discreet enough not to disrupt 
the map (most of the time). I would envision a small, round, red disk 
that is large enough to contain any information that is relevant, noting 
more (probably no. dives in the cluster)? When zooming into a cluster, 
the individual dives eventually show up (perhaps with a custom symbol)??


The other concern that I have is that on several of the sites with these 
maps, the update of the map was rather slow when one zooms or pans 
(using quad core CPU). But I am not aware how much of that is due to 
Internet access behind the scene and how much of it is due to slow 
updates in itself.


Kind regards,

willem


___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 7 July 2017 at 18:08, Davide DB  wrote:
> On 7 July 2017 at 15:38, Lubomir I. Ivanov  wrote:
>> quickly sketched the cluster idea.
>>
>> here is a preview:
>> https://www.dropbox.com/s/6gdygq68uba2pg8/cluster_clicks.png?dl=0
>>
>> each dive/marker now now has max. depth value. by clicking a cluster
>> on the map at the bottom of the Qt UI it says "selected a cluster with
>> N dives and average depth M".
>> so basically stats are possible for a dive cluster.
>>
>> if anyone wants i can share the source.
>>
>
> Very nice.
> This could be a starting point for a new "map view" feature in an
> upcoming Subsurface.
> In my idea we should just display clusters: when user click on one
> cluster the map zoom showing eventually other cluster until at a
> certain zoom level he sees only dives.

the above is already support in the demo.

> In this Map view if I hoover on a dive spot I get a small box with
> main dive details (just imagine what we have for a item in our mobile
> dive list). The dive name it's a link/button which brings me to the
> dive details. I mean that I will abandon the Map View and I swith to
> our classic view.
>

also supported. in the demo currently if you click on a marker on the
map a dive in the dive list is selected.
which means that the application receives notifications about the
selected dive from the map.

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Davide DB]

On 7 July 2017 at 15:38, Lubomir I. Ivanov  wrote:
> quickly sketched the cluster idea.
>
> here is a preview:
> https://www.dropbox.com/s/6gdygq68uba2pg8/cluster_clicks.png?dl=0
>
> each dive/marker now now has max. depth value. by clicking a cluster
> on the map at the bottom of the Qt UI it says "selected a cluster with
> N dives and average depth M".
> so basically stats are possible for a dive cluster.
>
> if anyone wants i can share the source.
>

Very nice.
This could be a starting point for a new "map view" feature in an
upcoming Subsurface.
In my idea we should just display clusters: when user click on one
cluster the map zoom showing eventually other cluster until at a
certain zoom level he sees only dives.
In this Map view if I hoover on a dive spot I get a small box with
main dive details (just imagine what we have for a item in our mobile
dive list). The dive name it's a link/button which brings me to the
dive details. I mean that I will abandon the Map View and I swith to
our classic view.

To avoid to forget this idea I copied it on Github:

https://github.com/Subsurface-divelog/subsurface/issues/471

Bye

-- 
Davide
https://vimeo.com/bocio/videos
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 7 July 2017 at 16:54, Dirk Hohndel  wrote:
>
>> On Jul 7, 2017, at 6:38 AM, Lubomir I. Ivanov  wrote:
>>
>> On 7 July 2017 at 00:48, Lubomir I. Ivanov  wrote:
>>> On 7 July 2017 at 00:26, Davide DB  wrote:
 http://i.imgur.com/o8IP4Ho.png
>>>
>>> this is possible with the so called MarkerCluster libarary:
>>> https://developers.google.com/maps/documentation/javascript/marker-clustering
>>> https://github.com/googlemaps/v3-utility-library/tree/master/markerclusterer
>>>
>>> i haven't tried it, but Google promote it and it seems easy to use.
>>>
>>
>> quickly sketched the cluster idea.
>>
>> here is a preview:
>> https://www.dropbox.com/s/6gdygq68uba2pg8/cluster_clicks.png?dl=0
>>
>> each dive/marker now now has max. depth value. by clicking a cluster
>> on the map at the bottom of the Qt UI it says "selected a cluster with
>> N dives and average depth M".
>> so basically stats are possible for a dive cluster.
>>
>> if anyone wants i can share the source.
>
> This looks neat. But it's still WebKit based, correct? So we replace Marble,
> but we still need WebKit.
>

yes, but it's not tightly coupled with the web page renderer.
so it can be ported eventually to QWebEngine.
it seems that  the only thing that will break is the JS <-> C++
communication, as QWebEngine uses a different method TMK.

