Re: [Denemo-devel] New mingw branch

[Richard Shann]

On Fri, 2013-03-01 at 14:43 +0000, Richard Shann wrote:
> I have some (good?) news, one of the xp machines I tested rc10 on was
> just now executing it perfectly: I think this machine exhibited the
> "lost fonts" bug when I last tested it.
> The "lost fonts" bug seems to come and go at random. When I had it a
> while back on the vista laptop it affected a very ancient denemo
> 0.8.22
> or some such that I keep available. That is, it is in the environment,
> not the program.
> 
> So rc10 may be working on more windows systems than I thought. But it
> does seem to be capable of rendering windows unstable in some way. 

I think this losing fonts may be due to the installer doing something
wrong writing the font keys in the windows registry - that would explain
why it affected an already installed program. Why it spontaneously
righted itself is down to some periodic housekeeping that windows does,
I guess. As these tests are done on machines which have had Denemo
installed and crudely deleted multiple times, the real-world experience
of folk may be different.
If I can just get to build the installer, I could do some tests on the
nsis spec - at the moment I am stuck with the denemo configure stage not
finding portmidi... it seems to be not in the gub tree, yet is listed as
a dependency in denemo.py

Richard