trunk make check error

[Jason Moxham]

creating t-get_d
gcc -std=gnu99 -O2 -m64 -march=nocona -mtune=nocona -o .libs/t-aors
t-aors.o ../../tests/.libs/libtests.a /root/mpir/mpir/trunk/.libs/libmpir.so ../../.libs/libmpir.so
gcc -std=gnu99 -O2 -m64 -march=nocona -mtune=nocona -o .libs/t-cmp
t-cmp.o ../../tests/.libs/libtests.a /root/mpir/mpir/trunk/.libs/libmpir.so ../../.libs/libmpir.so
gcc -std=gnu99 -O2 -m64 -march=nocona -mtune=nocona -o .libs/t-inp_str
t-inp_str.o ../../tests/.libs/libtests.a /root/mpir/mpir/trunk/.libs/libmpir.so ../../.libs/libmpir.so
gcc -std=gnu99 -O2 -m64 -march=nocona -mtune=nocona -o .libs/t-equal
t-equal.o ../../tests/.libs/libtests.a /root/mpir/mpir/trunk/.libs/libmpir.so ../../.libs/libmpir.so
creating t-aors
gcc -std=gnu99 -O2 -m64 -march=nocona -mtune=nocona -o .libs/t-set_str
t-set_str.o ../../tests/.libs/libtests.a /root/mpir/mpir/trunk/.libs/libmpir.so ../../.libs/libmpir.so
creating t-cmp
creating t-inp_str
creating t-equal
creating t-set_str
gcc -std=gnu99 -O2 -m64 -march=nocona -mtune=nocona -o .libs/t-set_f
t-set_f.o ../../tests/.libs/libtests.a /root/mpir/mpir/trunk/.libs/libmpir.so ../../.libs/libmpir.so
creating t-set_f
make[4]: Leaving directory `/root/mpir/mpir/trunk/tests/mpq'
make check-TESTS
make[4]: Entering directory `/root/mpir/mpir/trunk/tests/mpq'
tests_memory_end(): not all memory freed
80 blocks remaining
/bin/sh: line 4: 22647 Aborted ${dir}$tst
FAIL: t-aors
tests_memory_end(): not all memory freed
4 blocks remaining
/bin/sh: line 4: 22672 Aborted ${dir}$tst
FAIL: t-cmp
tests_memory_end(): not all memory freed
5 blocks remaining
/bin/sh: line 4: 22697 Aborted ${dir}$tst
FAIL: t-cmp_ui
PASS: t-cmp_si
PASS: t-equal
tests_memory_end(): not all memory freed
19 blocks remaining
/bin/sh: line 4: 22768 Aborted ${dir}$tst
FAIL: t-get_d
PASS: t-get_str
PASS: t-inp_str
PASS: t-md_2exp
PASS: t-set_f
PASS: t-set_str
=============================================================
4 of 11 tests failed
Please report to http://groups.google.co.uk/group/mpir-devel/
=============================================================
make[4]: *** [check-TESTS] Error 1
make[4]: Leaving directory `/root/mpir/mpir/trunk/tests/mpq'
make[3]: *** [check-am] Error 2
make[3]: Leaving directory `/root/mpir/mpir/trunk/tests/mpq'
make[2]: *** [check-recursive] Error 1
make[2]: Leaving directory `/root/mpir/mpir/trunk/tests'
make[1]: *** [check-recursive] Error 1
make[1]: Leaving directory `/root/mpir/mpir/trunk'
make: *** [check] Error 2

This is on nehalem linux , but it looks like a general error

[Bill Hart]

Hmm, doesn't happen on my k8. Either I didn't check something in or
there's a bug in something which is system specific.

Anyhow, at least it is only a memory leak. Should be easy to track
down. I'll check it out tomorrow. Gotta go to sleep soon. Got a weird
sleeping schedule at present.

What's a bit weird is the functions it occurs in. Can't be the mulmid
code as it isn't used. Could be in the new toom code, but I just did a
quick check of that and nothing is wrong there.

The other possibility is the new xgcd code. Anyhow I'll check tomorrow.

Bill.

2009/7/24 Jason Moxham <ja...@njkfrudils.plus.com>:

[Jason Moxham]

On cuda(k10) and sage(penryn) I get a syntax error

I notice
#include "gmp.h" in the files , don't know what this header file is does ?
perhaps try mpir.h

Jason

[Jason Moxham]

On Friday 24 July 2009 13:18:20 Jason Moxham wrote:
> On cuda(k10) and sage(penryn) I get a syntax error
>
> I notice
> #include "gmp.h" in the files , don't know what this header file is does ?
> perhaps try mpir.h
>

That fixes the syntax error , we are now just left with the mem leak below

[Bill Hart]

There was a leak in lgcd which is used by a lot of mpq functions. I've
fixed it in trunk. Do you want to try it and see if some of those
errors go away.

[Jason Moxham]

make check now passes