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Dirk Hohndel]

> On Jul 7, 2017, at 6:38 AM, Lubomir I. Ivanov  wrote:
> 
> On 7 July 2017 at 00:48, Lubomir I. Ivanov  wrote:
>> On 7 July 2017 at 00:26, Davide DB  wrote:
>>> http://i.imgur.com/o8IP4Ho.png
>> 
>> this is possible with the so called MarkerCluster libarary:
>> https://developers.google.com/maps/documentation/javascript/marker-clustering
>> https://github.com/googlemaps/v3-utility-library/tree/master/markerclusterer
>> 
>> i haven't tried it, but Google promote it and it seems easy to use.
>> 
> 
> quickly sketched the cluster idea.
> 
> here is a preview:
> https://www.dropbox.com/s/6gdygq68uba2pg8/cluster_clicks.png?dl=0
> 
> each dive/marker now now has max. depth value. by clicking a cluster
> on the map at the bottom of the Qt UI it says "selected a cluster with
> N dives and average depth M".
> so basically stats are possible for a dive cluster.
> 
> if anyone wants i can share the source.

This looks neat. But it's still WebKit based, correct? So we replace Marble,
but we still need WebKit.

/D
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 7 July 2017 at 00:48, Lubomir I. Ivanov  wrote:
> On 7 July 2017 at 00:26, Davide DB  wrote:
>> http://i.imgur.com/o8IP4Ho.png
>
> this is possible with the so called MarkerCluster libarary:
> https://developers.google.com/maps/documentation/javascript/marker-clustering
> https://github.com/googlemaps/v3-utility-library/tree/master/markerclusterer
>
> i haven't tried it, but Google promote it and it seems easy to use.
>

quickly sketched the cluster idea.

here is a preview:
https://www.dropbox.com/s/6gdygq68uba2pg8/cluster_clicks.png?dl=0

each dive/marker now now has max. depth value. by clicking a cluster
on the map at the bottom of the Qt UI it says "selected a cluster with
N dives and average depth M".
so basically stats are possible for a dive cluster.

if anyone wants i can share the source.

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 7 July 2017 at 00:26, Davide DB  wrote:
> http://i.imgur.com/o8IP4Ho.png

this is possible with the so called MarkerCluster libarary:
https://developers.google.com/maps/documentation/javascript/marker-clustering
https://github.com/googlemaps/v3-utility-library/tree/master/markerclusterer

i haven't tried it, but Google promote it and it seems easy to use.

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Davide DB]

http://i.imgur.com/o8IP4Ho.png

On 6 July 2017 at 15:07, Lubomir I. Ivanov  wrote:
> On 6 July 2017 at 00:20, Lubomir I. Ivanov  wrote:
>>
>> i would like to point out that i've noticed something else about the
>> google maps inside this latest QWebKit.
>> there are many threads online about the issue of google map tiles not
>> updating while the user is dragging the map or while the map is
>> animating towards a new spot.
>> this seems to be the case here as well, but none of the suggested
>> fixes online seem to fix it.
>>
>> for me this doesn't seem like a big showstopper, but it's best if
>> others test the demo app too.
>> in the meantime i can try contacting the google maps devs to see if
>> they have an easy solution.
>>
>
> https://issuetracker.google.com/issues/35826647
>
> lubomir
> --
> ___
> subsurface mailing list
> subsurface@subsurface-divelog.org
> http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface



