January 2010 - gsll-devel - mailman3.common-lisp.net

Re: [Gsll-devel] patch to export random number distribution symbols
by Liam Healy 17 Jan '10

17 Jan '10

OK I have changed the type symbols so that they are in the keyword package. Thanks for the patches. Liam On Mon, Jan 11, 2010 at 12:13 AM, Russell Kliese <russell(a)kliese.id.au> wrote: > There is no objection from me to putting them in the keyword package. > > Russell > > Liam Healy wrote: >> >> No, in fact when I first looked at this patch I wondered >> myself why I hadn't put them in the keyword package. >> I guess it was because these were originally functions >> with their own names, e.g. sample-uniform, so I just >> carried the symbol over when they became methods >> of #'sample. Is the consensus that they be in the keyword >> package? >> >> Liam >> >> On Thu, Jan 7, 2010 at 3:23 PM, James Wright <james(a)chumsley.org> wrote: >> >>> >>> A related question for Liam: It seems like keywords would make more >>> sense for the selector symbols (e.g., :uniform rather that >>> 'gsl:uniform). Is there any particular reason that the selection >>> symbols are in the GSL package instead? >>> >>> Thanks, >>> James >>> >>> On Wed, Jan 6, 2010 at 7:12 PM, Russell Kliese <russell(a)kliese.id.au> >>> wrote: >>> >>>> >>>> It appears that the symbols used to select the various random number >>>> sample >>>> functions are not exported. I was getting "The symbol "UNIFORM" is not >>>> external in the GSLL package." when I ran the following code. >>>> >>>> (clc:clc-require :gsll) >>>> (defparameter *rng* (gsl:make-random-number-generator >>>> gsl:+default-type+)) >>>> (gsl:sample *rng* 'gsl:uniform) >>>> >>>> I've created a patch to merge into the git repository, but I'm not a >>>> Lisp >>>> expert, so my solution may be wrong. >>>> >>>> Thanks to all who have contributed to this very useful library. >>>> >>>> Russell >>>> >> >> _______________________________________________ >> Gsll-devel mailing list >> Gsll-devel(a)common-lisp.net >> http://common-lisp.net/cgi-bin/mailman/listinfo/gsll-devel >> > >

1 0

Re: [Gsll-devel] Introducing "Grid Structured Data"
by Tamas Papp 17 Jan '10

17 Jan '10

On Sat, 16 Jan 2010, Mirko Vukovic wrote: > How would one handle this: > > Suppose I have an vector V of containers C (in this particular case, these > are objects of same type that hold data in matrix form) > v=#(c0 c1 c2 c3 c4) > > I extract a matrix from a C like (m c1), and I want the (aref (m C) 0 5). > > Can xarray let me do (magic-command v 2 0 5) to extract the 0 5 element from > the matrix in (aref v 2) (in this case, c2)? > > I know I can write a function to do that. But does xarray provide a > built-in unified mechanism to allow rows and columns of the stored matrix? > Maybe that is part of your response above? xref is a generic function, dispatching on its first argument. So if you have a vector, that is a subtype of array, so xref will call the method for arrays, which will just signal an error for the extra subscripts. You can deal with this by wrapping your vector in a class, and then dispatching on that, defining a method for xref that will use the extra subscripts to look up elements in the container. Of course, one could change the semantics and deal with extra subscripts by looking stuff up recursively. At first glance, that looks like a "clever trick", ie something appealing in a particular situation, with the possibility of leading to subtle bugs in general. R is full of these. Generally, I am very hostile to "clever" semantics: I always prefer to see an error, instead of the interface guessing that I meant something else. If you still want magic-command, it can be done like this (untested): (defun xref* (object &rest subscripts) (let ((rank (xrank object))) (if (< rank (length subscripts)) (apply #'xref* (apply #'xref object (subseq subscripts 0 rank)) (nthcdr rank subscripts)) (xref object subscripts)))) Tamas

