[Pdl-devel] PDL 2.026 released to CPAN

Dear PDL users, Ed recently released PDL 2.026 to CPAN. (https://metacpan.org/release/PDL) There are several notable changes since the last announcement to this email list, of PDL 2.020 on 2019-12-11. Thanks go to @mohawk2 (Ed), @wmlb (Luis Mochán), @newville (Matt Newville), and @sebastic (B

[Pdl-devel] problems with PDL on ActivePerl

After updating the install webpage for SciPDL I was looking at updates to the instructions for other operating systems. Starting with Windows. I'm going to include the not-so-new development that Strawberry Perl has a PDL-specific edition. ActivePerl is another matter. According to https://

Re: [Pdl-devel] PDL and its pre-req PerlIO::Layers fail to build in macOS 10.13.3

Hi Will, This doesn't really seem to be a PDL issue. And you don't say whether you've been able to install on prior versions of macOS. But my PerlIO::Layers installed from MacPorts just fine on macOS 10.12.x (Sierra), and it doesn't look like MacPorts does anything special to make it work. A

Re: [Pdl-devel] unstable SF git servers and github mirroring

Hi Christian, If the developers implemented the instructions to push to both SourceForge and GitHub at the same time described here (https://sourceforge.net/p/forge/documentation/GitHub%20Importer/#source-code ), wou

[Pdl-devel] Help debugging warp2d on Windows

If you run PDL on Windows, could you try building and running tests on the fix_warp2d branch? It passes tests on my Mac, but is failing our automated test suite on AppVeyor using Strawberry Perl 5.26.1.1. In particular t/image2d.t is failing with: > Free to wrong pool 39a0f0 not 8127396a153af

Re: [Pdl-devel] 'cat' output datatype based on first input

> On Jul 17, 2017, at 11:13 AM, Derek Lamb wrote: > > That way cat(@list_of_piddles)->dog would actually be a symmetric operation. I guess I should clarify that I mean that this would be symmetric for the values returned, but the datatypes of the piddles returned by dog would of

[Pdl-devel] 'cat' output datatype based on first input

If we create a few piddles of different types and cat them together, the output datatype is the same as the first piddle: ~~~perl $by=xvals(byte,5)+253; $so=xvals(short,5)+32766; $lo=xvals(long,5)+32766; $fl=float(xvals(5)+0.2); $c1 = cat($by,$so,$lo,$fl); $c2 = cat($lo,$so,$fl,$by); $c3 = cat($f

Re: [Pdl-devel] Question on the constructor

> On Jul 16, 2016, at 9:07 PM, Zakariyya Mughal wrote: > > On 2016-07-16 at 17:07:53 -0600, Derek Lamb wrote: >> Though I am not too familiar with the constructor code, the fact that it >> passes all tests with all combinations of BADVAL_USENAN and/or BADVAL_PERPDL &

Re: [Pdl-devel] Question on the constructor

th this > "bitwise equality" ( == or !finite(...) ) as opposed to > just "numeric equality" (==). > > I will squash it before merging. > > Cheers, > - Zaki Mughal > > On 2016-06-27 at 21:53:57 -0600, Craig DeForest wrote: >> Aw crud. I will have a l

Re: [Pdl-devel] Question on the constructor

FWIW the commit d12d7b1 "fix skip count for no-badvals case" (which I take to be the end of the line on this topic, since it has been merged into master) is not passing test t/constructor.t when BADVAL_USENAN and/or BADVAL_PER_PDL is true: see https://travis-ci.org/PDLPorters/pdl/builds/13740508

[Pdl-devel] PP Pars modifiers

The docs for PDL::PP talk about two modifiers in the Pars section: [o] designates an output piddle [t] designates a temporary piddle In the PDL source tree I also see [nc] [oca] [phys] I could trudge through the source to figure it out I suppose, but maybe somebody could just explain those thr

Re: [Pdl-devel] sourceforge

Apparently there's been an acquisition, some soul searching, and maybe SF is heading down a better path now: https://www.reddit.com/r/sysadmin/comments/4n3e1s/the_state_of_sourceforge_since_its_acquisition_in/ > On Jun 12, 2016, at 9:08 PM, Craig DeForest wrote: > > Hear, hear! I’m all for it.

Re: [Pdl-devel] SOLVED: Problem building PDL-Drawing-Prima-0.11

, I'll > need to update the examples in the PDL::PP chapter of the book. > > David > > On Sun, Jun 12, 2016 at 10:18 AM, Derek Lamb <mailto:de...@boulder.swri.edu>> wrote: > The fact that pp_done() was called automatically by PP was never a documented > fe

Re: [Pdl-devel] SOLVED: Problem building PDL-Drawing-Prima-0.11

The fact that pp_done() was called automatically by PP was never a documented feature—in fact it was directly contradicted by the documentation. Sorry to cause this headache—I didn't get any comments yea or nay on the bug report or before the branch was merged into master. Though probably you

Re: [Pdl-devel] 20th Anniversary OMG

Karl, Do you happen to have laying around anywhere a copy of the original perlDL article? The Perl Journal stopped publication ~10 years ago and Dr. Dobb's Journal website has been hosting the articles. However the Dr. Dobb's Journal website ceased updates at the end of 2014. So it might be

Re: [Pdl-devel] cleanups

$cr() = (ar+tt*ai)/dn; cheers, Derek > On Jun 3, 2016, at 10:44 PM, Derek Lamb wrote: > > 1) was addressed in the deprecate-finite branch that was recently merged into > master post-2.016. > > 3) is addressed in the types-printconv branch that has not yet been

Re: [Pdl-devel] cleanups

1) was addressed in the deprecate-finite branch that was recently merged into master post-2.016. 3) is addressed in the types-printconv branch that has not yet been merged—needs some code review to make sure it will pass on 32-bit and 64-bit platforms with different values of ivsize, etc. It h

[Pdl-devel] tautological out-of-range comparison in PDL::Bad

I get the following error (16 times; 2 for each datatype) from Bad.xs when building PDL with clang on my Mac: > Bad.xs:9487:19: warning: comparison of constant -1 with expression of type > 'pdl_datatypes' is always true > [-Wtautological-constant-out-of-range-compare] > if ( val.type !

[Pdl-devel] unhelpful error when pp_def'ing BadCode section

I'm writing some new code to deal handle the new GSL SF Legendre functions (see https://sourceforge.net/p/pdl/bugs/407/ ). This is my first time really diving into PP. The signature of the new function is: Signature: gsl_sf_legendre_array(double x(); double [o]y(n); char norm; int lmax; int

Re: [Pdl-devel] Constructing multi-level perl list from a PDL

yes: use 'unpdl'. > On Apr 14, 2016, at 1:04 PM, dhunt wrote: > > > Hi all: I've got what might seem like a silly question: Is there any > PDL routine which will > convert an N-dimensional PDL to a corresponding N-dimensional perl list > structure? > > The 'pdl' constructor does the inverse

[Pdl-devel] code review request (potential license issue)

I created a pull request (from branch lmfit) on GitHub earlier this week for some changes to the PDL::Fit::LM module. Most (all?) of the changes to the module itself are documentation. But since that module did not have any tests in the suite, I added t/lm.t. If one is going to test the fitti

Re: [Pdl-devel] Mac: New SciPDL with PDL-2.015 for preliminary testing

Hi Karl, I am testing this on my Mac, OS 10.10.5 (Yosemite). The results were mixed. Unpacking went fine. > Example commands for install (for hardcore unix people only, mistypes can be > dangerous!): > > % cd / > % tar xvfz /path/to/tarball.tar.gz > > > And to test: > > % source /Applicat

Re: [Pdl-devel] question about github / SF interoperability

nd an FF-only merge, not with a merge > commit :-) You'll then be able to tidily delete your sf#396 branch, on the SF > repo, which will then get reflected on the GH repo in due course. > > Ed > > -Original Message- From: Derek Lamb > Sent: Friday, September 18,

[Pdl-devel] question about github / SF interoperability

Question about PDL's GitHub / SF interoperability. A couple weeks ago I pushed the branch sf#396 to SF. Soon thereafter it showed up on GitHub. I then initiated a pull requested from GitHub. After a while it passed all the checks, so I let it sit for a week or so. Seeing no up- or down-vote

[Pdl-devel] demo cartography failure

I'm getting a failure when doing 'demo cartography' from the perldl shell with 2.00716 installed. It fails on the first line: $coast = earth_coast() -> glue( 1, graticule(15,1) ); OOPS!!! Something went wrong, please make a bug report!: PDL: PDL::Primitive::append(a,b,c): Parameter 'b': Mism

Re: [Pdl-devel] 2.007_16 has a bit too much 'use warnings'

please ignore the '#comment' gripe: that's actually a documented feature that I completely forgot about. > On Apr 20, 2015, at 5:07 PM, Derek Lamb wrote: > > So finally instead of just trying builds, I actually installed 2.007_16 and > started using it today. >

[Pdl-devel] 2.007_16 has a bit too much 'use warnings'

So finally instead of just trying builds, I actually installed 2.007_16 and started using it today. I do not like it: too many warnings. Let's start a shell: > derek@turul:~$ pdl > Name "PERLDL::HISTFILESIZE" used only once: possible typo at > /Users/derek/local/bin/perldl line 21. > Name "PDL

Re: [Pdl-devel] CPAN clean install 2.008-rc1 fails, missing Module::Compile

or some time (after I badgered Ingy to fix & release it) and works > for me on Linux and Windows. > > -Original Message- From: Derek Lamb > Sent: Thursday, April 16, 2015 10:18 PM > To: pdl-devel > Subject: Re: [Pdl-devel] CPAN clean install 2.008-rc1 fails,miss

Re: [Pdl-devel] CPAN clean install 2.008-rc1 fails, missing Module::Compile

As a follow-up, the problem was that Module::Compile did not pass its own tests, so didn't install. The PDL build saw this, but proceeded anyway. I've posted the issue to Module::Compile's Github issue tracker. > On Apr 15, 2015, at 3:05 PM, Derek Lamb wrote: > > I a

[Pdl-devel] CPAN clean install 2.008-rc1 fails, missing Module::Compile

I attempted a clean install of 2.008-rc1 on a OS X 10.10.1 VM, via CPAN. I completely cleaned ~/.cpan and started a new local::lib. 'make' works OK. 'make test' passes, but complained about unsatisfied "REQUIRED prerequisites" (are there any other kind?), specifically Module::Compile was miss

[Pdl-devel] 2 slices.pd questions

1) In the current git, where index() is pp_def'd in slices.pd, there is a funny $a =BadCode => at line 239 that I don't understand. It was introduced by this commit [1]. My guess is the $a is wrong but harmless, but I wanted to check. 2) Also at line 3238 of slices.pd there is a declarati

Re: [Pdl-devel] Proj presently broken

Hi Ed, I found a little time to look at this again. I activated proj @4.9.1 in Macports: $ port installed proj The following ports are currently installed: proj @4.8.0_1 proj @4.9.1_0 (active) 1) On branch master at commit d9a474e, during Makefile.PL, I get Building PDL::GIS::Proj. Turn o

Re: [Pdl-devel] Proj presently broken

Ditto: building Lib/Transform/Proj4 without Alien is causing me difficulty, too. This is with Rob's latest revert commit (931dcc). My case is different: I got my Proj4 from Macports. Just upgraded to the new version 4.9.1, and PDL started throwing errors during make. Two items of note from t

[Pdl-devel] branch clang_warnings

I just created a branch clang_warnings that is focused on squashing the sometimes nit-picky warnings clang generates when PDL is compiled. In fact, probably >90% of the warnings have already been squashed. Before rebasing onto master, I would appreciate tests and code review from both clang an

Re: [Pdl-devel] t/ops.t fail on 64-bit Strawberry Perl

Ed: I am fine TODO-ing those failing tests. I see you've already done that. Thanks. Chris: My hunch is that it doesn't have anything to do with 64-bit index support, rather just what datatypes can hold how big of values on what kind of Perls. My first idea was to TODO those tests for Perls w

Re: [Pdl-devel] [Pdl-general] CHM/PDL-2.007_12.tar.gz released to CPAN

Hi Rob, Somewhere in that commit message I think was something like "I have no idea what will happen on a 32-bit system. Good luck!" So obviously I hit something there. Will look into it in the next day or two. cheers, Derek > On Mar 6, 2015, at 9:08 PM, sisyph...@optusnet.com.au wrote: > >

Re: [Pdl-devel] fitpoly1d()

Hi Karl, Maybe if we had a test for threaded fitpoly1d we would have caught this 4.5 years ago! Looks like the problem is with the change from PDL::Slatec::matinv to PDL::MatrixOps::inv [8adb0b] . Running your ex

Re: [Pdl-devel] PDL::NiceSlice engine is slow (again)

On Feb 26, 2015, at 5:12 AM, Chris Marshall wrote: > > perldlpp.pl doesn't do the Filter::Simple filtering, just classic... oh, ok. I guess I didn't look at it that closely. Now that I am, I don't see how that happens, since that script just calls PDL::NiceSlice::perldlpp, which is in NiceSl

Re: [Pdl-devel] PDL::NiceSlice engine is slow (again)

derek/Build/pdl-code/blib/lib/PDL/Graphics/PGPLOT/Window.pm Looks sane to me. > On Feb 25, 2015, at 5:44 PM, Ed wrote: > > Pulling in installed, working version? > > From: Derek Lamb <mailto:de...@boulder.swri.edu> > Sent: Thursday, February 26, 2015 12:32 AM > To: p