-- 
Davide
https://vimeo.com/bocio/videos
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 6 July 2017 at 00:20, Lubomir I. Ivanov  wrote:
>
> i would like to point out that i've noticed something else about the
> google maps inside this latest QWebKit.
> there are many threads online about the issue of google map tiles not
> updating while the user is dragging the map or while the map is
> animating towards a new spot.
> this seems to be the case here as well, but none of the suggested
> fixes online seem to fix it.
>
> for me this doesn't seem like a big showstopper, but it's best if
> others test the demo app too.
> in the meantime i can try contacting the google maps devs to see if
> they have an easy solution.
>

https://issuetracker.google.com/issues/35826647

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 5 July 2017 at 01:22, Lubomir I. Ivanov  wrote:
> here is an animated gif of it in action -> extract the gif from the
> zip file and drag & drop it in a browser to view it.
> https://www.dropbox.com/s/oinkj5jjvjaf3yy/WebKitGoogleMapsAnimatedGifPreview.zip?dl=0
>
> this is the source code:
> https://www.dropbox.com/s/e9fbgcbjd8zyk3s/WebKitGoogleMapsSrc.zip?dl=0
>
> to build:
> extract the source zip in ~/
> cd ~/
> qmake
> make
>
> the binary needs "libeay", "libssl", "ssleay" the same way the current
> webkit in subsurface needs them.
>
> random notes:

i would like to point out that i've noticed something else about the
google maps inside this latest QWebKit.
there are many threads online about the issue of google map tiles not
updating while the user is dragging the map or while the map is
animating towards a new spot.
this seems to be the case here as well, but none of the suggested
fixes online seem to fix it.

for me this doesn't seem like a big showstopper, but it's best if
others test the demo app too.
in the meantime i can try contacting the google maps devs to see if
they have an easy solution.

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 5 July 2017 at 01:53, Davide DB  wrote:
> On 5 July 2017 at 00:22, Lubomir I. Ivanov  wrote:
>>
>> here is an animated gif of it in action -> extract the gif from the
>> zip file and drag & drop it in a browser to view it.
>> https://www.dropbox.com/s/oinkj5jjvjaf3yy/WebKitGoogleMapsAnimatedGifPreview.zip?dl=0
>
>
> Lubomir, do you pretend I dive into that river? :)
>

doesn't look like the best diving spot. i agree. :)

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Davide DB]

On 5 July 2017 at 00:22, Lubomir I. Ivanov  wrote:
>
> here is an animated gif of it in action -> extract the gif from the
> zip file and drag & drop it in a browser to view it.
> https://www.dropbox.com/s/oinkj5jjvjaf3yy/WebKitGoogleMapsAnimatedGifPreview.zip?dl=0


Lubomir, do you pretend I dive into that river? :)

PS
Its' time to refresh an old idea I proposed years ago...

-- 
Davide
https://vimeo.com/bocio/videos
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 4 July 2017 at 20:11, Lubomir I. Ivanov  wrote:
> On 4 July 2017 at 19:26, Dirk Hohndel  wrote:
>>
>>
>> I like Lubomir's idea of just using a web browser pointing at an 
>> appropriately
>> customized web page. Let's see where this takes us. We only use Marble on
>> the desktop, so the restrictions with showing web pages don't bother us.
>>
>
> other than that the custom POC is almost done. i need like 2-3 hours
> more, so it's going to be finished later today or tomorrow.
>

here is an animated gif of it in action -> extract the gif from the
zip file and drag & drop it in a browser to view it.
https://www.dropbox.com/s/oinkj5jjvjaf3yy/WebKitGoogleMapsAnimatedGifPreview.zip?dl=0

this is the source code:
https://www.dropbox.com/s/e9fbgcbjd8zyk3s/WebKitGoogleMapsSrc.zip?dl=0

to build:
extract the source zip in ~/
cd ~/
qmake
make

the binary needs "libeay", "libssl", "ssleay" the same way the current
webkit in subsurface needs them.

