About ReplacingGMPNotes

Brian Hulley brianh at metamilk.com
Sun Sep 16 07:00:48 EDT 2007

Roberto Zunino wrote:
> I noticed the following on
> http://hackage.haskell.org/trac/ghc/wiki/ReplacingGMPNotes
> ===
> The LGPL licensing situation may be worse: even though The Glasgow
> Haskell Compiler License is essentially a "free software" license
> (BSD3), according to paragraph 2 of the LGPL, GHC must be distributed
> under the terms of the LGPL!
> ===
> This seems to be inaccurate to me. The LGPLv2.1, section 6, states that,
> if you statically-link, it is enough to provide the GHC source (even
> under BSD), o re-linkable object code. I think the newer LGPL has a
> similar clause.
Well, I think that's what the quote you quoted says: when you distribute 
GHC, you have to distribute it under the restrictions imposed by the 
LGPL, namely to provide relinkable object code or source + permission 
for reverse engineering.

This means that if you want to incorporate GHC into a proprietary 
product you would have to provide re-linkable object code for that 
product as well as permission to reverse engineer the code. This is of 
course not a problem for some people but it is nevertheless a 
restriction that could bite if other issues prevent you being able to 
release your code under such conditions.

It also means that if you want to develop a proprietary app for Windows, 
the fact that GMP is hardwired into the runtime means that everything 
you build, even if it doesn't need Integer, has to be released with 
relinkable object code + reverse engineering permissions.

All I'm saying is: please don't dissuade people from the effort to 
replace GMP with a BSD3 alternative, since I'm sure a lot of people 
would **love** this to happen as soon as possible.


More information about the Glasgow-haskell-users mailing list