e.g
> $ dig +short ssl.example.com
> 123.456.789.012
> $ curl -I https://123.456.789.012/foo.bar
> curl: (51) SSL: certificate subject name 'ssl.example.com' > does not match target host name '123.456.789.012'
Besides that, he makes several good points which I think should speak for themselves.
It also concerns me that the author of security conscious software isn't concerned by this.
That said, I do consider this to be a temporary measure while the software is maturing to the point that I can get it accepted into the mainstream distros and build proper win/mac packages. That would be better for everyone, both from a usability and security standpoint.
We are a very small team (usually just me) and these things take time.
https://pagekite.net/wiki/Howto/GNULinux/DebianPackage/ https://pagekite.net/wiki/Howto/GNULinux/RpmPackage/
... this works and is the most conveninient way to use PageKite if you are treating it as part of the OS (so always on, exposing local servers).
However, in the context of security and trust, you really should think thrice before installing 3rd party repos like mine - once you have added my repo to your OS, I can upgrade any package on your computer with whatever malicious content I want. So folks who don't trust me and my curl-bash hack really shouldn't trust this sort of solution either.
This is why getting third parties to build, package and sign my software and getting into the public repos is the "correct" solution, but doing so didn't make sense while I was rapidly iterating. Things are starting to stabalize now, so I will probably be revisiting this very soon.
It looks scary, but it really isn't. Or rather, it's no scarier than most of the other ways to add software to the common desktop operating systems.
Edit: and for those who care, there are .rpm and .deb packages available, or you can grab the source from github. Pick your poison.