random notes:
- there seems to be some delay until the map loads for the first time,
i think we can listen for that
- also the first markers are not added immidiatelly (maybe it's my internet)
- better animations can be added when clicking on markers, so that the
map zooms and pans smoothly. the built-in animation isn't great for
that.
- the green tick images are just downloaded from the internet. a
marker image as Qt resource should be used instead
- the info boxes contents can be customized
- if someone has taken photos near a dive site - even the street view
thing works (by dragging the street view "human figure" from the
right).
- this example is running using a public API key that i've created,
this probably needs to be changed to another google account.
- localization is supported - e.g. passing "en_US"

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 4 July 2017 at 19:26, Dirk Hohndel  wrote:
>
>
> I like Lubomir's idea of just using a web browser pointing at an appropriately
> customized web page. Let's see where this takes us. We only use Marble on
> the desktop, so the restrictions with showing web pages don't bother us.
>

ultimatelly if QtLocation supports satellite images it might be the
better solution.
other than that the custom POC is almost done. i need like 2-3 hours
more, so it's going to be finished later today or tomorrow.

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Dirk Hohndel]

> On Jul 4, 2017, at 9:19 AM, Thiago Macieira  wrote:
>> I was told on the Qt list that there is a MUCH newer branch that I should
>> try - they have updated to a current WebKit. The thing that's odd... between
>> 5.7.1 (which the Windows binaries used to be based on) and 5.9 there
>> shouldn't really have been any WebKit changes... so why does it crash now.
> 
> There are a number of minor changes submitted to the 5.8 and 5.9 branches by 
> people like Konstantin, Allan and me. Mine were mostly just to make it, but 
> looking at the diff there appear to be a couple of backports from upstream.

OK. For now I have given up on upgrading the libraries in the Windows
binary. BLE isn't supported, anyway, so whatever. I'll stick with 5.7.1 as
that creates working builds for me.

I'll revisit this when I have an excess of time on my hands :-)

>>> i mean i can possibly find the offending line of code, but i won't
>>> really have a good fix for it. after seeing that unpredictable Marble
>>> multi-threading crash on Windows, i think we are entering the
>>> territory of big, complicated but unmaintained libraries that simply
>>> may not run with the latest Qt version. Qt 6.0 is coming too...
>> 
>> I think they are working on 5.10 ...
> 
> Qt 6.0 will be discussed at the Contributor Summit this year in Berlin. The 
> current idea is that 5.12 will be LTS and then 6.0. That would put 6.0 about 
> two years away from now.

Interesting. Will 6.0 once again be incompatible as the major version bump
implies? What will change?

>>> so we need to:
>>> 1) move away from these unmaintained libraries
>>> WebKit - currently we are stuck with WebKit on Windows
>>> Grantlee - we are stuck with Grantlee due to the template syntax which
>>> the users are already using
>>> Marble - we probably should start using a 2D map of sorts
>> 
>> WebKit - I think we are stuck with that for a while
>> Marble - Tomaz has this working with the latest, maintained Marble
>> Grantee - I have this working on Mac
> 
> What happened to the idea of using QtLocation for maps?

We talked about this here a few times.
We absolutely have to have satellite images. Maps are useless for a dive log.
And QtLocation appears to only support maps. So no go.

I like Lubomir's idea of just using a web browser pointing at an appropriately
customized web page. Let's see where this takes us. We only use Marble on
the desktop, so the restrictions with showing web pages don't bother us.

>>> BTW you can try building this QtWebKit port as it's actively updated:
>>> https://github.com/annulen/webkit/releases
>> 
>> Yes, and that appears to be in the 5.212 branch of Qt that I need to
>> find the time to investigate...
> 
> 5.212 builds with CMake, not qmake anymore. My recommendation is that you 
> build it after all the rest of Qt.

Hmm. I still haven't looked at this. Too many parts of the code that are in
motion, too many platforms, too little help, too little time...