1 0

[Gsll-devel] Introducing "Grid Structured Data"
by Liam Healy 09 Jan '10

09 Jan '10

Based on several expressed wishes for the ability to create, compose and extract array-like objects and pieces of such objects, I have introduced the "Grid Structured Data" collection at http://repo.or.cz/w/gsd.git, and rewritten GSLL to be built on top of that. The GSLL interface is the same as before, but now it is possible to do subarrays, concatenation, slices, transpose, etc. on arrays (both CL arrays and GSLL marrays). There is some documentation for gsd in gsd/documentation/grid/index.html which describes how it works. There is more work to be done, but as it is now, it provides functions that people have asked for to create and manipulate marrays. If you don't need that, you can go on as before and everything is the same. Liam

3 2

[Gsll-devel] patch to export random number distribution symbols
by Russell Kliese 09 Jan '10

09 Jan '10

It appears that the symbols used to select the various random number sample functions are not exported. I was getting "The symbol "UNIFORM" is not external in the GSLL package." when I ran the following code. (clc:clc-require :gsll) (defparameter *rng* (gsl:make-random-number-generator gsl:+default-type+)) (gsl:sample *rng* 'gsl:uniform) I've created a patch to merge into the git repository, but I'm not a Lisp expert, so my solution may be wrong. Thanks to all who have contributed to this very useful library. Russell diff --git a/random/bernoulli.lisp b/random/bernoulli.lisp index 5e6ba48..e3259a7 100644 --- a/random/bernoulli.lisp +++ b/random/bernoulli.lisp @@ -20,6 +20,7 @@ (in-package :gsl) +(export 'bernoulli) (defmfun sample ((generator random-number-generator) (type (eql 'bernoulli)) &key probability) diff --git a/random/beta.lisp b/random/beta.lisp index f830f50..e63ab29 100644 --- a/random/beta.lisp +++ b/random/beta.lisp @@ -20,6 +20,7 @@ (in-package :gsl) +(export 'beta) (defmfun sample ((generator random-number-generator) (type (eql 'beta)) &key a b) "gsl_ran_beta" diff --git a/random/cauchy.lisp b/random/cauchy.lisp index ef8c44d..88913a7 100644 --- a/random/cauchy.lisp +++ b/random/cauchy.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'cauchy) (defmfun sample ((generator random-number-generator) (type (eql 'cauchy)) &key a) "gsl_ran_cauchy" diff --git a/random/chi-squared.lisp b/random/chi-squared.lisp index 4c0b2cc..6f2ed09 100644 --- a/random/chi-squared.lisp +++ b/random/chi-squared.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'chi-squared) (defmfun sample ((generator random-number-generator) (type (eql 'chi-squared)) &key nu) "gsl_ran_chisq" diff --git a/random/dirichlet.lisp b/random/dirichlet.lisp index 17a7668..1618e48 100644 --- a/random/dirichlet.lisp +++ b/random/dirichlet.lisp @@ -22,6 +22,7 @@ ;;; /usr/include/gsl/gsl_randist.h +(export 'dirichlet) (defmfun sample ((generator random-number-generator) (type (eql 'dirichlet)) &key alpha (theta (vdf (dim0 alpha)))) diff --git a/random/discrete.lisp b/random/discrete.lisp index c0bf6c0..c0dd53a 100644 --- a/random/discrete.lisp +++ b/random/discrete.lisp @@ -35,6 +35,7 @@ generally as ``weights'')---the preprocessor will normalize appropriately. This return value is used as an argument to #'discrete.") +(export 'discrete) (defmfun sample ((generator random-number-generator) (type (eql 'discrete)) &key table) diff --git a/random/exponential-power.lisp b/random/exponential-power.lisp index bcc9eee..a5f4844 100644 --- a/random/exponential-power.lisp +++ b/random/exponential-power.lisp @@ -22,6 +22,7 @@ ;;; /usr/include/gsl/gsl_randist.h +(export 'exponential-power) (defmfun sample ((generator random-number-generator) (type (eql 'exponential-power)) &key a b) diff --git a/random/exponential.lisp b/random/exponential.lisp index 574a2af..940497f 100644 --- a/random/exponential.lisp +++ b/random/exponential.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'exponential) (defmfun sample ((generator random-number-generator) (type (eql 'exponential)) &key mu) diff --git a/random/fdist.lisp b/random/fdist.lisp index 9710abf..cfd22c2 100644 --- a/random/fdist.lisp +++ b/random/fdist.