Re: GNU Win32? Not anymore.

[Klee Dienes <klee@mit.edu>]

> OK. Why is there no binary package?

For some reason it keeps getting removed from the distribution.  I've
uploaded it two or three times now, actually; I have no idea what
could be going wrong.

I'll re-build the binary packages and re-upload them as soon as I can
get a working development environment again.

There is a mailing list for the software, <gnu-win32@cygnus.com>.
There are at least a few other people out there who are upset by the
Cygnus licensing; I'm not sure how willing they are to back up their
irritation with real code.  My experience with the readers of the
gnu-win32 list is that (with a few notable exceptions), most of them
don't really understand the "fix-it-yourself" aspect of free software
--- they tend to do a lot more complaining than contributing.

Some assorted thoughts on the issue:

* GNU/Win32 is more than just the winsup library

  The bulk of the work in the cygwin32 project was in the modification
  of the compiler tools and GNU toolchain to support Win32
  executables.  All of that code has been contributed into the
  mainstream sources and is covered by the GPL. Compared to the
  toolchain work, the winsup library is actually a small piece of any
  GNU/Win32 porting effort.

* The winsup library is not all that hard to replace

  My experience with trying to port dpkg to Cygwin32 was that the winsup
  library itself was incomplete and rather hastily written, and that
  furthermore its development progress suffered a huge blow when Steve
  Chamberlain stopped working on the project.

  Writing a replacement winsup library will be tricky, but is not all
  that enormous an amount of work --- there are, after all, not all
  that many syscalls out there to emulate.

* Cygnus is not our enemy

  I think Cygnus has made a serious lapse in judgment in their
  handling of the winsup licensing (personally, I think they made it a
  while ago, when they switched from the LGPL to the GPL for winsup
  licensing).

  Their closed-door development policy regarding GCC and GDB also
  leaves much to be desired, and it's increasingly easy to come up with
  conspiracy theories regarding their core tools release engineering
  policies.

  But despite all this, I think it's important to put their actions
  in proper context.  Cygnus has done a tremendous amount of work on
  the GNU toolchain to support Windows NT as part of their Cygwin32
  work, all of which is released under the GPL, and has done an
  enormous amount of work to support free software in general.  

  Even if they have lost the faith recently (which is not yet
  certain), they are still vastly more in tune with the goals of the
  free software community than any other company out there, and we
  should support them and work with them whenever possible (I agree
  that the winsup library is not currently such a case).

* Debian GNU/Win32 is an important project  
  
  I believe that creating a solid Debian GNU/Win32 distribution is an
  incredibly important goal for the Debian project, whether we can do
  it in partnership with Cygnus, or whether we are forced to do it on
  our own.

  There are millions of developers currently using and developing for
  the Windows platform.  Many of them might well be sympathetic to the
  goals of free software, but have not yet been exposed to it, 
  or are too entrenched in the use of proprietary tools to really
  appreciate the benefits.

  Porting Debian to this platform allows these people to use and
  contribute to the development of free software even in 
  circumstances where it is not practially or politically feasible for
  them to switch directly to a Linux or other non-Win32 system, and
  opens up for us a huge market of potential developers.

  Furthermore, porting to GNU/Win32 is a great test of the strength of
  our distribution and of our packaging system.  I expect that the
  problems we find and fix in a GNU/Win32 port will teach us a lot
  about ways to make our distribution stronger in general.

  One of my my major goals for dpkg is to make it a single source and
  binary packaging format capable of being used on all architectures and
  by all platforms.  I would like a single Debian source package to be
  able to be used as the single upstream source, and to be able to be
  built automatically for all possible platforms and in all necessary
  formats, including foreign packaging formats.

  Obviously, we have a long way to go before this goal is achieved.
  But I think that porting Debian to GNU/Win32 would be a good step
  along the way, as well as provide us the potential for a lot of new
  converts to the free software movement, even if we do end up having
  to re-write the winsup layer ourselves before we are finished.
  I don't have the time to do it myself, but I'd be happy to help and
  support anyone who does, at least so far as the packaging tools are
  concerned.


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org . 
Trouble?  e-mail to templin@bucknell.edu .