So 5.212 is just QWebKit that I build against an existing Qt build?
Of course, I will still have to build from source for Linux and Android, it 
seems,
as in both cases 5.9.1 is missing BLE patches from Alex that we need...

/D
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Linus Torvalds]

On Tue, Jul 4, 2017 at 9:19 AM, Thiago Macieira  wrote:
>
> What happened to the idea of using QtLocation for maps?

As noted elsewhere, we require high-quality satellite imagery for dive
site location - seeing shallow reefs under water world-wide and things
like smallish coastal features. That effectively means google maps
data.

I see maptype::SatelliteMapDay, but no actual documentation or hints
about how well it actually works (googling for images I didn't find
anything that looked promising)

   Linus
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Thiago Macieira]

On segunda-feira, 3 de julho de 2017 14:32:35 PDT Dirk Hohndel wrote:
> Grumble.
> 
> I was told on the Qt list that there is a MUCH newer branch that I should
> try - they have updated to a current WebKit. The thing that's odd... between
> 5.7.1 (which the Windows binaries used to be based on) and 5.9 there
> shouldn't really have been any WebKit changes... so why does it crash now.

There are a number of minor changes submitted to the 5.8 and 5.9 branches by 
people like Konstantin, Allan and me. Mine were mostly just to make it, but 
looking at the diff there appear to be a couple of backports from upstream.

> > i mean i can possibly find the offending line of code, but i won't
> > really have a good fix for it. after seeing that unpredictable Marble
> > multi-threading crash on Windows, i think we are entering the
> > territory of big, complicated but unmaintained libraries that simply
> > may not run with the latest Qt version. Qt 6.0 is coming too...
> 
> I think they are working on 5.10 ...

Qt 6.0 will be discussed at the Contributor Summit this year in Berlin. The 
current idea is that 5.12 will be LTS and then 6.0. That would put 6.0 about 
two years away from now.

> > so we need to:
> > 1) move away from these unmaintained libraries
> > WebKit - currently we are stuck with WebKit on Windows
> > Grantlee - we are stuck with Grantlee due to the template syntax which
> > the users are already using
> > Marble - we probably should start using a 2D map of sorts
> 
> WebKit - I think we are stuck with that for a while
> Marble - Tomaz has this working with the latest, maintained Marble
> Grantee - I have this working on Mac

What happened to the idea of using QtLocation for maps?

> > BTW you can try building this QtWebKit port as it's actively updated:
> > https://github.com/annulen/webkit/releases
> 
> Yes, and that appears to be in the 5.212 branch of Qt that I need to
> find the time to investigate...

5.212 builds with CMake, not qmake anymore. My recommendation is that you 
build it after all the rest of Qt.

-- 
Thiago Macieira - thiago (AT) macieira.info - thiago (AT) kde.org
   Software Architect - Intel Open Source Technology Center

___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 4 July 2017 at 00:55, Dirk Hohndel  wrote:
>
>> On Jul 3, 2017, at 2:49 PM, Lubomir I. Ivanov  wrote:
>>
>> On 4 July 2017 at 00:37, Dirk Hohndel  wrote:
>>>
 On Jul 3, 2017, at 2:34 PM, Linus Torvalds  
 wrote:

 On Mon, Jul 3, 2017 at 2:21 PM, Lubomir I. Ivanov  
 wrote:
>
> Marble - we probably should start using a 2D map of sorts

 Note that it's not just "2D map of sorts".

 It needs to be something with good satellite imagery. Which presumably
 means "google maps" in practice.