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'fdist) (defmfun sample ((generator random-number-generator) (type (eql 'fdist)) &key nu1 nu2) "gsl_ran_fdist" diff --git a/random/flat.lisp b/random/flat.lisp index 5017d88..4841913 100644 --- a/random/flat.lisp +++ b/random/flat.lisp @@ -22,6 +22,7 @@ ;;; /usr/include/gsl/gsl_randist.h +(export 'flat) (defmfun sample ((generator random-number-generator) (type (eql 'flat)) &key a b) "gsl_ran_flat" diff --git a/random/gamma.lisp b/random/gamma.lisp index 0327a04..d02a4c4 100644 --- a/random/gamma.lisp +++ b/random/gamma.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'gamma) (defmfun sample ((generator random-number-generator) (type (eql 'gamma)) &key a b) "gsl_ran_gamma" @@ -37,6 +38,7 @@ is known as the Erlang distribution. The variates are computed using the algorithms from Knuth (vol 2).") +(export 'gamma-mt) (defmfun sample ((generator random-number-generator) (type (eql 'gamma-mt)) &key a b) "gsl_ran_gamma_mt" diff --git a/random/gaussian-bivariate.lisp b/random/gaussian-bivariate.lisp index b00bf77..2c72ecd 100644 --- a/random/gaussian-bivariate.lisp +++ b/random/gaussian-bivariate.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h +(export 'gaussian-bivariate) (defmfun sample ((generator random-number-generator) (type (eql 'gaussian-bivariate)) &key sigma-x sigma-y rho) diff --git a/random/gaussian-tail.lisp b/random/gaussian-tail.lisp index e4881e5..f2b0817 100644 --- a/random/gaussian-tail.lisp +++ b/random/gaussian-tail.lisp @@ -22,6 +22,7 @@ ;;; /usr/include/gsl/gsl_randist.h +(export 'gaussian-tail) (defmfun sample ((generator random-number-generator) (type (eql 'gaussian-tail)) &key a sigma) @@ -50,6 +51,7 @@ for a Gaussian tail distribution with standard deviation sigma and lower limit a, using the formula given for gaussian-tail.") +(export 'ugaussian-tail) (defmfun sample ((generator random-number-generator) (type (eql 'ugaussian-tail)) &key a) diff --git a/random/gaussian.lisp b/random/gaussian.lisp index af4009d..d714a3d 100644 --- a/random/gaussian.lisp +++ b/random/gaussian.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'gaussian) (defmfun sample ((generator random-number-generator) (type (eql 'gaussian)) &key sigma) @@ -48,6 +49,7 @@ "Compute the probability density p(x) at x for a Gaussian distribution with standard deviation sigma.") +(export 'gaussian-ziggurat) (defmfun sample ((generator random-number-generator) (type (eql 'gaussian-ziggurat)) &key sigma) @@ -60,6 +62,7 @@ Marsaglia-Tsang ziggurat method. The Ziggurat algorithm is the fastest available algorithm in most cases.") +(export 'gaussian-ratio-method) (defmfun sample ((generator random-number-generator) (type (eql 'gaussian-ratio-method)) &key sigma) @@ -71,6 +74,7 @@ "Compute a Gaussian random variate using the Kinderman-Monahan-Leva ratio method.") +(export 'ugaussian) (defmfun sample ((generator random-number-generator) (type (eql 'ugaussian)) &key) "gsl_ran_ugaussian" (((mpointer generator) :pointer)) @@ -89,6 +93,7 @@ equivalent to the #'gaussian-pdf with a standard deviation of one, sigma = 1.") +(export 'ugaussian-ratio-method) (defmfun sample ((generator random-number-generator) (type (eql 'ugaussian-ratio-method)) &key) diff --git a/random/generators.lisp b/random/generators.lisp index d0177fb..5e2f968 100644 --- a/random/generators.lisp +++ b/random/generators.lisp @@ -77,6 +77,7 @@ (:documentation "Sample from the probability distribution.")) +(export 'uniform) (defmfun sample ((source random-number-generator) (type (eql 'uniform)) &key) "gsl_rng_uniform" (((mpointer source) :pointer)) :definition :method @@ -91,6 +92,7 @@ (the maximum number of bits that can be portably represented in a single :ulong.") +(export 'uniform>0) (defmfun sample ((source random-number-generator) (type (eql 'uniform>0)) &key) "gsl_rng_uniform_pos" (((mpointer source) :pointer)) :definition :method @@ -102,6 +104,7 @@ type 'uniform until a non-zero value is obtained. You can use this function if you need to avoid a singularity at 0.0.") +(export 'uniform-fixnum) (defmfun sample ((source random-number-generator) (type (eql 'uniform-fixnum)) &key upperbound) diff --git a/random/geometric.lisp b/random/geometric.lisp index 26ba486..e38106b 100644 --- a/random/geometric.lisp +++ b/random/geometric.lisp @@ -22,6 +22,7 @@ ;;; /usr/include/gsl/gsl_randist.h +(export 'geometric) (defmfun sample ((generator random-number-generator) (type (eql 'geometric)) &key probability) diff --git a/random/gumbel1.lisp b/random/gumbel1.lisp index 89b0ca0..fe33b90 100644 --- a/random/gumbel1.lisp +++ b/random/gumbel1.