[argyllcms] Re: Argyllcms 1.0.1 packaged in fedora-devel

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Mon, 28 Jul 2008 03:26:37 +1000

Frederic Crozat wrote:

When I read this, I have the feeling you don't want your software being easily
available to millions of Linux users out there. Too bad :(

I don't see any point in the interested users getting something that
doesn't work. It creates disappointment and aggravation for all concerned.

I'm not interested in Windows (or MacOS X), since we don't have any
control on the "platform".

No, but I am, since I support a wider audience. You at least need to
respect this.

Sorry, but my goal is to create a "clean" platform, where this kind of
hacks are not needed. If you were distributing a patch of your change
on libusb in your tarball, at least, we wouldn't have to hunt after
those changes.

You seem to be going out of your way to break things, with the
justification of being "clean" (whatever that means!). I don't
see any point being "clean" and not working. It may be aesthetically
pleasing, but it's also useless. Since I have little idea
what particular version of libusb various platforms use and
have no desire to track this sort of thing (development
is hard enough as it is!) it's difficult to provide patches,
and far simpler and safer to provide a cohesive version that
is known to work.

On Linux, we (collectively) have the possibility to "fix" the
platform. I can understand why you are not interested
in improving the plaform, but don't blame people who want to do it.
And it would be nice to not make it harder to
fix it.

It's not the platform, it's the upstream (libusb-win32) library.

Where there is no technical justification for not using the system provided
libraries (tifflib), I've accommodated this in the V1.0.0 release
(see the whole new Jambase and Jamfile arrangement).
There are technical reasons for paying the small price to use
the libusb supplied with Argyll, since the operation of certain
instruments is sensitive to it, and this is the version that
I've tested against. When and if a working and widely supported
"standard" version of libusb becomes available (seems a way off given
the state of libusb V1, libusb-win32 V1 and openusb), then I'll
certainly be happy to switch to it, rather than maintaining
the libraries provided with Argyll.

Graeme Gill.

Other related posts: