0.3.9 tarballs

[Luigi Ballabio]

Hi all,
        near-final 0.3.9 tarballs are available in  
<http://quantlib.org/prerelease/>. Please try them out and report any  
problems.

Thanks,
        Luigi

----------------------------------------

The purpose of abstraction is not to be vague, but to create a new
semantic level in which one can be absolutely precise.
-- W.E. Dijkstra

[Dirk Eddelbuettel]

(Sorry for the delay, but I was on vaction til Sunday.)

On 1 April 2005 at 09:05, Luigi Ballabio wrote:
|
| Hi all,
| near-final 0.3.9 tarballs are available in  
| <http://quantlib.org/prerelease/>. Please try them out and report any  
| problems.

Ok, after having had to take a pass with the earlier golden master due to the
missing Boost fix from their CVS in the corresponding Debian package (which
was later corrected; thanks to Luigi for sending over the patch), I have
prepare a release candidate for Debian's unstable flavour which is currently
being uploaded. Given that QuantLib 0.3.8 is in Debian's testing
distribution, we may as well try the new release candidate.

Cheers, Dirk

--
Better to have an approximate answer to the right question than a precise
answer to the wrong question.  --  John Tukey as quoted by John Chambers

[Dirk Eddelbuettel]

On 4 April 2005 at 22:54, Dirk Eddelbuettel wrote:
| Ok, after having had to take a pass with the earlier golden master due to the
| missing Boost fix from their CVS in the corresponding Debian package (which
| was later corrected; thanks to Luigi for sending over the patch), I have
| prepare a release candidate for Debian's unstable flavour which is currently
| being uploaded. Given that QuantLib 0.3.8 is in Debian's testing
| distribution, we may as well try the new release candidate.

That worked swimmingly -- the 0.3.9 release candidate was built on all
architectures within a day.

Is there a time frame for when 0.3.9 will hit the limelight?

Dirk

--
Better to have an approximate answer to the right question than a precise
answer to the wrong question.  --  John Tukey as quoted by John Chambers

[Luigi Ballabio]

On Apr 9, 2005, at 3:51 PM, Dirk Eddelbuettel wrote:

> That worked swimmingly -- the 0.3.9 release candidate was built on all
> architectures within a day.

Good.

> Is there a time frame for when 0.3.9 will hit the limelight?

Well, that would depend on when Nando finds the time to build the
Windows installer :)

Later,
        Luigi

[Dirk Eddelbuettel]

On 10 April 2005 at 11:11, Luigi Ballabio wrote:
|
| On Apr 9, 2005, at 3:51 PM, Dirk Eddelbuettel wrote:
|
| > That worked swimmingly -- the 0.3.9 release candidate was built on all
| > architectures within a day.
|
| Good.
|
| > Is there a time frame for when 0.3.9 will hit the limelight?
|
| Well, that would depend on when Nando finds the time to build the
| Windows installer :)

If we have otherwise frozen tarballs (esp. on the Unix side of things), maybe
Nando could issue call for attention so that us porters can get the binaries
ready prior to the announcement.

It would suit me greatly as the rc release would otherwise migrate into
testing, I'd rather replace it in unstable with a final 0.3.9.

Dirk

--
Better to have an approximate answer to the right question than a precise
answer to the wrong question.  --  John Tukey as quoted by John Chambers

[Dirk Eddelbuettel]

So I had a go at rolling the Debian packages of QL forward (to the rc of
2005-05-12, taken yesterday).  New packages have been uploaded, autobuilders
are busy.  

But Ruby failed to built -- it croaks after the tests from setup.rb. From the
build log:

[...]
dpkg-source: building quantlib-ruby using existing quantlib-ruby_0.3.8.rc.20050407.orig.tar.gz
dpkg-source: building quantlib-ruby in quantlib-ruby_0.3.8.rc.20050407-1.diff.gz
dpkg-source: building quantlib-ruby in quantlib-ruby_0.3.8.rc.20050407-1.dsc
 debian/rules build
dpkg-architecture: warning: no utmp entry available and LOGNAME not defined; using uid of process (0)
dh_testdir
touch quantlib_wrap.cpp
CXX="g++" \
CFLAGS="-O2 " \
CXXFLAGS="-O2 " ruby setup.rb build  
Building extension...
creating Makefile
make[1]: Entering directory `/tmp/buildd/quantlib-ruby-0.3.8.rc.20050407'
g++ -fPIC -Wall -g -O2  -fPIC -O2    -I. -I/usr/lib/ruby/1.8/i386-linux -I/usr/lib/ruby/1.8/i386-linux -I.   -I/usr/include -Wno-uninitialized -Wno-unused -O2  -DBOOST_DISABLE_THREADS -c quantlib_wrap.cpp
g++ -shared  -L"/usr/lib" -o QuantLibc.so quantlib_wrap.o  -lruby1.8  -L/usr/lib -lQuantLib-0.3.9 -lpthread -ldl -lcrypt -lm   -lc
make[1]: Leaving directory `/tmp/buildd/quantlib-ruby-0.3.8.rc.20050407'
touch build-stamp
 debian/rules binary
dpkg-architecture: warning: no utmp entry available and LOGNAME not defined; using uid of process (0)
ruby setup.rb test
Building extension...
creating Makefile
make[1]: Entering directory `/tmp/buildd/quantlib-ruby-0.3.8.rc.20050407'
make[1]: Nothing to be done for `all'.
make[1]: Leaving directory `/tmp/buildd/quantlib-ruby-0.3.8.rc.20050407'
Testing QuantLib-Ruby...
Loaded suite QuantLib test suite
Started
Testing date ranges: .
Testing observability of stocks: .
Testing observability of market elements: .
Testing observability of market element handles: .
Testing segment integral: .
Testing 1-D solvers: .
Testing observability of forward-spreaded term structure: .
Testing observability of implied term structure: .
Testing observability of zero-spreaded term structure: .

