On Thursday, 19 July 2012 07:35:15 UTC+8, Jean-Pierre Flori wrote:
>
>
> In fact I somehow solved the problem.
> I think there is some black magic dll import/export stuff missing to let 
> ld link the files together on Cygwin.
>
> As a first solution, adding -Wl,--enable-auto-import to 
> extra_compiler_flags in module_list.py seems to be sufficient to fix the 
> problem.
>

As far as I recall, we did something like this to a number of packages 
already last year...

 

> This let the farey stuff compile.
> Hopefully this will be the last step before Sage completely compiles, with 
> obviously no guarantee that it actually starts afterward :)
>
> A more proper fix could be to let Cython add _declspec(_dllimport) 
> (approximate spelling...) to the produced header file, not sure about that.
>

it could be the case. IMHO Cython is not tested on Cygwin, although I might 
be wrong there.
Last time I checked (cygwin version 1.7.10 ?) Sage's Cython 
building dynamic modules, even very simple ones, did not work well due to 
Cygwin
fork problems. 
Even now if I try easy_install cython on Cygwin, I get an error indicating 
that it has no idea about Cygwin, basically, complaining about absence of 
Microsoft's C compiler.

Dima
 

>
> By the way, i guess the farey lines are a missing proper "depends" section 
> including farey.hpp and sl2z.hpp, but that's not the source of my problems.
>

-- 
-- 
To post to this group, send an email to [email protected]
To unsubscribe from this group, send an email to 
[email protected]
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org



Reply via email to