Hazen,
I like the plflush() calls. That helps a lot. However, I am not able
to get the cursor routine to work. Executing your sample script gives
the following result:
snip
Click on the plot
Use of uninitialized value in join or string at plplot_test.pdl line 29,
STDIN line 3.
Use of
I'm not trying to use PDL::Minuit, but during testing of the latest CVS
found that it failed make test. However, running $ perl -Mblib
t/minuit.t looks like it passes (at least I see ok 1, etc). The
relevant output of 'make test' is below.
--Derek
t/matrixops...ok
Someone posted a question on the PDL wiki asking why the PDL 2.4.3
release is marked UNAUTHORIZED on CPAN. I found this thread:
http://www.nntp.perl.org/group/perl.module-authors/2007/05/msg5487.html
which suggests that it has to do with changing release managers and
permissions on PAUSE. I
With the current CVS, when I do
$ perl -e 'use PDL::FFT; use PDL::Slatec;'
I get
Prototype mismatch: sub PDL::fft: none vs (;@) at
/usr/local/lib/perl/5.8.8/PDL/Slatec.pm line 308.
Prototype mismatch: sub main::fft: none vs (;@) at -e line 1
In PDL::Slatec, the relevant code is:
@EXPORT_OK =
Sisyphus wrote:
- Original Message - From: Derek Lamb [EMAIL PROTECTED]
.
.
I didn't have an exceptionally good reason for using File::Temp
here--my primary concern was to get minuit.t to stop dropping a
logfile into the PDL top-level build directory, and also to not sent
Sisyphus wrote:
- Original Message - From: Derek Lamb [EMAIL PROTECTED]
.
.
I think an acceptable solution would be to revert back to the old
behavior, but add an unlink command at the end of the test script.
Heh ... still won't unlink :-)
On winders, perl can't unlink a file
On my Linux machine with Perl 5.8.6 and EU::MM 6.44 and the latest PDL
CVS, make test does the standard test of the t/*.t files and passes, but
then traverses the directory tree looking for other tests to run. It's
failing on Graphics/TriD/OpenGL/test.pl with:
make[3]: Entering directory
Chris Marshall wrote:
With the new requirements for sf.net account
access, I don't think I was able to use anonymous
cvs login for snapshot download the last time
I tried.
If someone could verify the restriction, it may
be time to add a daily cvs snapshot to our pdl
downloads page. The
Sisyphus wrote:
Hi,
In some discussions with Jan Dubois regarding ActiveState ppm builds he
pointed out (wrt PDL-2.4.3):
-- quote --
I noticed one other issue with the PDL Makefile.PL: It will try
to use ExtUtils::F77 before CPAN.pm gets a chance to build and install
it, so it decides
Chris Marshall wrote:
That is the question...
BADVAL has been enabled for automated testing in
the PDL-2.4.3_03 and above developers' releases.
The tests have been passing so far.
It would be possible to make BADVAL the default
setting for PDL-2.4.4. That would make possible
more
working on it now. I'll update the webpage this morning as well.
Derek
Chris Marshall wrote:
PDL-2.4.4 has just been uploaded to CPAN. If someone knows
how to put up a download of the release from sourceforge, I would
appreciate it.
Thanks,
Chris
Some snafu with you being on the project but not having any technician
or admin privileges on the SF tracker. I changed your status so now you
can be assigned tasks, hopefully that allows you to close the bug report
(by changing the Status and Resolution fields). I'll make sure you're
on the
Chris Marshall wrote:
Frossie wrote:
Xavier Calbet wrote on February 26:
I have been trying to edit the wiki pages with no success.
I get the following message:
Could not get main lock: File exists
Am I doing something wrong?
Well, it's kind of a long story. I
Tim Jenness wrote:
Now that sourceforge has git support:
http://apps.sourceforge.net/trac/sourceforge/wiki/Git
what's the feeling on migrating our CVS repository to git? We've just
done a major migration to git here (700 MB of git repositories
created) and whilst it's a bit of a
Tim Jenness wrote:
On Wed, 22 Apr 2009, Craig DeForest wrote:
Sorry I have been so unavailable-- first because I wantedto defer to more
senior admins and then because I am traveling. Once I am I really online
again (tonight) rather than on my phone I intend to mark Tim as an
Chris Marshall wrote:
Tim Jenness wrote:
On Apr 24, 2009, at 9:48 AM, Frossie wrote:
Tim Jenness wrote on April 24:
On Fri, 24 Apr 2009, Derek Lamb wrote:
In the sorry I'm late to the party category: are the other CVS
repos still around somewhere
Frossie wrote:
Tim Jenness wrote on May 21:
it's back. That currently means that the old PDL CVS is available as
well as the book and web pages.
