Re: GCC 3.4.0 build failed on interix3

[Christopher Faylor <cgf at alum dot bu dot edu>]

On Fri, Apr 23, 2004 at 05:39:51PM -0400, Ross Ridge wrote:
>By using your logic, GCC shouldn't support Cygwin.  Cygwin has helped sell
>many more copies of Windows than copies of Interex have ever been sold.

On Fri, Apr 23, 2004 at 08:26:00PM -0400, Ross Ridge wrote:
>> > By using your logic, GCC shouldn't support Cygwin.
>> 
>> Cygwin is different because its runtime is itself GPL.
> 
>By Christopher Faylor's logic Cygwin is the same because they both help
>the sales of proprietary software.  Well, actually, no, Cygwin is 
>much worse by his logic.

First:

http://www.gnu.org/licenses/gpl-faq.html#TOCGPLIncompatibleLibs

Second, let's recap:

My original objection to Interix was that it was a proprietary add-on to
Windows and, as such, I thought it was against FSF policy for gcc to
support it for the same reason that it was against policy to support
U/WIN.  RMS (and Robert Dewar) disagreed since, unlike U/WIN, Interix
can be considered to be part of the OS and so, falls under the same
rule as when gcc runs on Tru64, Solaris, or HP/UX.  RMS represents the
FSF and so, obviously, has the final word.  I may not agree but my
disagreement is of no consequence.

Cygwin is a GPLed product.  While it runs under a proprietary OS, it is
no different from any other GPLed product which runs on a proprietary
OS.  Using your interpretation of my logic, no GNU tool should be
supported on any proprietary system, which clearly is ridiculous.

In the process of stupidly trying to recap my now-rendered-meaningless
objections to Interix, the argument was made that:

"The ubiquitous availability of GCC is more important for the process
of Free Software Enlightenment than the freedom of GCC itself."

This is a common observation in these FSF/GPL threads.  However, I
pointed out that the extension of this logic would mean that gcc
*should* be supported under U/WIN since, by doing so, gcc would become
more available.  This is a contradiction to RMS's veto of U/WIN, though.
It does not take into account the actual reason why U/WIN was pulled
from gcc.

Finally, IMO, the assertion that Cygwin has helped to sell many copies
of Windows is, IMO, extremely arguable.  I have a very hard time
believing that anyone has bought a copy of Windows because they know
they can get Cygwin on it.  If the existence of free tools was a major
deal for someone, then downloading a linux or a *BSD would make a lot
more sense.  For no cost, you get a faster system with more tools and
more functionality than you would get if you bought Windows and
downloaded Cygwin.  I suspect that anyone educated enough to know about
Cygwin prior to purchasing Windows would also be aware of the existence
of Linux or BSD.

There are a few more standard responses that are certain to be made to
what I've said above but I'll stop here since I'm in off-topic territory
for the gcc mailing list.

cgf