On Tuesday 21 April 2009 17:47:23 Bill Hart wrote:
> Great, so we can do 1.1.1.
>
> Let's do the following:
>
> * I'll issue a release candidate (hopefully later today).
> * Jeff could you verify that the release candidate works on your
> machine as expected.
> * Jeff could you also check the build proceeds on Windows on the
> machines you'd like to test on as we didn't do a full round of testing
> for 1.1 on Windows.
> * Jason and I can verify ./config.guess still returns the correct
> thing on all the linux build machines on SkyNet and the other machines
> we have access to.
menas and fulvia are suspect , the rest of skynet is OK and my own boxes.
I'll look at it now.
>
> Michael reports there were no issues with MPIR 1.1 for Sage so this
> should remain true and likely MPIR 1.1.1 will end up in Sage 3,4.2 I
> guess.
>
> * I'll do the release on the website.
>
> Bill.
>
> 2009/4/21 Jason Moxham <ja...@njkfrudils.plus.com>:
> > On Tuesday 21 April 2009 16:04:00 Jeff Gilchrist wrote:
> >> On Tue, Apr 21, 2009 at 10:19 AM, Jason Moxham
> >>
> >> <ja...@njkfrudils.plus.com> wrote:
> >> > This is getting stranger , jeff1.c is or perhaps I should say supposed
> >> > to be exactly the same same as mpir generated {dummy}32.c
> >> >
> >> > Can you a diff for us
> >>
> >> This does not make any sense. You are right the only difference is
> >> that you added /* version 1 */ to the top of jeff1.c. Even when I add
> >> that line to dummy-278832.c making it identical it fails for that file
> >> but not for jeff1.c.
> >>
> >> [je...@tope mpir-1.1]$ md5sum dummy-278832.c jeff1.c
> >> 885509cb2c8cdf8904757c65718605ff dummy-278832.c
> >> 885509cb2c8cdf8904757c65718605ff jeff1.c
> >>
> >> [je...@tope mpir-1.1]$ cc dummy-278832.s jeff1.c -o test2
> >> jeff1.c:9: warning: return type defaults to ‘int’
> >> jeff1.c: In function ‘main’:
> >> jeff1.c:27: warning: implicit declaration of function ‘cpuid’
> >> jeff1.c:42: warning: implicit declaration of function ‘strcmp’
> >> jeff1.c:119: warning: implicit declaration of function ‘printf’
> >> jeff1.c:119: warning: incompatible implicit declaration of built-in
> >> function ‘printf’
> >>
> >> [je...@tope mpir-1.1]$ cc dummy-278832.s dummy-278832.c -o test2
> >> dummy-278832.c:9: warning: return type defaults to ‘int’
> >> dummy-278832.c: In function ‘main’:
> >> dummy-278832.c:27: warning: implicit declaration of function ‘cpuid’
> >> dummy-278832.c:42: warning: implicit declaration of function ‘strcmp’
> >> dummy-278832.c:119: warning: implicit declaration of function ‘printf’
> >> dummy-278832.c:119: warning: incompatible implicit declaration of
> >> built-in function ‘printf’
> >> /tmp/cco.v1c9C5: In function `main':
> >> /work/jeffg/mpir-1.1/dummy-278832.c:9: multiple definition of `main'
> >> /tmp/cco.v1c9C5:/work/jeffg/mpir-1.1/dummy-278832.c:9: first defined
> >> here /tmp/cco.v1c9C5: In function `main':
> >> (.text+0x5b): undefined reference to `cpuid'
> >> /tmp/cco.v1c9C5: In function `main':
> >> (.text+0x76): undefined reference to `cpuid'
> >> /tmp/cco.v1c9C5: In function `main':
> >> (.text+0x5b): undefined reference to `cpuid'
> >> /tmp/cco.v1c9C5: In function `main':
> >> (.text+0x76): undefined reference to `cpuid'
> >> collect2: ld returned 1 exit status
> >>
> >> Ah, I have something. For some reason if the .s file and the .c file
> >> have the same name it fails, if they are different it works. That is
> >> why the file from gmp and jeff1 worked because config.guess was still
> >> using the dumy-$ value so different names. If I try:
> >>
> >> [je...@tope mpir-1.1]$ cc jeff1.s jeff1.c -o test2
> >> jeff1.c:9: warning: return type defaults to ‘int’
> >> jeff1.c: In function ‘main’:
> >> jeff1.c:27: warning: implicit declaration of function ‘cpuid’
> >> jeff1.c:42: warning: implicit declaration of function ‘strcmp’
> >> jeff1.c:119: warning: implicit declaration of function ‘printf’
> >> jeff1.c:119: warning: incompatible implicit declaration of built-in
> >> function ‘printf’
> >> /tmp/cco.HDcIhW: In function `main':
> >> /work/jeffg/mpir-1.1/jeff1.c:9: multiple definition of `main'
> >> /tmp/cco.HDcIhW:/work/jeffg/mpir-1.1/jeff1.c:9: first defined here
> >> /tmp/cco.HDcIhW: In function `main':
> >> (.text+0x5b): undefined reference to `cpuid'
> >> /tmp/cco.HDcIhW: In function `main':
> >> (.text+0x76): undefined reference to `cpuid'
> >> /tmp/cco.HDcIhW: In function `main':
> >> (.text+0x5b): undefined reference to `cpuid'
> >> /tmp/cco.HDcIhW: In function `main':
> >> (.text+0x76): undefined reference to `cpuid'
> >> collect2: ld returned 1 exit status
> >>
> >> You can see it fails. So for some reason using the same name is bad.
> >> GMP 4.3.0 uses different names they use 0.s or 1.s and 2.c, maybe they
> >> have encountered this problem before.
> >>
> >> So to fix this problem you would need to change config.guess to use
> >> something like:
> >>
> >> dummy-278832.c
> >> dummy-2788320.s
> >> dummy-278864.c
> >> dummy-2788640.s
> >>
> >> instead of the same filename.
> >>
> >> Jeff.
> >
> > Brilliant , Thanks
> > What an error !!
> > I think we should report this pathscale , or see if they have bug report
> > allready
> >
> > I suspect that the cc "compiles" prog.c to prog.s and therefore
> > overwrites our prog.s , and when we link them we get two mains , and an
> > undefined reference to cpuid
> >
> > I put change in trunk and mpir-1.1 branch.
> >
> > Thanks
> > Jason
>
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"mpir-devel" group.
To post to this group, send email to mpir-devel@googlegroups.com
To unsubscribe from this group, send email to
mpir-devel+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/mpir-devel?hl=en
-~----------~----~----~----~------~----~------~--~---