 I absolutely depend on seeing the actual reefs and shoreline features
 when I check (and occasionally fix) the location. At some point we
 tried some other map thing (I think the marine version of
 openstreetmaps) and it may have been acceptable for boating navigation
 needs, but certainly not for placing dive dites.
>>>
>>> Correct. That's the main reason I'm loath to move away from Marble.
>>> QtLocation is cute, but has maps, not satellite imagery. Not useful.
>>> Same goes for openstreetmaps based stuff. It's got to be Bing or Google
>>> satellite images.
>>>
>>
>> i suggested this mid-june:
>> http://lists.subsurface-divelog.org/pipermail/subsurface/2017-June/028642.html
>>
>> simply a web page with google maps and some javascript that can be
>> called from c++.
>>
>> but instead of WebEngine use WebKit (for now)
>>
>
> I could have sworn that I responded... but can't find it in the archive.
>
> The challenge is "how do we place the dive site markers on that web page"?
>
> We can't use this on mobile (but we don't have Marble there, anyway, so no
> loss). But if you can work around that and have a POC, I'd love that.
>

ok, i can try creating a standalone POC project for that.

lubomit
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Dirk Hohndel]

> On Jul 3, 2017, at 2:49 PM, Lubomir I. Ivanov  wrote:
> 
> On 4 July 2017 at 00:37, Dirk Hohndel  wrote:
>> 
>>> On Jul 3, 2017, at 2:34 PM, Linus Torvalds  
>>> wrote:
>>> 
>>> On Mon, Jul 3, 2017 at 2:21 PM, Lubomir I. Ivanov  
>>> wrote:
 
 Marble - we probably should start using a 2D map of sorts
>>> 
>>> Note that it's not just "2D map of sorts".
>>> 
>>> It needs to be something with good satellite imagery. Which presumably
>>> means "google maps" in practice.
>>> 
>>> I absolutely depend on seeing the actual reefs and shoreline features
>>> when I check (and occasionally fix) the location. At some point we
>>> tried some other map thing (I think the marine version of
>>> openstreetmaps) and it may have been acceptable for boating navigation
>>> needs, but certainly not for placing dive dites.
>> 
>> Correct. That's the main reason I'm loath to move away from Marble.
>> QtLocation is cute, but has maps, not satellite imagery. Not useful.
>> Same goes for openstreetmaps based stuff. It's got to be Bing or Google
>> satellite images.
>> 
> 
> i suggested this mid-june:
> http://lists.subsurface-divelog.org/pipermail/subsurface/2017-June/028642.html
> 
> simply a web page with google maps and some javascript that can be
> called from c++.
> 
> but instead of WebEngine use WebKit (for now)
> 

I could have sworn that I responded... but can't find it in the archive.

The challenge is "how do we place the dive site markers on that web page"?

We can't use this on mobile (but we don't have Marble there, anyway, so no
loss). But if you can work around that and have a POC, I'd love that.

/D

___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 4 July 2017 at 00:37, Dirk Hohndel  wrote:
>
>> On Jul 3, 2017, at 2:34 PM, Linus Torvalds  
>> wrote:
>>
>> On Mon, Jul 3, 2017 at 2:21 PM, Lubomir I. Ivanov  
>> wrote:
>>>
>>> Marble - we probably should start using a 2D map of sorts
>>
>> Note that it's not just "2D map of sorts".
>>
>> It needs to be something with good satellite imagery. Which presumably
>> means "google maps" in practice.
>>
>> I absolutely depend on seeing the actual reefs and shoreline features
>> when I check (and occasionally fix) the location. At some point we
>> tried some other map thing (I think the marine version of
>> openstreetmaps) and it may have been acceptable for boating navigation
>> needs, but certainly not for placing dive dites.
>
> Correct. That's the main reason I'm loath to move away from Marble.
> QtLocation is cute, but has maps, not satellite imagery. Not useful.
> Same goes for openstreetmaps based stuff. It's got to be Bing or Google
> satellite images.
>

i suggested this mid-june:
http://lists.subsurface-divelog.org/pipermail/subsurface/2017-June/028642.html

simply a web page with google maps and some javascript that can be
called from c++.

but instead of WebEngine use WebKit (for now)

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Dirk Hohndel]

> On Jul 3, 2017, at 2:34 PM, Linus Torvalds  
> wrote:
> 
> On Mon, Jul 3, 2017 at 2:21 PM, Lubomir I. Ivanov  wrote:
>> 
>> Marble - we probably should start using a 2D map of sorts
> 
> Note that it's not just "2D map of sorts".
> 
> It needs to be something with good satellite imagery. Which presumably
> means "google maps" in practice.
> 
> I absolutely depend on seeing the actual reefs and shoreline features
> when I check (and occasionally fix) the location. At some point we
> tried some other map thing (I think the marine version of
> openstreetmaps) and it may have been acceptable for boating navigation
> needs, but certainly not for placing dive dites.

