Re: [linux-minidisc] qhimdtransfer windows autoupdate feature

[John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>]

On 01/23/2013 05:10 PM, Thomas Arp wrote:
> Maybe we could add the WinSparkle.def definition file to our sources and than use libtool from a system call to create the lib.

Yes, that would be acceptable.

> Alternativly we can offer a download link (at CompilingWithMingw/precompiled binaries) for the libwinsparkle.a file
> (and, of course, the WinSparkle binaries and headers). We are doing this for all other precompiled binaries for windows.

No, this will never happen. Having to rely on external binaries means 
the sources cannot be compiled from the original source tarball which 
makes them unfit for any distribution willing to package the software.

> > > I also created a sample appcast rss feed winsparkle checks for new
> > > versions on my web server [2].
> >
> > Looks very simple, didn't know that. Cool.
>
> Yes, it is very simple, for sparkle/winsparkle the "enclosure" section in an item is the only thing we need.
> The rest is just for compatibility reasons to create a valid rss feed.

Ok.

> > > QHiMDUpdate base class can be used for mac autoupdate feature using
> > > sparkle, too.
> >
> > Even better.
>
> The QHiMDUpdate base class is just a dummy class doing nothing.
> WinSparkle uses most of the Sparkle member names, so the subclass for mac shouldn´t be much different to this for windows.
> It is important to reimplement the public virtual members to leave the calling routines from the mainwindow platform independent.

Understood.

> > I think this is something we should merge next - before the
> > NetMD integration - as the latter involves much more changes and needs
> > way more review.
> >
> > You should also split up your NetMD changes into more, smaller patches.
> > Reviewing a 100k patch isn't really feasible.
> O.K., if we decide to merge the autoupdate mechanism next i have to recreate the netmd integration patch. I will split it than.

In any case, the autoupdate stuff must not break any existing things. 
Most importantly, it must not break the possibility to compile the 
binaries from the original source tarball.

Adrian

--
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer - glaubitz@debian.org
`. `'   Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913