Somebody had better replace the CVS link on the PDL with the git one
though, or some poor bugger will get confused when looking for the
Karl Glazebrook wrote:
No let's NOT do that!
I actually think git/cvs is actually total over-kill for a set of WWW pages
and is not the best approach. The only reason they were put there originally
was simply to put them in a public place where people can edit, but git/cvs
presents a
Derek Lamb wrote:
Karl Glazebrook wrote:
No let's NOT do that!
I actually think git/cvs is actually total over-kill for a set of WWW pages
and is not the best approach. The only reason they were put there originally
was simply to put them in a public place where people can edit
On Fri, 2010-01-01 at 13:23 -0500, Chris Marshall wrote:
Derek, let me know if you need anything
else to update the web page.
nope. all done!
Maybe you
could document what is needed on the wiki
so others could help out.
could do, but it's mostly download the WWW module from cvs and
In February 2009 there was a discussion on the perldl list about the
behavior of rint. One thing that Rob pointed out was that the GNU
implementation of rint used by math.pd does not perform as expected on
all half-integer values. In particular, even half-integers
On 05/17/2010 11:07 AM, Daniel Carrera wrote:
On Mon, 17 May 2010 09:43:52 -0500
David Mertensdcmertens.p...@gmail.com wrote:
On Mon, May 17, 2010 at 8:36 AM, Matthew Kenworthy
kenwor...@strw.leidenuniv.nl wrote:
I'd be good with pulling the website over to git, though.
On 07/08/2010 02:01 PM, Daniel Carrera wrote:
Since there don't seem to be any major issues with the reorgnized
site, I've gone ahead and pushed the changes to the main site. In
addition to the suggestions from Matt and David, I made a few further
improvements to the Reference section, but there
On Thu, 2010-08-12 at 22:30 +0200, Daniel Carrera wrote:
I have one more question: What will happen if I add and remove some
files in Basic/Pod? A quick test suggests that this isn't a problem.
Even the Index.html file seems correct. Is there anything I should
keep in mind when I add/delete
(1) is fixed. It brings you to PDL::Index now. From there you can get to
PDL::Tutorials, QuickStart, etc.
(2) doesn't throw the error for me, but it doesn't show the tacks on the map
either. Maybe Puneet knows more?
I asked Daniel to write such documentation, mentioning that we don't all
(a) should be fixed the next time I do a docs upload (tonight, perhaps?).
(b) is fixed.
On Oct 6, 2010, at 2:38 PM, Douglas Burke wrote:
I just noticed that
a)
The Image Processing section at
http://pdl.perl.org/?docs=Modulestitle=PDL::Modules#image_processing
ends with the
Hi all,
To clarify, there are two bugs outstanding:
1) plplot.t segfaulting (or more) on linux, windows (I haven't tried Mac),
after the recent constructor bug fix (between 2.4.7_002 and _003).
2) plplot.t segfaulting on Mac OS X ONLY when plshades is called, and this was
reported by Puneet
yeah, sorry, that was a lazy bug report on my part. Christians original
message, his own response with a patch, and your OK (except for a couple casts
I think) starts here:
http://mailman.jach.hawaii.edu/pipermail//pdl-porters/2010-August/003189.html
Derek
On Mar 9, 2011, at 4:22 AM,
Can somebody with admin access kill this git repository? All web development
is taking place in pdl-www.