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'geometric) (defmfun sample ((generator random-number-generator) (type (eql 'gumbel1)) &key a b) "gsl_ran_gumbel1" diff --git a/random/gumbel2.lisp b/random/gumbel2.lisp index 13eda47..ea69218 100644 --- a/random/gumbel2.lisp +++ b/random/gumbel2.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'gumbel2) (defmfun sample ((generator random-number-generator) (type (eql 'gumbel2)) &key a b) "gsl_ran_gumbel2" diff --git a/random/hypergeometric.lisp b/random/hypergeometric.lisp index 0771c72..a307cea 100644 --- a/random/hypergeometric.lisp +++ b/random/hypergeometric.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'hypergeometric) (defmfun sample ((generator random-number-generator) (type (eql 'hypergeometric)) &key n1 n2 tt) diff --git a/random/landau.lisp b/random/landau.lisp index 635c4de..087775a 100644 --- a/random/landau.lisp +++ b/random/landau.lisp @@ -22,6 +22,7 @@ ;;; /usr/include/gsl/gsl_randist.h +(export 'landau) (defmfun sample ((generator random-number-generator) (type (eql 'landau)) &key) "gsl_ran_landau" diff --git a/random/laplace.lisp b/random/laplace.lisp index e0089c9..a8d7d37 100644 --- a/random/laplace.lisp +++ b/random/laplace.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'laplace) (defmfun sample ((generator random-number-generator) (type (eql 'laplace)) &key a) "gsl_ran_laplace" diff --git a/random/levy.lisp b/random/levy.lisp index df84970..9a9589b 100644 --- a/random/levy.lisp +++ b/random/levy.lisp @@ -22,6 +22,7 @@ ;;; /usr/include/gsl/gsl_randist.h +(export 'levy) (defmfun sample ((generator random-number-generator) (type (eql 'levy)) &key c alpha) "gsl_ran_levy" @@ -40,6 +41,7 @@ For \alpha < 1 the tails of the distribution become extremely wide. The algorithm only works for 0 < alpha <= 2.") +(export 'levy-skew) (defmfun sample ((generator random-number-generator) (type (eql 'levy-skew)) &key c alpha beta) diff --git a/random/logarithmic.lisp b/random/logarithmic.lisp index dcdc3ff..d8999ef 100644 --- a/random/logarithmic.lisp +++ b/random/logarithmic.lisp @@ -22,6 +22,7 @@ ;;; /usr/include/gsl/gsl_randist.h +(export 'logarithmic) (defmfun sample ((generator random-number-generator) (type (eql 'logarithmic)) &key probability) diff --git a/random/logistic.lisp b/random/logistic.lisp index fed8882..b2ec5b5 100644 --- a/random/logistic.lisp +++ b/random/logistic.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'logistic) (defmfun sample ((generator random-number-generator) (type (eql 'logistic)) &key a) "gsl_ran_logistic" diff --git a/random/lognormal.lisp b/random/lognormal.lisp index eeacb13..6594532 100644 --- a/random/lognormal.lisp +++ b/random/lognormal.lisp @@ -20,6 +20,7 @@ (in-package :gsl) +(export 'lognormal) (defmfun sample ((generator random-number-generator) (type (eql 'lognormal)) &key zeta sigma) diff --git a/random/multinomial.lisp b/random/multinomial.lisp index 5677767..24336b8 100644 --- a/random/multinomial.lisp +++ b/random/multinomial.lisp @@ -22,6 +22,7 @@ ;;; /usr/include/gsl/gsl_randist.h +(export 'multinomial) (defmfun sample ((generator random-number-generator) (type (eql 'multinomial)) &key sum probabilities diff --git a/random/negative-binomial.lisp b/random/negative-binomial.lisp index 4e7350a..bddcb66 100644 --- a/random/negative-binomial.lisp +++ b/random/negative-binomial.lisp @@ -26,6 +26,7 @@ ;;;; Negative binomial ;;;;**************************************************************************** +(export 'negative-binomial) (defmfun sample ((generator random-number-generator) (type (eql 'negative-binomial)) &key probability n) @@ -70,6 +71,7 @@ ;;;; Pascal ;;;;**************************************************************************** +(export 'pascal) (defmfun sample ((generator random-number-generator) (type (eql 'pascal)) &key probability n) diff --git a/random/pareto.lisp b/random/pareto.lisp index e1753e2..d5dea2b 100644 --- a/random/pareto.lisp +++ b/random/pareto.