Finished in 4.872506 seconds.

9 tests, 0 assertions, 0 failures, 0 errors
setup.rb:229: [BUG] Segmentation fault
ruby 1.8.2 (2005-04-11) [i386-linux]

make: *** [test-stamp] Aborted (core dumped)
pbuilder: Failed autobuilding of package
 -> Aborting with an error
 -> unmounting dev/pts filesystem
 -> unmounting proc filesystem
 -> cleaning the build env
    -> removing directory /var/local/cache/pbuilder/build//13139 and its subdirectories

This is on Debian unstable in a fresh chroot with the QL rc 2005-04-12.  

Luigi -- any idea?  Python builds and tests fine (once TermStructureTest is
comemnted out, as has been required for a while).

Thanks, Dirk

--
Better to have an approximate answer to the right question than a precise
answer to the wrong question.  --  John Tukey as quoted by John Chambers

[Luigi Ballabio]

On 04/14/05 14:20:18, Dirk Eddelbuettel wrote:
>
> So I had a go at rolling the Debian packages of QL forward (to the rc of
> 2005-05-12, taken yesterday).  New packages have been uploaded,
> autobuilders are busy.
>
> But Ruby failed to built -- it croaks after the tests from setup.rb.
>
> Luigi -- any idea?

Not really. It works on my Sarge box.


> Python builds and tests fine (once TermStructureTest
> is comemnted out, as has been required for a while).

Did you try adding -DBOOST_DISABLE_THREADS to CXXFLAGS when compiling the  
wrappers?

Later,
        Luigi

----------------------------------------

Within C++, there is a much smaller and cleaner language struggling to get  
out.
-- Bjarne Stroustrup

[Dirk Eddelbuettel]

On 15 April 2005 at 07:59, Luigi Ballabio wrote:
|
| On 04/14/05 14:20:18, Dirk Eddelbuettel wrote:
| >
| > So I had a go at rolling the Debian packages of QL forward (to the rc of
| > 2005-05-12, taken yesterday).  New packages have been uploaded,
| > autobuilders are busy.
| >
| > But Ruby failed to built -- it croaks after the tests from setup.rb.
| >
| > Luigi -- any idea?
|
| Not really. It works on my Sarge box.

May have to do with running in a chroot. I could make it go away by not run
the test suite. Am in a hurry now so that will have to do...
 
| > Python builds and tests fine (once TermStructureTest
| > is comemnted out, as has been required for a while).
|
| Did you try adding -DBOOST_DISABLE_THREADS to CXXFLAGS when compiling the  
| wrappers?

No, apparently not. Will try. Incidentally, ruby (1.8) adds that, may come
from the Debian compiler default configs for it.

Regards, Dirk

--
Better to have an approximate answer to the right question than a precise
answer to the wrong question.  --  John Tukey as quoted by John Chambers

[Luigi Ballabio]

On 04/15/05 14:13:08, Dirk Eddelbuettel wrote:

> >
> | > Python builds and tests fine (once TermStructureTest
> | > is comemnted out, as has been required for a while).
> |
> | Did you try adding -DBOOST_DISABLE_THREADS to CXXFLAGS when compiling
> the
> | wrappers?
>
> No, apparently not. Will try. Incidentally, ruby (1.8) adds that, may  
> come from the Debian compiler default configs for it.

No, it comes from me having added it :)
Curiously enough---as we had the opposite problem with Python---I had  
freezes with QL-Ruby on Sarge which you didn't have on Sid. Adding the  
define cured them.

Later,
        Luigi

----------------------------------------

Hanlon's Razor:
         Never attribute to malice that which is adequately explained
         by stupidity.