thanks,
Derek
___
PDL-porters mailing list
PDL-porters@jach.hawaii.edu
http://mailman.jach.hawaii.edu/mailman/listinfo/pdl-porters
Hi folks,
You can see the changes to the website at http://pdl.perl.org/future. Tonight
or tomorrow morning I will push that to the main site. So any
questions/comments/changes should be sent to me ASAP. The only thing that
remains to be done from my list is upload the 2.4.8 docs, which I'm
I just noticed that the default PDL download
(http://sourceforge.net/projects/pdl/) for OS X is still 2.4.7, presumably
because SciPDL hasn't been built for 2.4.8 yet. Is someone working on that? I
don't see any instructions for doing that anywhere--perhaps we can put them on
the wiki? The
Hi Daniel,
I think your question reveals a misunderstanding of how PDL development
proceeds as compared to other (much larger) open source projects. There's no
governing board or core people who make the decisions beyond what you see on
this list. There's no official Road Map. Right now
On my Mac OS 10.6.7, Perl 5.12.3 via macports, and all external libraries
enabled (PGPLOT, PLplot, GSL, Proj, Minuit, basically and by definition
anything that shows up in the pdl.perl.org online docs) I get no more test
failures than normal with the current git.
Derek
On Jun 26, 2011, at
moving this question pdl-porters, since it's devel-centric.
Doug probably just forgot to push. This reminds me--there is a
pdl-graphics-plplot repo that I believe David started with the idea of having
that be the first module completely factored out of the source tree. That repo
hasn't been
Perl can not find the Inline module on my system:
$ perl -MInline -e '1;'
Can't locate Inline.pm in @INC (@INC contains: blah blah blah
When I run PDL's Makefile.PL, it tells me:
Checking if your kit is complete...
Looks good
Warning: prerequisite Inline 0.43 not found.
Writing Makefile for
On Sep 10, 2011, at 7:43 AM, chm wrote:
On 9/10/2011 9:01 AM, David Mertens wrote:
Of all of the kitchen sink that constitutes PDL, the Inline requirement
demands the most from CPAN (unless Text::Balanced takes more). It is also a
very important component of PDL, a major differentiating
I've never seen a stock Perl on a Mac with pthreads enabled. For example:
$ perl -MConfig -e 'print $Config::Config{libs},\n;'
-lgdbm -ldbm -ldl -lm -lutil -lc
Notice no -lpthread in there. I do have /usr/include/pthread.h and
/usr/lib/libpthread.dylib, but Perl doesn't know about them, as
of redoing
the config time checks to operate more correctly.
Thanks,
Chris
On Fri, Nov 18, 2011 at 1:54 PM, Derek Lamb de...@boulder.swri.edu wrote:
I've never seen a stock Perl on a Mac with pthreads enabled. For example:
$ perl -MConfig -e 'print $Config::Config{libs},\n;'
-lgdbm -ldbm -ldl
wrote:
- Original Message - From: Derek Lamb
The last PLplot test Fails to crash with POSIX threads still fails.
What's the nature of the failure ? Is the size of the file greater than
600_000 ?
Is fork() implemented on Mac perls as a proper Unix fork ?
On Windows the same test
As described in the ExtUtils::MakeMaker docs here:
http://search.cpan.org/~mschwern/ExtUtils-MakeMaker/lib/ExtUtils/MakeMaker.pm#Module_Meta-Data_(META_and_MYMETA),
EU::MM will generate MYMETA.yml and MYMETA.json files. Since PDL has many
many Makefile.PL's (110!), these get generated for
generate
all those files but I didn't know what changed or
what to do about it. Please fix in git
--Chris
On 12/21/2011 6:25 PM, Derek Lamb wrote:
As described in the ExtUtils::MakeMaker docs here:
http://search.cpan.org/~mschwern/ExtUtils-MakeMaker/lib/ExtUtils/MakeMaker.pm#Module_Meta
:
Does the fix still work with EU::MM 6.31 that we recently upgraded
to as the minimum version number? If not, you may need to add
a conditional check.
--Chris
On Thu, Dec 22, 2011 at 1:57 AM, Derek Lamb de...@boulder.swri.edu wrote:
Looks like that was added in EU::MM 6.57_02 (2010 September
According to Luis Mochan's recent message to the perldl list (March 8th 2012),
this is fixed with Perl 5.15.8 and later. So between Perl 5.10.1 (or maybe .0)
and 5.15.8 there's this problem (in the debugger only?).
