On Saturday 22 May 2010 03:22:40 Bill Hart wrote:
> On 22 May 2010 03:17, Jason Moxham <[email protected]> wrote:
> > On Friday 21 May 2010 14:16:49 Bill Hart wrote:
> >> Regarding time frame for the release, I really hope there are plans to
> >> get mpir-2.0.1 out the door as soon as possible.
> >>
> >> I know Sage has been waiting on these gcd normalisation issues to be
> >> fixed for absolutely ages. They still use mpir-1.2.2 !!!
> >>
> >> The significant holdups I see are:
> >>
> >> * broken with gcc 4.5.0 on ia64  - probably not an MPIR bug (may have
> >> to blacklist this gcc on ia64 along with gcc 4.3.2 on all 64 bit
> >> machines)
> >
> > this is fixed
> >
> >> * some issue on another skynet machine - something to do with the
> >> setup of the machine
> >
> > this is fixed except for a purely local issue(I think) wrong linker
> > installed see this from configue log on fulvia
> >
> > hecking if gcc -std=gnu99 supports -c -o file.o... yes
> > checking whether the gcc -std=gnu99 linker
> > (/usr/local/binutils-2.20.1/x86_64- SunOS-core2-gcc-4.4.3/bin/ld)
> > supports shared libraries... yes
> > checking whether -lc should be explicitly linked in... yes
> > checking dynamic linker characteristics... solaris2.10 ld.so
> >
> > where gcc is version 4.5.0 , but the linker is 4.4.3
> >
> >> * add copyi.c copyd.c to linux build (this is done by adding them to
> >> mpn_functions in configure.in and running autoconf, automake, ON
> >> BOXEN!!)
> >
> > I can do this now , easy peasey lemon squeezey :)
> >
> >> * document the normalisation changes in mpir.texi (if this hasn't
> >> already been done - I forgot if I did it or not)
> >
> > Is there a spec , or do we just state we are now the same as old gmp and
> > gmp post 5.0.1?
>
> Simply take the definitions from the GMP docs for mpn_gcdext. The new
> test code tests precisely this spec and I've tested hundreds of
> millions of normalisations to make sure it is right.
>
> >> I would focus attention on these so the release can be made.
> >
> > There is the rosemary make tune issue (although isn't rosemary a virtual
> > machine?)
>
> I don't know where Rosemary is. Minh?
>
> > And what do we what to do about the mpn_sqr(_n) names ? I cant figure out
> > how to duplicate the symbols in libtool.
> > I could do it in asm like this.
> > rename our sqr as "internal_sqr_fn"
> > define in asm two duplicate symbols "mpn_sqr" and "mpn_sqr_n" and just
> > point both of these to the internal_sqr_fn" symbol (perhaps need one jmp
> > instruction?)
>
> The only thing you can really do is change the name internally to
> mpn_sqr. Put the #define the other way around in mpir.h.
>

but this will break apps that are dynamically linked with the old way of 
mpn_sqr_n , rather than the new apps that use mpn_sqr , we will of course 
update the so-numbers . If that is acceptable ?

> I think Brian's solution would work, but it introduces a function
> pointer I think, so it's not efficient.
>

I think Brian sol would work , mostly , but because I can see no other project 
using it I'm not convinced it wont cause problems later . 


> >> Bill.
> >>
> >> On 21 May 2010 13:07, Jason Moxham <[email protected]> wrote:
> >> > On Friday 21 May 2010 07:54:49 Cactus wrote:
> >> >> On May 21, 12:24 am, Jason Moxham <[email protected]> wrote:
> >> >> > On Wednesday 19 May 2010 12:49:51 Bill Hart wrote:
> >> >> > > Hi all,
> >> >> > >
> >> >> > > In the past few months I've been increasingly busy with my
> >> >> > > research and looking at my calendar for the next few months (urgh
> >> >> > > is it even years), I see I am going to have not as much time as I
> >> >> > > would like to work on bignum stuff.
> >> >> > >
> >> >> > > I think it is time to pass on the responsibilities I've had
> >> >> > > within the project to some others.
> >> >> > >
> >> >> > > I'll still be available to ask stuff and I'm sure I'll be able to
> >> >> > > contribute some code from time to time.
> >> >> > >
> >> >> > > I understand that the 2.0.1 release is quite close and I'll still
> >> >> > > be available to help Minh through the process.
> >> >> > >
> >> >> > > The logical person to be in charge of linux stuff is Jason Moxham
> >> >> > > (who is also capable of making releases) and of course Brian is
> >> >> > > as always, in charge of Windows stuff. For contributions for
> >> >> > > upcoming releases, please talk to these guys. I'm sure they'll
> >> >> > > update everyone with any plans for releases as they come to
> >> >> > > light.
> >> >> > >
> >> >> > > Bill.
> >> >> >
> >> >> > OK , I'm happy to do it.
> >> >> > From memory I believe this is what we have up and coming...
> >> >> >
> >> >> > v2.0.1 is a gcd normalization and a few minor bugletts
> >> >> >
> >> >> > v2.1.0 is a MSVC vs2010 update (assuming everything appears on
> >> >> > time)
> >> >> >
> >> >> > There are some minor x86_64 assembler tweeks/updates , linux
> >> >> > portions ready to go , and windows versions depending on how the
> >> >> > MSVC update goes. My windows box is still in the shop so I can't do
> >> >> > the windows bit for the moment.
> >> >>
> >> >> Let me know when you add these to SVN and I'll look at doing it.
> >> >>
> >> >> When you say 'minor tweeks/updates', do you mean a few individual
> >> >> instruction changes that won't need full conversion or full 'Linux to
> >> >> Windows' assembler code conversion?
> >> >>
> >> >>     Brian
> >> >
> >> > Both , but mainly full conversion , I could do the simpler ones if I
> >> > had my windows box back from the shop , I might get it back in time
> >> > depends on what sort of time frame we expect for the release. if not
> >> > we can delay until the next release.
> >> >
> >> > Jason
> >> >
> >> >
> >> > --
> >> > You received this message because you are subscribed to the Google
> >> > Groups "mpir-devel" group. To post to this group, send email to
> >> > [email protected]. To unsubscribe from this group, send
> >> > email to [email protected]. For more options,
> >> > visit this group at http://groups.google.com/group/mpir-devel?hl=en.
> >
> > --
> > You received this message because you are subscribed to the Google Groups
> > "mpir-devel" group. To post to this group, send email to
> > [email protected]. To unsubscribe from this group, send email
> > to [email protected]. For more options, visit this
> > group at http://groups.google.com/group/mpir-devel?hl=en.

-- 
You received this message because you are subscribed to the Google Groups 
"mpir-devel" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/mpir-devel?hl=en.

Reply via email to