Package Details: quantlib 1.27-1

Git Clone URL: https://aur.archlinux.org/quantlib.git (read-only, click to copy)
Package Base: quantlib
Description: A free/open-source library for quantitative finance.
Upstream URL: http://quantlib.org
Licenses: BSD
Submitter: None
Maintainer: xgdgsc (thrasibule)
Last Packager: thrasibule
Votes: 11
Popularity: 0.000000
First Submitted: 2007-01-01 07:43 (UTC)
Last Updated: 2022-07-26 18:01 (UTC)

Latest Comments

a.kudelin commented on 2022-07-26 11:24 (UTC) (edited on 2022-07-26 12:32 (UTC) by a.kudelin)

@thrasibule, I found that the current package configuration makes some tests from the internal suite fail. I have already reported the problem on github. Also it would be nice to add the flag -mno-avx.

thrasibule commented on 2022-07-26 05:55 (UTC)

I'll bump it. What loss of precision are you talking about, and where do you fix it in your patch?

a.kudelin commented on 2022-07-25 20:05 (UTC)

Dear maintainers,
Please consider my patch bumping the version and fixing the loss of precision:
https://pastebin.com/xRG6xnsz

thrasibule commented on 2019-03-15 00:47 (UTC)

@retronostalgia getting rid of boost it the goal, but right now QuantLib still uses it even with --enable-std-function and enable-std-classes, so can't quite get rid of it yet. But downstream users of the library can get rid of boost, like RQuantLib if they fixed their code.

retronostalgia commented on 2019-03-14 22:16 (UTC)

@thrasibule I see - thanks for the info. For now, I think it's simplest for me to recompile with boost. Although - if boost ends up not being needed, going forward, should it be removed from the dependency list?

Anyway, if it's useful for anyone, here is the issue I opened on GitHub originally: https://github.com/eddelbuettel/rquantlib/issues/129

thrasibule commented on 2019-03-14 18:53 (UTC)

@retronostalgia --enable-std-class was a typo, so it doesn't do anything. What happens is that quantlib now allows to be compiled without using boost (which is what the settings --enable-std-function and --enable-std-classes do). The RQuantLib library has boost hardcoded which is why it doesn't work. You can use my branch of rquantlib https://github.com/thrasibule/rquantlib/tree/no_boost which has a fix in. Or you can recompile quantlib with boost like you did.

retronostalgia commented on 2019-03-14 18:13 (UTC)

@thrasibule thanks. I've also discovered that if I modify the current PKGBUILD to use the configure options of the previous PKGBUILD (i.e. change --enable-std-classes back to --enable-std-class and delete --enable-std-function) that RQuantLib works for me again. I don't know what those options do, however...

thrasibule commented on 2019-03-11 20:20 (UTC)

@retronostalgia You need this patch: https://github.com/thrasibule/rquantlib/commit/c458c09c94f38c5335197741320635e17c03d7ca or install my fork of rquantlib...

retronostalgia commented on 2019-03-11 19:58 (UTC)

After upgrading to 1.15-1, the R package that calls QuantLib (RQuantlib) no longer loads or compiles. 1.14 worked fine.

The RQuantLib load error is: Error: package or namespace load failed for ‘RQuantLib’ in dyn.load(file, DLLpath = DLLpath, ...): unable to load shared object '~/R/x86_64-pc-linux-gnu-library/3.5/RQuantLib/libs/RQuantLib.so': ~/R/x86_64-pc-linux-gnu-library/3.5/RQuantLib/libs/RQuantLib.so: undefined symbol: _ZNK8QuantLib9CPICoupon15checkPricerImplERKN5boost10shared_ptrINS_21InflationCouponPricerEEE

...and the first of many RQuantLib compile errors is:

affine.cpp: In function ‘void calibrateModel2(const boost::shared_ptr<QuantLib::ShortRateModel>&, const std::vector<boost::shared_ptr<QuantLib::BlackCalibrationHelper> >&, QuantLib::Real, Rcpp::NumericVector&, Rcpp::NumericVector&, Rcpp::NumericVector&)’: affine.cpp:37:88: error: no matching function for call to ‘QuantLib::ShortRateModel::calibrate(const std::vector<boost::shared_ptr<QuantLib::BlackCalibrationHelper> >&, QuantLib::LevenbergMarquardt&, QuantLib::EndCriteria)’ model->calibrate(helpers, om,QuantLib:: EndCriteria(400,100,1.0e-8, 1.0e-8, 1.0e-8));

The RQuantLib maintainer recommended I report the issue here. Let me know if you need any more info.

thrasibule commented on 2017-03-24 15:15 (UTC)

@xgdgsc, sorry I just saw your response. I can be a co-maintainer. I use QuantLib regularly, so I can help with version bumps and the like.

xgdgsc commented on 2017-03-03 13:05 (UTC)

@thrasibule Thanks. Added. Would you like to become a co-mantainer?

thrasibule commented on 2017-03-02 17:52 (UTC) (edited on 2017-03-02 18:30 (UTC) by thrasibule)

Could you add the following options to configure? --enable-intraday --enable-openmp --disable-static The first one enable high precision dates, and the second one parallelize some code. The last one is to speed up compile. No need to build the static archive since it's not installed anyway. Also you might want to remove the -j`nproc` option. The right parameter should already be passed by the MAKEFLAGS variable in makepkg.conf. (and nproc might not be the right number anyway, with distcc for instance).

commented on 2012-06-01 07:04 (UTC)

I do not use this package now, please adopt it.

commented on 2011-05-28 15:03 (UTC)

Updated to version 1.1

commented on 2011-05-25 05:14 (UTC)

Does not build the package: if g++ -DHAVE_CONFIG_H -I. -I. -I../../ql -I../.. -I../.. -march=i686 -mtune=generic -O2 -pipe -MT BermudanSwaption.o -MD -MP -MF ".deps/BermudanSwaption.Tpo" -c -o BermudanSwaption.o BermudanSwaption.cpp; \ then mv -f ".deps/BermudanSwaption.Tpo" ".deps/BermudanSwaption.Po"; else rm -f ".deps/BermudanSwaption.Tpo"; exit 1; fi In file included from ../../ql/math/interpolations/all.hpp:18:0, from ../../ql/math/all.hpp:31, from ../../ql/quantlib.hpp:46, from BermudanSwaption.cpp:22: ../../ql/math/interpolations/multicubicspline.hpp:276:33: error: reference 'y2_' cannot be declared 'mutable' [-fpermissive] ../../ql/math/interpolations/multicubicspline.hpp:277:34: error: reference 'v_' cannot be declared 'mutable' [-fpermissive] ../../ql/math/interpolations/multicubicspline.hpp:336:33: error: reference 'y2_' cannot be declared 'mutable' [-fpermissive] ../../ql/math/interpolations/multicubicspline.hpp:337:34: error: reference 'v_' cannot be declared 'mutable' [-fpermissive] ../../ql/math/interpolations/multicubicspline.hpp:337:39: error: reference 'v1_' cannot be declared 'mutable' [-fpermissive] ../../ql/math/interpolations/multicubicspline.hpp:337:45: error: reference 'v2_' cannot be declared 'mutable' [-fpermissive] make[2]: *** [BermudanSwaption.o] Error 1

lrm commented on 2010-07-31 16:38 (UTC)

Took over maintainership and updated to the latest, 1.0.1.