Derek
On Apr 11, 2012, at 11:46 PM, Karl Glazebrook wrote:
Hi everyone
I
You don't pull a specific branch. If your copy of the repo is in sync with
origin, then can't you just do a checkout of the new branch?
git branch -a
(lists the branches, including the remote-tracking branches, and puts an * in
front of the current branch you are on--I see a fix-PP-commentary)
Hi David,
I noticed yesterday that there are some funny strings in the documentation for
PDL::Core. Look for the pattern 'PDL::Core.pm' when doing 'pdldoc PDL::Core'.
git-blame says that these crept in with this commit db274c02 to Core.pm.PL on
January 5th of this year--looks to be something
Sorry Rob if this seems elementary:
1) With an up-to-date copy of the PDL git repository, I can launch 'gitk' from
the command line while in that directory and it will show me the commits,
branches, patches, etc. Not sure if that's available on Windows.
2) Or you can click Brows Git from the
With PDL v.2.4.11_001, in the perlDL shell v1.355:
pdl $x1 = xvals(10)
pdl $x2 = xvals(20)
pdl $p = zeroes($x1)
pdl $p-where($x1==4)+=1
pdl $p-where($x2==8)+=2
pdl
and
whereND should give the same results for 1D data.
Derek
On Dec 5, 2012, at 4:31 PM, David Mertens wrote:
Are you proposing that we tone-down the admonition in the docs, or add a
warning to where, or both?
David
On Dec 5, 2012 4:43 PM, Derek Lamb de...@boulder.swri.edu wrote
Hi Rob,
I would be very surprised if the NO_MYMETA line did that on the
PDL::Transform::Proj4 Makefile.PL and on no others. That line is just to
suppress all of the Writing Proj4.yml lines in the Makefile.PL stage (and
also those useless files themselves), and should be the same for about 40
Well, I feel like there should be at least one stick-in-the-mud!
Joel, probably a lot of us could give you this advice: don't get too caught up
in PDL stuff if you're trying to finish your dissertation. There's plenty of
time to screw around with code later.
I don't have a GitHub account yet,
Maggie had that issue with PDL::Stats IIRC. I think her workaround was running
scantree after install. See add_doc.pl (which is referenced in her
Makefile.PL) at http://cpansearch.perl.org/src/MAGGIEXYZ/PDL-Stats-0.6.2/
Note: if you re-install PDL after installing these external modules, it
I saw in the PDL-2.4.11_010 release notes the addition of PDL::IO::IDL for
reading IDL sav files. Cool. There's a note in git that IDL (or whatever
they're called these days) has opened up the IDL sav file license. Do we have
any documentation on that? The only sav files I have contain the
Rob,
With the latest CPAN of PDL::Graphics::Gnuplot I think I got the same error at
test #7, but when I upgraded to the latest git version it passed OK. So if you
still see the error you might want to pull the latest P::G::G from git.
It might not be that PLplot is hanging--the first time I
PLplot did this just last week:
http://sourceforge.net/mailarchive/message.php?msg_id=30601328
Alan had quite the verification process, apparently. But they use SVN.
Verifying git repo integrity should be much simpler--just a sha1 hash of a
pre-conversion clone, and a post-conversion clone,
I'm preparing the HTML documentation to upload to the PDL website. In the past
I have only put up what comes in the main PDL distribution, which seemed to
make sense. But now there are significant modules outside the main
distribution (Stats, Prima, Gnuplot, Simple). Simple is needed on Page
, Mar 26, 2013 at 2:19 PM, Derek Lamb de...@boulder.swri.edu wrote:
I'm preparing the HTML documentation to upload to the PDL website. In the
past I have only put up what comes in the main PDL distribution, which
seemed to make sense. But now there are significant modules outside the
main
I don't have ActivePerl or a Solaris system, but it looks like there are three
sources of build error on Solaris with ActivePerl
(http://code.activestate.com/ppm/PDL/). There are a few timeout errors, which
I imagine will be fixed soon. There is a pthreadBarf test failure on 32-bit
with
On Apr 4, 2013, at 11:02 AM, Craig DeForest wrote:
If I make test with the current PDL git all the slatec tests (e.g.
limits_round.t) skip with PDL::Slatec not available even though Slatec *is*
available. They are failing with a dynamic loading problem (unable to find
libgfortran).
Good catch, Dima!
I had noticed something like this before, but never attributed it to NiceSlice.
I just thought that in my 4000-line file Emacs or Perl lost count or
something, so I would always look in the line or two around the reported error,
which was annoying but apparently never enough
On Apr 12, 2013, at 4:02 PM, Dima Kogan wrote:
Hi.
