Testing on Windows 10

[Dirk Hohndel]

On Sun, Sep 27, 2015 at 06:26:10PM +1000, Rick Walsh wrote:
> Hi,
> 
> I thought I might as well start a new thread to report on testing on
> Windows 10, as the current one dealt mostly with building and creating a
> working installer.  That works - here's my summary of actually testing the
> latest windows build (subsurface-4.4.97-81-gc4eb9571bea0).  My system is an
> HP laptop running Windows 10, core i7 processor, 1366 x 768 display.

Thanks - this is much appreciated. I have a similar laptop now (but didn't
take it with me on my trip - one laptop is enough to lug around the world).

> What works:
> Installation

Well, you get the scary warning. I need to plop down another bunch of
money to get a signing certificate so I can finally sign our Windows
binaries, just like I sign our Mac binaries. Open source is such a
rewarding (and expensive) hobby... :-)

> Native Bluetooth download with Shearwater Petrel 2
> Facebook logging in and posting this morning's dive
> Friends commenting that the software used to post the dive profile looks
> impressive :)

Awesome!

> Create dive site, manually entering coordinates
> Create dive site, using Marble to pick location

Good.

> Save to cloud (same computer, xml log file, and network connection that I
> can't get to work for save to cloud when I boot Fedora)
> Open from cloud
> Make change, save to cloud again
> Plan a dive for Friday
> Save planned dive

That's a good list. I like your report so far...

> Problems:
> Bluetooth download the old way (create COM5 device, etc.) failed, and
> Subsurface crashed.  Terminal output:
> C:\Program Files (x86)\Subsurface>
> Map theme file does not exist: ""
> Found 126 GeoNode object LEAKS!
> No GeoNode object leak!
> Unable to match dive 'program.divelog' (subsurface)
> Unable to match dive 'version.divelog' (3)
> Unable to match autogroup 'model.divecomputerid.' (Shearwater Petrel)
> Unable to match autogroup 'deviceid.divecomputerid.' (dc701c49)
> Unable to match autogroup 'serial.divecomputerid.d.' (390e6bd0)
> Unable to match autogroup 'firmware.divecomputerid.' (29)
> Found 99 GeoNode object LEAKS!
> Set the current dive site: 0
> No GeoNode object leak!
> [5.130899] ERROR: The semaphore timeout period has expired (121) [in
> /data/win/libdivecomputer/src/serial_win32.c:152 (serial_open)]
> [5.131164] ERROR: Failed to open the serial port. [in
> /data/win/libdivecomputer/src/custom_serial.c:69 (dc_serial_native_open)]

So IIRC Windows 10 has a bug in the BT serial emulation code. I know that
we even mention that in the ReleaseNotes. I'm rather surprised that this
would lead to a crash. Can you look at the eventmgr and try to get what
miniscule information Windows provides us in a crash? Can you reproduce
the problem?

> Printing (to pdf) - choosing print selected option with 10 dives selected,
> the profile for the first dive was repeated for every selected dive.  The
> notes and data for the first two printed dives were correct, but completely
> empty for the rest.

Oh wow, that's a new one. But hold on. 10 dives? I don't think we have
such a standard template. Is this a template you did yourself?

I just did a quick check here with the six dive template and things seem
about right... Can you tell us more about this?

> Swedish translation - strings not translated:
> In menus:
> Save to cloud storage
> Find moved images
> Under
> Redo
> Edit dive
> Edit dive in planner
> Edit device names
> Share on
> Save, Discard, Cancel buttons (new log from file menu)
>
> Preferences not translated:
> No default file, Local default file, Cloud storage default file
> Unit system
> Location Display  - note this string uses incorrect capitalization (Display
> shouldn't be capitalized)
> traditional (dms) - capitalization in English
> Standard CCR setpoint för dykplannering (half translated)
> UI language
> Facebook - log in to use your Facebook account with Subsurface (or is this
> page set by system default?)
> Georeference (all strings on this page)
> Discard, Apply (buttons)
> 
> Choose Bluetooth download mode (dive computer download dialog)
> 
> Planner window:
> based on Buhlmann ZHL-16B with (Planner notes)
> based on VPM-B at nominal conservatism (Planner notes)
> Save, Cancel (buttons)
> 
> I can usually read Swedish without too much trouble, writing, speaking and
> listening it is more than clear I am not a native speaker.  I am not the
> right person to do the translations.

I haven't pulled the latest translations in a couple of days and yes, I
think most of the languages still are missing a few strings. What's more
important is to make sure that all of these strings are marked for
translation. Can you check that in the sources?

We have several Swedish speakers on the team who I'm hopeful will be
willing to deal with the translations once we know all the strings are on
transifex.

Thanks again, Rick. Much appreciated.

/D