Correct. That's the main reason I'm loath to move away from Marble.
QtLocation is cute, but has maps, not satellite imagery. Not useful.
Same goes for openstreetmaps based stuff. It's got to be Bing or Google
satellite images.

/D
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Linus Torvalds]

On Mon, Jul 3, 2017 at 2:21 PM, Lubomir I. Ivanov  wrote:
>
> Marble - we probably should start using a 2D map of sorts

Note that it's not just "2D map of sorts".

It needs to be something with good satellite imagery. Which presumably
means "google maps" in practice.

I absolutely depend on seeing the actual reefs and shoreline features
when I check (and occasionally fix) the location. At some point we
tried some other map thing (I think the marine version of
openstreetmaps) and it may have been acceptable for boating navigation
needs, but certainly not for placing dive dites.

Linus
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Dirk Hohndel]

> On Jul 3, 2017, at 2:21 PM, Lubomir I. Ivanov  wrote:
> 
> On 3 July 2017 at 18:28, Dirk Hohndel  wrote:
>> Lubomir,
>> 
>> I'm trying to do too many things at once... I updated the build tools and
>> most libraries for the Windows builds to current versions.
>> 
>> http://subsurface-divelog.org/downloads/test/subsurface-4.6.4-310-g6451adfec112.exe
>> 
>> Crashes for me at startup in QtCore - before it gets any useful information
>> out. The Event Log isn't useful, either.
>> 
>> Could you take a look?
>> 
> 
> i'm checking this source, but not sure if it's the same on your end:
> http://download.qt.io/community_releases/5.9/5.9.0-final/
> 
> it crashes after the second call to WebCore::initializeWebCoreQt().
> (Source/WebKit/qt/WebCoreSupport/InitWebCoreQt.cpp)
> the first call succeeds and the second call should be a NOP (return if
> the method was already called) and that's what happens.
> 
> a prime suspect for me is something which creates a QWebPageAdapter
> class instance (it has a call to WebCore::initializeWebCoreQt() in the
> constructor) and then it tries to QObject::connect() something related
> to ConnectionType and crashes.
> QWebPageAdapter is a class which is used *a lot*, so hard to tell what
> causes that.

Grumble.

I was told on the Qt list that there is a MUCH newer branch that I should
try - they have updated to a current WebKit. The thing that's odd... between
5.7.1 (which the Windows binaries used to be based on) and 5.9 there shouldn't
really have been any WebKit changes... so why does it crash now.

:-(

> i mean i can possibly find the offending line of code, but i won't
> really have a good fix for it. after seeing that unpredictable Marble
> multi-threading crash on Windows, i think we are entering the
> territory of big, complicated but unmaintained libraries that simply
> may not run with the latest Qt version. Qt 6.0 is coming too...

I think they are working on 5.10 ...


> i also saw that Grantlee report in the OSX thread.
> 
> so we need to:
> 1) move away from these unmaintained libraries
> WebKit - currently we are stuck with WebKit on Windows
> Grantlee - we are stuck with Grantlee due to the template syntax which
> the users are already using
> Marble - we probably should start using a 2D map of sorts

WebKit - I think we are stuck with that for a while
Marble - Tomaz has this working with the latest, maintained Marble
Grantee - I have this working on Mac

> 2) maintain them our self
> we already do that for Marble partially, but its hard as these are
> complicated projects.
> QtWebKit in particular is a pretty big project.

