Subsurface ApplImage
Dirk Hohndel
dirk at hohndel.org
Thu Nov 5 21:03:25 PST 2015
Probono - question for you almost at the end...
On Thu, Nov 05, 2015 at 08:48:32PM +0200, Willem Ferguson wrote:
> Here are some comments, running the ApplImage in a terminal. Tested
> following actions:
Thanks for testing this, Willem!
> I suspect several of issues below are known conditions.
>
> Dive import from Galileo Sol. Apparently ok.
> ISSUE: Dive import from Shearwater: Checkbox for Bluetooth is not shown in
> the import panel, so cannot import via Bluetooth.
So this seems to imply that I built without BT (and I'm fairly sure I
didn't) or that the BT initialization fails. Was there anything useful on
the console? Have you tried running with '-v' (I can't remember right now
if we give more detailed debug messages with -v...)
> Adding notes for imported dives: ok
>
> Adding locality for imported dives through notes panel: ok
> HOWEVER, terminal message:
> qt.network.ssl: QSslSocket: cannot resolve SSLv2_client_method
> qt.network.ssl: QSslSocket: cannot resolve SSLv2_server_method
Those are normal. Qt tries to dynamically load some functions and can't
find them. I find this a bit disconcerting but it appears to do that and
still appears to work. I get these on Mac and ArchLinux with native builds
as well.
> Adding equipment information:
> Add cylinders. ok
> Add weights. ok
>
> Info tab appears reasonable.
> Stats tab appears reasonable.
> Photos tab ok.
> Extra data tab not tested.
>
> Dive planner:
> ISSUE: Dive planner: Small issue: when Gradient factors are changed, the
> profile immediately adapts but the texy at the top op the profile is not
> updated.
> For instance if one changes: 45/55 to 45/75. then the text
> remains at 45/55.
Does the same thing work with the natively built binary?
> As far as I can see rest of planner ok.
> ISSUE: When returning from planner to dive log, the header for the dive
> planner is not replaced.
> For instance, if last dive plan was for VPM +4, changing back
> to dive notes keeps VPM+4 on top of profile.
I think that's a bug in current master, independent of the AppImage.
> Preferences panel:
> ISSUE: When changing gradient factors in Preferences panel, GF text for dive
> profile in log is not updated. No message in terminal.
> ISSUE: Facebook Preferences tab is gone.
Both are known issues with master and will show up in a native build as
well.
> Dive list manipulations:
> Tested all actions in context menu, including importing images from file. No
> problems.
> Did not test load images from web.
>
> FILE dropdown menu
> Save as -> HTML, see following warnings:
> copy of "/media/willem/TOSHIBA EXT/Fotos/Skotland.2015 Okt/Oban
> diving/darktable_exported/none_01.jpg" to
> "/home/willem/Desktop/obantest.html_files//photos/none_01.jpg" failed
> copy of "/media/willem/TOSHIBA EXT/Fotos/Skotland.2015 Okt/Oban
> diving/darktable_exported/none.jpg" to
> "/home/willem/Desktop/obantest.html_files//photos/none.jpg" failed
> BUT: images appear to be exported correctly and HTML appears correct.
That's strange. So the images show up in exactly those paths, but we get
these warnings?
> Save to Cloud:
> ISSUE: Appears to happen (showing progress bar) but Subsurface hangs with
> very high CPU usage imediately afterwards. Have to use OS to force quit.
If you can reproduce this (I tried here and as usual this works fairly
quickly and without a problem for me), could you attach with a debugger
and figure out what Subsurface is doing?
We definitely need to spend some time profiling the load/save from/to
cloud code. There clearly is something odd going on and it doesn't seem to
happen with reasonably fast connections. I know I had some problems on
Bonaire and Palau, but there things simply timed out...
> VIEW dropdown menu: All ok. Did not test next dc/previous dc.
>
> Load images from file
> ISSUE: When loading images, a double-click on a thumbnail does not open a
> full-size view of the image. Terminal messages:
> QXcbWindow: Unhandled client message: "_GTK_LOAD_ICONTHEMES"
> QXcbWindow: Unhandled client message: "_GTK_LOAD_ICONTHEMES"
> (Subsurface_4.5.1.78_x86_64.AppImage:6974): GLib-GIO-CRITICAL **:
> g_dbus_connection_register_object: assertion 'G_IS_DBUS_CONNECTION
> (connection)' failed
> (Subsurface_4.5.1.78_x86_64.AppImage:6974): GLib-GIO-CRITICAL **:
> g_dbus_connection_register_object: assertion 'G_IS_DBUS_CONNECTION
> (connection)' failed
Probono - we use Qt to open the files. It looks like that message doesn't
get passed on to the OS correctly. Any idea?
Willem - a HUGE thank you for that detailed test. This was really really
helpful.
/D
More information about the subsurface
mailing list