On Feb 24, 9:25 pm, "Georg S. Weber" <georgswe...@googlemail.com>
wrote:
> Hi Michael,
Hi Georg,
> this got me a bit high-spirited
:)
>, so I assembled another singular.spkg
> with some bigger cleanups w.r.t memory allocator usage. I did a
> complete build with this spkg dropped into Sage 3.3 instead of the
> original spkg on my OS X 32bit box, and the build and subsequent
> "testlong" runs just finished without any failure.
Ok, as mentioned earlier I just got to Athens a few hours ago and am
catching up, but I will take this for a spin tomorrow.
> If you could give that one another try on some idle OS X 64 box --- if
> there is something new to be seen, we should get even further. Yes, I
> saw the message from Carl. But OO programming and garbage collector
> stuff in not the kind of programming land I'm acquainted to. So while
> his approach is from "above", mine is from "below" --- let's squeeze
> hard together. But if the approach "from below" yields nothing, the I
> fear others (Carl, Robert, ...) have to step in.
Ok, 64 bit OSX might still contain a mix of problems.
> This spkg might also be worth a shot for some renewed Cygwin
> experimentation. I tried hard to get away all proprietary functions
> called "malloc" and being deing defined by using the std header
> <stdlib.h> although the implementation is proprietary and inside
> libomalloc. Maybe that was a problem for making a DLL on Cygwin out of
> Singular/omalloc. Let's see (there's no hurry).
The cygwin problem was that the memory manager was never initialized,
due to hard coded flags via numerical values for certain RTLD_* flags,
but that is fixed now. That it worked at all on OSX is minor miracle
and we can probably thank the OSX linker/dlopen implementation for
that.
> Oh, I should mention that this spkg is uploaded on sage.math under /
> home/weberg/spkg/singular-3-0-4-4-20080711.p3test01.spkg
>
> Cheers,
> gsw
Cheers
Michael
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