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'pareto) (defmfun sample ((generator random-number-generator) (type (eql 'pareto)) &key a b) "gsl_ran_pareto" diff --git a/random/poisson.lisp b/random/poisson.lisp index 3bb2f2d..36c512d 100644 --- a/random/poisson.lisp +++ b/random/poisson.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'poisson) (defmfun sample ((generator random-number-generator) (type (eql 'poisson)) &key mu) "gsl_ran_poisson" diff --git a/random/rayleigh-tail.lisp b/random/rayleigh-tail.lisp index a848456..3416606 100644 --- a/random/rayleigh-tail.lisp +++ b/random/rayleigh-tail.lisp @@ -22,6 +22,7 @@ ;;; /usr/include/gsl/gsl_randist.h +(export 'rayleigh-tail) (defmfun sample ((generator random-number-generator) (type (eql 'rayleigh-tail)) &key a sigma) diff --git a/random/rayleigh.lisp b/random/rayleigh.lisp index 4855b22..e68fefa 100644 --- a/random/rayleigh.lisp +++ b/random/rayleigh.lisp @@ -22,6 +22,7 @@ ;;; /usr/include/gsl/gsl_randist.h +(export 'rayleigh) (defmfun sample ((generator random-number-generator) (type (eql 'rayleigh)) &key sigma) diff --git a/random/shuffling-sampling.lisp b/random/shuffling-sampling.lisp index 541e2ca..fc26b09 100644 --- a/random/shuffling-sampling.lisp +++ b/random/shuffling-sampling.lisp @@ -22,6 +22,7 @@ ;;; These are currently defined only for vectors. +(export 'shuffle) (defmfun sample ((generator random-number-generator) (type (eql 'shuffle)) &key base) @@ -40,6 +41,7 @@ permutations with equal probability, assuming a perfect source of random numbers.") +(export 'choose-random) (defmfun sample ((generator random-number-generator) (type (eql 'choose-random)) &key src (dest (dim0 src)) @@ -68,6 +70,7 @@ same relative order as those in src. You will need to call with 'shuffle if you want to randomize the order.") +(export 'random-sample) (defmfun sample ((generator random-number-generator) (type (eql 'random-sample)) &key src (dest (dim0 src)) diff --git a/random/spherical-vector.lisp b/random/spherical-vector.lisp index c55ea86..0fa5533 100644 --- a/random/spherical-vector.lisp +++ b/random/spherical-vector.lisp @@ -24,6 +24,7 @@ ;;; No test for #'direction-Nd yet. +(export 'direction-2d) (defmfun sample ((generator random-number-generator) (type (eql 'direction-2d)) &key) "gsl_ran_dir_2d" @@ -36,6 +37,7 @@ two dimensions. The vector is normalized such that |v|^2 = x^2 + y^2 = 1.") +(export 'direction-2d-trig-method) (defmfun sample ((generator random-number-generator) (type (eql 'direction-2d-trig-method)) &key) "gsl_ran_dir_2d_trig_method" @@ -48,6 +50,7 @@ two dimensions. The vector is normalized such that |v|^2 = x^2 + y^2 = 1. Uses trigonometric functions.") +(export 'direction-3d) (defmfun sample ((generator random-number-generator) (type (eql 'direction-3d)) &key) "gsl_ran_dir_3d" @@ -64,6 +67,7 @@ projected along any axis is actually uniform (this is only true for 3 dimensions).") +(export 'direction-Nd) (defmfun sample ((generator random-number-generator) (type (eql 'direction-Nd)) &key vector) "gsl_ran_dir_nd" diff --git a/random/tdist.lisp b/random/tdist.lisp index aa530d5..804b731 100644 --- a/random/tdist.lisp +++ b/random/tdist.lisp @@ -22,6 +22,7 @@ ;;; /usr/include/gsl/gsl_randist.h +(export 'tdist) (defmfun sample ((generator random-number-generator) (type (eql 'tdist)) &key nu) "gsl_ran_tdist" diff --git a/random/weibull.lisp b/random/weibull.lisp index ea00a79..82dce05 100644 --- a/random/weibull.lisp +++ b/random/weibull.lisp @@ -23,6 +23,7 @@ ;;; /usr/include/gsl/gsl_randist.h ;;; /usr/include/gsl/gsl_cdf.h +(export 'weibull) (defmfun sample ((generator random-number-generator) (type (eql 'weibull)) &key a b) "gsl_ran_weibull"