Massively insanely huge. But it actually is maintained. Just not
as well as one might hope :-/

> BTW you can try building this QtWebKit port as it's actively updated:
> https://github.com/annulen/webkit/releases

Yes, and that appears to be in the 5.212 branch of Qt that I need to
find the time to investigate...

> 3) call the library developers to test and patch against the latest Qt
> NOTE: i talked to annulen (maintainer from the above link) and i don't
> think he even tests his Windows binaries...release works, debug
> doesn't.
> 
> 4) stick with an older Qt that works
> not great as we miss new features in new Qt versions.

For Windows that's actually an option. BLE isn't supported, anyway,
and that's the main reason we really want 5.9+ on the other platforms

> i wish that Qt didn't break older libraries and everything just worked... :\

Me too. And I want a pony.

/D
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 3 July 2017 at 18:28, Dirk Hohndel  wrote:
> Lubomir,
>
> I'm trying to do too many things at once... I updated the build tools and
> most libraries for the Windows builds to current versions.
>
> http://subsurface-divelog.org/downloads/test/subsurface-4.6.4-310-g6451adfec112.exe
>
> Crashes for me at startup in QtCore - before it gets any useful information
> out. The Event Log isn't useful, either.
>
> Could you take a look?
>

i'm checking this source, but not sure if it's the same on your end:
http://download.qt.io/community_releases/5.9/5.9.0-final/

it crashes after the second call to WebCore::initializeWebCoreQt().
(Source/WebKit/qt/WebCoreSupport/InitWebCoreQt.cpp)
the first call succeeds and the second call should be a NOP (return if
the method was already called) and that's what happens.

a prime suspect for me is something which creates a QWebPageAdapter
class instance (it has a call to WebCore::initializeWebCoreQt() in the
constructor) and then it tries to QObject::connect() something related
to ConnectionType and crashes.
QWebPageAdapter is a class which is used *a lot*, so hard to tell what
causes that.

i mean i can possibly find the offending line of code, but i won't
really have a good fix for it. after seeing that unpredictable Marble
multi-threading crash on Windows, i think we are entering the
territory of big, complicated but unmaintained libraries that simply
may not run with the latest Qt version. Qt 6.0 is coming too...

i also saw that Grantlee report in the OSX thread.

so we need to:
1) move away from these unmaintained libraries
WebKit - currently we are stuck with WebKit on Windows
Grantlee - we are stuck with Grantlee due to the template syntax which
the users are already using
Marble - we probably should start using a 2D map of sorts

2) maintain them our self
we already do that for Marble partially, but its hard as these are
complicated projects.
QtWebKit in particular is a pretty big project.

BTW you can try building this QtWebKit port as it's actively updated:
https://github.com/annulen/webkit/releases

3) call the library developers to test and patch against the latest Qt
NOTE: i talked to annulen (maintainer from the above link) and i don't
think he even tests his Windows binaries...release works, debug
doesn't.

4) stick with an older Qt that works
not great as we miss new features in new Qt versions.

i wish that Qt didn't break older libraries and everything just worked... :\

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: latest Windows binary crashes

[Lubomir I. Ivanov]

On 3 July 2017 at 18:28, Dirk Hohndel  wrote:
> Lubomir,
>
> I'm trying to do too many things at once... I updated the build tools and
> most libraries for the Windows builds to current versions.
>
> http://subsurface-divelog.org/downloads/test/subsurface-4.6.4-310-g6451adfec112.exe
>
> Crashes for me at startup in QtCore - before it gets any useful information
> out. The Event Log isn't useful, either.
>
> Could you take a look?
>

ok, i can take a look in about an hour.

lubomir
--
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

latest Windows binary crashes

[Dirk Hohndel]

Lubomir,

I'm trying to do too many things at once... I updated the build tools and most 
libraries for the Windows builds to current versions.

http://subsurface-divelog.org/downloads/test/subsurface-4.6.4-310-g6451adfec112.exe
 


Crashes for me at startup in QtCore - before it gets any useful information 
out. The Event Log isn't useful, either.

Could you take a look?

Thanks

/D___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface