[openlp-dev] Trouble with pyinstaller/ Jenkins account enable request.

[Chris Reed]

For reference, it works when I downgraded from pyinstaller 3.3 to 3.1.1 
- though I also needed a pyinstaller hook for 
sqlalchemy.sql.default_comparator.  Not sure where to stick that, so 
I've just put it in hook-openlp.py for now.

Now to work out why it's crashing when I shut down OpenLP.


On 19/10/2016 13:59, Chris Reed wrote:
> Presumably if I am able to start a job in Jenkins, that'll build a 
> .exe for me, which is a way of getting changes into a testable 
> state..?  Could Raoul/ someone enable my account in Jenkins - cr212 - 
> to do builds?  Thanks.
>
> FYI, I'm working on a patch ( 
> http://bazaar.launchpad.net/~cr212/openlp/pptpresenterview/revision/2700 
> ) for adding a configuration option to use the powerpoint presenter 
> view.  Rationale: we want auto-advancing slides for our notices, and 
> powerpoint won't advance slides if it doesn't have focus.  By putting 
> the presenter view up, it avoids operator error by hiding the OpenLP 
> window with the presenter view, so less danger of losing focus from ppt.
>
> It works for me if I'm running from the checkout, but I'm struggling 
> to build a pyinstaller executable to test in an existing 
> installation.  I've checked out the packaging branch, copied 
> windows\config.ini.default to config.ini, and run:
>
> C:\Users\Chris\dev\packaging\windows>\Python34\python.exe 
> windows-builder.py -u -b c:\\Users\\Chris\\dev\openlp
>
> This produces an OpenLP.exe in the dist output, but running it seems 
> to not do anything, nor does putting it into a portable installation - 
> no splash screen and no OpenLP starts up.  Running ProcessMonitor 
> shows it is running and scanning through the plugins, but no sign of a 
> window.  Any pointers?
>
> Also, I've looked at the existing tests for the powerpoint controller, 
> but there doesn't seem to be any practical way of adding a test for 
> this patch - the bulk of it is making sure the right window has focus, 
> and it looks like the existing tests don't even run against a real 
> powerpoint.  Is just manual testing of such a change acceptable?
>
> Many thanks.
>