Re: Sage 4.3.4.alpha0 released

[Kiran Kedlaya]

As expected based on my experience with 4.3.3, I got a build error
building 4.3.4.alpha0, though this time it was a linking error with gd
rather than cddlib. Again, this is Fedora 10 on a 64-bit system, but
on a 32-bit network, so there exist 32-bit libraries findable via NFS
(under /usr/local). But this time I don't see any obvious evidence
that this is the source of the trouble.
-----
/bin/sh ./libtool --tag=CC --mode=link gcc -fPIC -g -I"/scratch/
sage-4.3.4.alph
a0/local/include" -I/scratch/sage-4.3.4.alpha0/local/include/
freetype2/ -L/scra
tch/sage-4.3.4.alpha0/local/lib -Wl,--rpath -Wl,/scratch/
sage-4.3.4.alpha0/local
/lib -L/scratch/sage-4.3.4.alpha0/local/lib -o annotate
annotate.o ./libgd.la
-lfontconfig -lfreetype -lpng12 -lz -lm
gcc -fPIC -g -I/scratch/sage-4.3.4.alpha0/local/include -I/scratch/
sage-4.3.4.al
pha0/local/include/freetype2/ -Wl,--rpath -Wl,/scratch/
sage-4.3.4.alpha0/local/l
ib -o .libs/annotate annotate.o -L/scratch/sage-4.3.4.alpha0/local/
lib ./.libs/
libgd.so -lfontconfig /scratch/sage-4.3.4.alpha0/local/lib/
libfreetype.so /scrat
ch/sage-4.3.4.alpha0/local/lib/libpng12.so -lz -lm -Wl,--rpath -Wl,/
scratch/sage
-4.3.4.alpha0/local/lib
./.libs/libgd.so: undefined reference to `libiconv'
./.libs/libgd.so: undefined reference to `libiconv_close'
./.libs/libgd.so: undefined reference to `libiconv_open'
collect2: ld returned 1 exit status
-----
Full log posted (for now) at http://math.mit.edu/~kedlaya/sage434a.log.

Kiran

On Mar 3, 1:05 pm, Minh Nguyen <nguyenmi...@gmail.com> wrote:
> Hi folks,
>
> This release incorporates many combinatorics tickets positively
> reviewed during and/or before Sage Days 20.
>

[Kiran Kedlaya]

As predicted, that doesn't seem to help.

Kiran

On Mar 3, 11:27 pm, Minh Nguyen <nguyenmi...@gmail.com> wrote:
> Hi Kiran,
>
> On Thu, Mar 4, 2010 at 3:12 PM, Minh Nguyen <nguyenmi...@gmail.com> wrote:
>
> <SNIP>
>
> > See ticket #8432 for an implementation of this
> > approach.
>
> From your install log, I see that your build also has gotten past
> installing iconv. This makes me doubt that the solution at #8432 would
> solve the build problem you reported. In any event, I would like to
> know if the solution at #8432 works for your particular system.
>
> --
> Regards
> Minh Van Nguyen

[David Joyner]

On a 10.6.2 mac, the following tests failed:

The following tests failed:

sage -t "devel/sage/sage/categories/examples/finite_semigroups.py"
sage -t "devel/sage/sage/categories/finite_semigroups.py"
sage -t
"local/lib/python2.6/site-packages/sagenb-0.7.5.1-py2.6.egg/sagenb/misc/sageinspect.py"
sage -t
"local/lib/python2.6/site-packages/sagenb-0.7.5.1-py2.6.egg/sagenb/notebook/interact.py"

For example

sage -t "devel/sage/sage/categories/examples/finite_semigroups.py"
**********************************************************************
File "/Users/wdj/sagefiles/sage-4.3.4.alpha0/devel/sage/sage/categories/examples/finite_semigroups.py",
line 75:
sage: S.j_transversal_of_idempotents()
Expected:
['acb', 'ac', 'ab', 'bc', 'a', 'c', 'b']
Got:
['a', 'acb', 'ac', 'ab', 'cb', 'c', 'b']
**********************************************************************
1 items had failures:
1 of 14 in __main__.example_1
***Test Failed*** 1 failures.

Should I post the log somewhere?

On Wed, Mar 3, 2010 at 1:43 PM, Minh Nguyen <nguye...@gmail.com> wrote:
> Hi folks,
>
> On Thu, Mar 4, 2010 at 5:05 AM, Minh Nguyen <nguye...@gmail.com> wrote:
>
> <SNIP>
>
>> * The following tests failed on sage.math:
>
> They also fail on bsd.math and rosemary.math. This issue is now
> tracked at ticket #8430:
>
> http://trac.sagemath.org/sage_trac/ticket/8430

>
> --
> Regards
> Minh Van Nguyen
>

> --
> You received this message because you are subscribed to the Google Groups "sage-release" group.
> To post to this group, send email to sage-r...@googlegroups.com.
> To unsubscribe from this group, send email to sage-release...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/sage-release?hl=en.
>
>

[Kiran Kedlaya]

Looking at the iconv part of my install log, I see some odd lines in
the configure script output,
such as:

checking if gcc supports -c -o file.o... rm: cannot remove
`conftest*': No such
file or directory
yes

That makes me wonder whether iconv is getting built correctly in this
environment.

Kiran

On Mar 3, 11:27 pm, Minh Nguyen <nguyenmi...@gmail.com> wrote:
> Hi Kiran,
>
> On Thu, Mar 4, 2010 at 3:12 PM, Minh Nguyen <nguyenmi...@gmail.com> wrote:
>
> <SNIP>
>
> > See ticket #8432 for an implementation of this
> > approach.
>
> From your install log, I see that your build also has gotten past
> installing iconv. This makes me doubt that the solution at #8432 would
> solve the build problem you reported. In any event, I would like to
> know if the solution at #8432 works for your particular system.
>

[Pat LeSmithe]

On 03/06/2010 06:55 PM, Kiran Kedlaya wrote:
> Looking at the iconv part of my install log, I see some odd lines in
> the configure script output,
> such as:
>
> checking if gcc supports -c -o file.o... rm: cannot remove
> `conftest*': No such
> file or directory
> yes

I see this line repeated three times in what appears to be a successful
build of iconv on sage.math.

> That makes me wonder whether iconv is getting built correctly in this
> environment.

What happens if you add

----with-libiconv-prefix="$SAGE_LOCAL"

to

./configure --prefix="$SAGE_LOCAL" --without-jpeg --without-x
--with-zlib="$SAGE_LOCAL" --with-freetype="$SAGE_LOCAL" --without-xpm

in

gd-2.0.35.p4/spkg-install

?

[Kiran Kedlaya]

If I make that change, then the odd lines go away, but the build still
fails in the same spot.

Kiran