[gecode-users] Adventures in installation

[Martin Mann]

Hi Christian,

mhh.. I am using cygwin too. I encountered several problems by the old shipped gcc too but until now I was lucky and I got everything working. But I ve got to admit I ve not found the time to check gecode 2.X.

I my opinion it sounds a bit hard not to support cygwins gcc but I see the problems of course. With a fast look into the net I ve not found anybody with a proper working gcc 4.x in cygwin. :(
Anyway, a direct support of cygwin without the msvc compiler would be great if the effort is acceptable. Would be a step to complicate a gecode usage for poor windows users as I am and may stop some people to give gecode a try! And that would be a pitty! ;)

Anyway I will have to try the msvc compiler if you do such a strong suggestion! Hopefully the error output will be a bit more expressive.

Hope it helps,

Martin


-------- Original-Nachricht --------
> Datum: Tue, 15 Jan 2008 21:18:16 +0100
> Von: "Christian Schulte" <cschulte at kth.se>
> An: "\'David Barton\'" <d.barton at edaptive.com>, users at gecode.org
> Betreff: Re: [gecode-users] Adventures in installation

> Dear David,
> 
> let me explain what happens here. I am a Cygwin user since a long time ;-(
> 
>  - The first thing you notice is that the gcc shipped with Cygwin rejects
> legal code. We might fix that for the 
>    next release, most likely (see below).
>  - The other thing is a bug in Gecode 2.0.0 that has been fixed for 2.0.1
> (it has to do with static linking, Guido is that true?).
> 
> But, actually, my recommendation is very simple: forget gcc on Cygwin as
> there is no need whatsoever to use it.
> 
> As said, I use cygwin every day but together with the freely available
> Express Edition of Microsoft's Visual C++ compiler. It works as follows:
> download, install, set up the environment that cygwin can pick up the
> necessary environment variables for the commandline version of the
> compiler
> (easy, has been discussed before on this list) and then ./configure CC=cl
> CXX=cl (the documentation discusses compiling for MSVC).
> 
> The benefits are enormous. Compilation is much much faster and so is the
> generated code.
> 
> I stopped using gcc on cygwin completely as it is so outdated and any
> improvement is not in sight. But I have to admit that even if a recent gcc
> would be available on cygwin I still would use MSVC.
> 
> I am not even sure that we should waste any effort on hacking around gcc's
> bugs under cygwin, maybe the best would be to just say on the
> documentation
> what I said in this email and state clearly that gcc is not supported on
> cygwin (as dynamic linking is broken anyway).
> 
> Would that be a hardship for any of the users?
> 
> Thanks for the feedback!
> Christian
> 
> --
> Christian Schulte, www.ict.kth.se/~cschulte/
> 
> 
> -----Original Message-----
> From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On Behalf
> Of David Barton
> Sent: Tuesday, January 15, 2008 3:48 PM
> To: users at gecode.org
> Subject: [gecode-users] Adventures in installation
> 
> Gentle gecode users:
> 
> I have a virgin machine on which I am attempting to load gecode.  Because
> it
> 
> is a virgin, beautiful as it is, I downloaded a new cygwin installation,
> so 
> it has the latest everything.  Downloading gecode and doing .configure, 
> everything went splendidly.  When I did a make, I got a g++ compile error
> on
> 
> ranges-diff.icc --- the goto statement at line 94 gave me one of those 
> strange C++ errors that you are branching to a label in a wrong dimension,
> or whatever they want to say.  I got that to work by putting a loop around
> the whole thing, moving the label down to the bottom of the loop, and 
> sticking a nonsense statement after the loop.  Fine.
> 
> Now, when it gets to the final load statement of examples/bacp.exe, I am 
> getting ten thousand undefined reference statements in 
> minimodel/bool-expr.cc and int/var/bool.icc.  Most of these are of
> BoolVars 
> of various sorts.  Those I *don't* know what to do about.  I suppose
> having 
> an example not load won't kill me, but I'd like to know what is going on.
> 
> Any assistance would be greatly appreciated.....
> 
> David L. Barton
> Director, Rosetta CoE
> EDAptive Computing, Inc.
> 1245-G Lyons Road, Dayton, OH 45458
> (937) 433-0477
> 
> CONFIDENTIALITY NOTICE: The information transmitted is intended only for
> the
> 
> person or entity to which it is addressed and may contain company 
> proprietary and/or privileged material. Any review, retransmission, 
> dissemination or other use of, or taking of any action in reliance upon, 
> this information by persons or entities other than the intended recipient
> is
> 
> prohibited. If you received this in error, please contact the sender and 
> destroy any copies of this document. 
> 
> 
> 
> _______________________________________________
> Gecode users mailing list
> users at gecode.org
> https://www.gecode.org/mailman/listinfo/gecode-users
> 
> 
> _______________________________________________
> Gecode users mailing list
> users at gecode.org
> https://www.gecode.org/mailman/listinfo/gecode-users