From d.trudgett at aamhatch.com.au Wed May 30 06:13:47 2007 From: d.trudgett at aamhatch.com.au (Trudgett, David) Date: Wed, 30 May 2007 16:13:47 +1000 Subject: [cl-mathstats-devel] Losing symbols on system re-load Message-ID: Hi there! Any help or hints to solve this little problem would be greatly appreciated! I'm just starting to use CL-MATHSTATS on Windows CLISP, but I am having a problem with losing symbols from CL-MATHSTATS when the system is reloaded using ASDF. About 30 symbols are lost (105 compared to 135), including ones like MEAN, MEDIAN, STANDARD-DEVIATION, STATISTICAL-SUMMARY (and obviously many others). Deleting the .fas files and re-loading gets back the lost symbols. There are various warnings during the initial compilation, but they don't seem to be related at all: WARNING: RESTART-CASE: restart cannot be invoked interactively because it is missing a :INTERACTIVE option: (USE-VALUE (X) X) and several like this: WARNING in MODE-INTERNAL in lines 319..376 : Unknown declaration VALUES. I fixed the latter type by adding CLISP to lines like: (declare #-(or DIGITOOL OPENMCL CLISP) (values g-score g-significance dof)) The CL-MATHSTATS symbol list (obtained with SLIME's APROPOS PACKAGE) with the 30 or so symbols missing is as follows: CL-MATHSTATS:+0DEGREES+ CL-MATHSTATS:+10DEGREES+ CL-MATHSTATS:+120DEGREES+ CL-MATHSTATS:+135DEGREES+ CL-MATHSTATS:+150DEGREES+ CL-MATHSTATS:+15DEGREES+ CL-MATHSTATS:+180DEGREES+ CL-MATHSTATS:+210DEGREES+ CL-MATHSTATS:+225DEGREES+ CL-MATHSTATS:+240DEGREES+ CL-MATHSTATS:+270DEGREES+ CL-MATHSTATS:+300DEGREES+ CL-MATHSTATS:+30DEGREES+ CL-MATHSTATS:+315DEGREES+ CL-MATHSTATS:+330DEGREES+ CL-MATHSTATS:+360DEGREES+ CL-MATHSTATS:+45DEGREES+ CL-MATHSTATS:+5DEGREES+ CL-MATHSTATS:+60DEGREES+ CL-MATHSTATS:+90DEGREES+ CL-MATHSTATS:+E+ CL-MATHSTATS:2FPI CL-MATHSTATS:AUTOCORRELATION CL-MATHSTATS:BETA CL-MATHSTATS:BETA-INCOMPLETE CL-MATHSTATS:BINOMIAL-CDF CL-MATHSTATS:BINOMIAL-CDF-EXACT CL-MATHSTATS:BINOMIAL-COEFFICIENT CL-MATHSTATS:BINOMIAL-COEFFICIENT-EXACT CL-MATHSTATS:BINOMIAL-PROBABILITY CL-MATHSTATS:BINOMIAL-PROBABILITY-EXACT CL-MATHSTATS:CHI-SQUARE-SIGNIFICANCE CL-MATHSTATS:COMBINATION-COUNT CL-MATHSTATS:CONFIDENCE-INTERVAL-T-SUMMARIES CL-MATHSTATS:CONVERT CL-MATHSTATS:CORRELATION CL-MATHSTATS:CORRELATION-FROM-SUMMARIES CL-MATHSTATS:CORRELATION-MATRIX CL-MATHSTATS:CROSS-CORRELATION CL-MATHSTATS:CROSS-PRODUCT CL-MATHSTATS:DEGREES->RADIANS CL-MATHSTATS:DIV2 CL-MATHSTATS:DOT-PRODUCT CL-MATHSTATS:ENSURE-FLOAT CL-MATHSTATS:ERROR-FUNCTION CL-MATHSTATS:ERROR-FUNCTION-COMPLEMENT CL-MATHSTATS:EXP2 CL-MATHSTATS:EXTRACT-UNIQUE-VALUES CL-MATHSTATS:F-MEASURE CL-MATHSTATS:F-SIGNIFICANCE CL-MATHSTATS:FACTORIAL CL-MATHSTATS:FACTORIAL-EXACT CL-MATHSTATS:FACTORIAL-LN CL-MATHSTATS:FPI CL-MATHSTATS:GAMMA-INCOMPLETE CL-MATHSTATS:GAMMA-LN CL-MATHSTATS:GAUSSIAN-CDF CL-MATHSTATS:GAUSSIAN-SIGNIFICANCE CL-MATHSTATS:LAGGED-CORRELATION CL-MATHSTATS:LINEAR-REGRESSION-BRIEF CL-MATHSTATS:LINEAR-REGRESSION-BRIEF-SUMMARIES CL-MATHSTATS:LINEAR-REGRESSION-MINIMAL CL-MATHSTATS:LINEAR-REGRESSION-MINIMAL-SUMMARIES CL-MATHSTATS:LINEAR-REGRESSION-VERBOSE CL-MATHSTATS:LINEAR-REGRESSION-VERBOSE-SUMMARIES CL-MATHSTATS:LINEAR-SCALE CL-MATHSTATS:LOG2 CL-MATHSTATS:MATRIX-MULTIPLY CL-MATHSTATS:MATRIX-TRACE CL-MATHSTATS:MOD2 CL-MATHSTATS:MULTIPLE-LINEAR-REGRESSION-ARRAYS CL-MATHSTATS:MULTIPLE-LINEAR-REGRESSION-BRIEF CL-MATHSTATS:MULTIPLE-LINEAR-REGRESSION-MINIMAL CL-MATHSTATS:MULTIPLE-LINEAR-REGRESSION-NORMAL CL-MATHSTATS:MULTIPLE-LINEAR-REGRESSION-VERBOSE CL-MATHSTATS:NORMALIZE-MATRIX CL-MATHSTATS:ON-INTERVAL CL-MATHSTATS:PARTIALS-FROM-PARENTS CL-MATHSTATS:PERMUTATION-COUNT CL-MATHSTATS:POISSON-CDF CL-MATHSTATS:R-SCORE CL-MATHSTATS:RADIANS->DEGREES CL-MATHSTATS:ROUND-TO-FACTOR CL-MATHSTATS:SAFE-EXP CL-MATHSTATS:SCHEFFE-TESTS CL-MATHSTATS:SMOOTH-4253H CL-MATHSTATS:SMOOTH-HANNING CL-MATHSTATS:SMOOTH-MEAN-2 CL-MATHSTATS:SMOOTH-MEAN-3 CL-MATHSTATS:SMOOTH-MEAN-4 CL-MATHSTATS:SMOOTH-MEAN-5 CL-MATHSTATS:SMOOTH-MEDIAN-2 CL-MATHSTATS:SMOOTH-MEDIAN-3 CL-MATHSTATS:SMOOTH-MEDIAN-4 CL-MATHSTATS:SMOOTH-MEDIAN-5 CL-MATHSTATS:SQUARE CL-MATHSTATS:STUDENTS-T-SIGNIFICANCE CL-MATHSTATS:SUM-OF-ARRAY-ELEMENTS CL-MATHSTATS:TIMES2 CL-MATHSTATS:TRANSPOSE-MATRIX CL-MATHSTATS:TRUNC2 CL-MATHSTATS:TRUNCATE-TO-FACTOR CL-MATHSTATS:UNDERFLOW-GOES-TO-ZERO CL-MATHSTATS:WITH-TEMP-TABLE CL-MATHSTATS:WITH-TEMP-VECTOR I haven't worked out an exact list of what is missing, yet, but can do so if that might help find the cause of the problem. Thanks, David Trudgett Attention Email Disclaimer Notice - This message is the property of AAMHatch. The information in this email is confidential and may be legally privileged. It is intended solely for the addressee. Access to this email by anyone else is unauthorised. If you are not the intended recipient, any disclosure, copying, distribution or any action taken or omitted to be taken in reliance on it is prohibited and may be unlawful. If you have received this message in error please notify AAMHatch immediately via email to mailadmin at aamhatch.com.au This email has been scanned and cleared by NetIQ Mail Marshal, however AAMHatch does not guarantee this message free of viruses, or interference. ________________________________ From gwking at metabang.com Wed May 30 13:52:13 2007 From: gwking at metabang.com (Gary King) Date: Wed, 30 May 2007 09:52:13 -0400 Subject: [cl-mathstats-devel] Losing symbols on system re-load In-Reply-To: References: Message-ID: Hi David, > Deleting the .fas files and re-loading gets back the lost symbols. This is what we call "an odd thing". Looking at the list you appended and revisiting the code leads me to suspect that the missing symbols are mostly (entirely) from the define-statistic macro. Which, I now see, expands into > (progn (eval-when (:compile-toplevel) (export 'mean)) > (defclass mean (simple-statistic) nil) > ... So that the symbol is only exported when the file is compiled ... I have _no_ idea how this has ever worked. Sigh. I'll be posting an update that corrects this shortly. (P.S., as for the values, I decided to just remove them instead of added clisp to the list...) -- Gary Warren King, metabang.com Cell: (413) 885 9127 Fax: (206) 338-4052 gwkkwg on Skype * garethsan on AIM From d.trudgett at aamhatch.com.au Wed May 30 23:57:34 2007 From: d.trudgett at aamhatch.com.au (Trudgett, David) Date: Thu, 31 May 2007 09:57:34 +1000 Subject: [cl-mathstats-devel] Losing symbols on system re-load Message-ID: Hi Gary, Thanks for your quick reply! Yes, I was doing some more poking around last night, and was able to fix the problem by adding the following exports to the package.lisp file: (:export #:data-length #:mean #:sum-of-squares #:variance #:standard-deviation #:skewness #:minimum #:maximum #:range #:quantile #:median #:trimmed-mean #:mode #:mode-for-continuous-data #:data-continuous-p #:smart-mode #:multiple-modes #:interquartile-range #:tukey-summary #:statistical-summary #:significance #:t-significance #:t-test-one-sample #:t-test #:d-test #:t-test-matched #:z-test-one-sample #:inner-product #:covariance #:confidence-interval #:confidence-interval-z #:confidence-interval-z-summaries #:confidence-interval-t #:confidence-interval-t-summaries #:confidence-interval-proportion #:scheffe-tests #:print-scheffe-table #:cross-product #:dot-product #:r-score #:difference-list #:sum-list #:chi-square-2x2-counts #:chi-square-2x2 #:chi-square-rxc-counts #:make-contingency-table #:chi-square-rxc #:g-test #:find-critical-value) (:export #:anova-one-way-variables #:anova-one-way-groups #:print-anova-table #:anova-two-way-groups #:make-3d-table #:anova-two-way-variables-unequal-cell-sizes #:anova-two-way-variables) (:export #:scalar-matrix-multiply #:multiply-matrices #:invert-matrix #:matrix-norm #:invert-matrix-iterate #:transpose-matrix #:normalize-matrix #:sum-of-array-elements) (:export #:reduce-matrix) (:export #:pythag-df #:svbksb-df #:svdcmp-df #:svzero-df #:svd-inverse-slow-df #:svd-inverse-fast-df #:pythag-sf #:svbksb-sf #:svdcmp-sf #:svzero-sf #:svd-inverse-slow-sf #:svd-inverse-fast-sf #:singular-value-decomposition #:svd-zero #:svd-back-substitute #:svd-solve-linear-system #:svdvar #:svd-matrix-inverse) Of course, whether all of these are supposed to be exported or not, I don't know. I'm also not sure if I duplicated any. > This is what we call "an odd thing". Yes, I couldn't think of a reason for it at the time. > So that the symbol is only exported when the file is compiled ... I > have _no_ idea how this has ever worked. Sigh. Almost qualifies for a bar or two of the Twilight Zone theme... :-) Anyway, thanks for your help. I'll check back in a few days or a week for the new source to keep my stuff in sync. Cheers, David -----Original Message----- From: Gary King [mailto:gwking at metabang.com] Sent: Wednesday, 30 May 2007 23:52 To: Trudgett, David Cc: 'cl-mathstats-devel at common-lisp.net' Subject: Re: [cl-mathstats-devel] Losing symbols on system re-load Hi David, > Deleting the .fas files and re-loading gets back the lost symbols. This is what we call "an odd thing". Looking at the list you appended and revisiting the code leads me to suspect that the missing symbols are mostly (entirely) from the define-statistic macro. Which, I now see, expands into > (progn (eval-when (:compile-toplevel) (export 'mean)) > (defclass mean (simple-statistic) nil) > ... So that the symbol is only exported when the file is compiled ... I have _no_ idea how this has ever worked. Sigh. I'll be posting an update that corrects this shortly. (P.S., as for the values, I decided to just remove them instead of added clisp to the list...) -- Gary Warren King, metabang.com Cell: (413) 885 9127 Fax: (206) 338-4052 gwkkwg on Skype * garethsan on AIM Attention Email Disclaimer Notice - This message is the property of AAMHatch. The information in this email is confidential and may be legally privileged. It is intended solely for the addressee. Access to this email by anyone else is unauthorised. If you are not the intended recipient, any disclosure, copying, distribution or any action taken or omitted to be taken in reliance on it is prohibited and may be unlawful. If you have received this message in error please notify AAMHatch immediately via email to mailadmin at aamhatch.com.au This email has been scanned and cleared by NetIQ Mail Marshal, however AAMHatch does not guarantee this message free of viruses, or interference. ________________________________