3 2

[Gsll-devel] Contributed patches
by Liam Healy 06 Jan '10

06 Jan '10

I notice there are a few emails recently with patches for GSLL, which I appreciate very much. I will get to them soon I promise. As a general practice if you are using git, if you could, please use git format-patch origin and attach all the files created (which start with a four-digit sequence number). This way, it will get merged into the repository at the right point and the patch author (you!) gets proper credit. This is no criticism of the patches received to date; they are as I say appreciated and I can apply the patches in the diff --git form; it's just a little nicer to using git format-patch and git am. Thanks. Liam

1 0

[Gsll-devel] clisp&cffi: broken foreign functions in saved images
by Sam Steingold 06 Jan '10

06 Jan '10

Hi, I do this: $ clisp -x '(asdf:operate 'asdf:load-op :gsll) (saveinitmem "gsll" :executable t)' $ ./gsll > > (gsll:make-chebyshev 1) FFI::FOREIGN-CALL-OUT: no dynamic object named "gsl_cheb_alloc" in library :DEFAULT this indicates a bug somewhere between cffi and gsll. specifically, the definition of the foreign function gsl_cheb_alloc should specify that it is to be found in "libgsl.so", not in :default. moreover, when I do the (not so) trivial thing: (ffi:default-foreign-language :stdc) (ffi:open-foreign-library "libgslcblas.so") (ffi:def-call-out gsl_cheb_alloc (:library "libgsl.so") (:arguments (n ffi:int)) (:return-type ffi:c-pointer)) (gsl_cheb_alloc 10) ;; ==> #<FOREIGN-ADDRESS #x0000000005503380> (saveinitmem "foo" :executable t) (quit) ./foo (ffi:open-foreign-library "libgslcblas.so") (gsl_cheb_alloc 10) #<FOREIGN-ADDRESS #x0000000001D9F4C0> i.e., if the foreign function definition includes the correct library name, the function will work automatically[*] even in the saved image. So, since I am not quite sure where the bug is hiding, I am reporting this to both gsll and cffi lists. Thanks. Sam. [*] apparently, libgsl.so cannot be loaded before libgslcblas.so is. how do people specify such dependencies?!

2 7