I'm trying to do something that feels like it should be built-in, but I
can't see a function for it. I have a 2D piddle, and I want to find the
index of the maximum value in this piddle. The index should be a 2-long
piddle or list or
There seems to be a couple problems with viewing our main code repo on the SF
site. I've filed a ticket:
http://sourceforge.net/p/forge/site-support/3729/
Derek
On Apr 15, 2013, at 8:23 AM, Chris Marshall wrote:
There was a problem with the sf.net git migration
of our main repo. I'll let
SF is doing a full repository refresh--I don't know if that will affect push
/ pulls, but if you get unexpected results that might be why. I'll post back
once I hear that it is resolved.
On Apr 21, 2013, at 10:09 PM, Derek Lamb wrote:
There seems to be a couple problems with viewing our main
In PDL::Ops, binary operators have this documentation in =for example (example
for plus):
$c = plus $a, $b, 0; # explicit call with trailing 0
$c = $a + $b; # overloaded call
$a-inplace-plus($b,0); # modify $a inplace
But when I do the first one, I get a usage error:
If I do a make clean, then cd to Basic and grep -iRn magicno *, I get some
stuff from Core/pdl.h.PL, pdlapi.c, pdlcore.c.PL, pdlthread.[c,h], and
Pod/Internals.pod. Does one of these ring a bell?
On May 5, 2013, at 9:02 PM, Craig DeForest wrote:
Could be I'm just being idiotic with
We have in PDL::Core functions to convert datatypes: float(), double(),
longlong(), ushort(), etc. I don't see a function to convert to PDL_Index
type--should there be? I know this is either long or longlong, but I thought
that something like pdl_index() might be useful. Obviously it
+ years ago but limiting
now.
On Thu, Jul 25, 2013 at 11:50 PM, Derek Lamb de...@boulder.swri.edu wrote:
We have in PDL::Core functions to convert datatypes: float(),
double(), longlong(), ushort(), etc. I don't see a function to
convert to PDL_Index type--should there be? I know
caveat: I know nothing about Moose related things yet.
The problem (if it is really) is not with the logical not operation in PDL.
One can do logical not on a piddle just fine (print not xvals(10);). The
problem is from the overload bool that happens in PDL::Core, thus the use of
if is the
Is that going to cause problems if the user turns off badvalue support at build
time?
On Aug 2, 2013, at 3:34 PM, Chris Marshall wrote:
Fix code layout for readability. -chm
On Fri, Aug 2, 2013 at 5:29 PM, Chris Marshall devel.chm...@gmail.com wrote:
It seems we haven't had a consensus on
Karl, Christian, etc (original PDL'ers) please weigh in:
I was looking at Basic.pm over the weekend and saw three functions that were
unfamiliar, perhaps because they are not documented anywhere: sec, ins, and
similar_assign.
1) sec returns a slice of a piddle based on a list of border
I think I can do it (I have all the external dependencies) but I have no idea
how--instructions from Matt K. or KGB would be really helpful!! Then we could
put them on the wiki.
http://sourceforge.net/apps/mediawiki/pdl/index.php?title=Developer%27s_Corner#Building_SciPDL_for_Release
is empty
who can help me, I'd appreciate it!
Cheers,
Matt
On 21 January 2014 21:33, Derek Lamb de...@boulder.swri.edu wrote:
Waiting on instructions from Matt. I'm sure he has notes somewhere but I
can't seem to pry them out of his hands.
MATT!! Even unpolished notes would give me
pointed to PDL source on sourceforge.
Best,
Maggie
On Tue, Jan 21, 2014 at 5:09 PM, Derek Lamb de...@boulder.swri.edu wrote:
thanks! nearly 1000 lines of POD: I'll see what I can do.
On Jan 21, 2014, at 2:36 PM, Matthew Kenworthy wrote:
Hi all,
I've been trying to build
, Chris Marshall wrote:
On Thu, Feb 20, 2014 at 9:30 AM, Derek Lamb de...@boulder.swri.edu wrote:
As an update:
I have a VirtualBox with OS X 10.9 Mavericks cleanly installed, all of the
development tools, and figuring out how to make a Mac package.
Unfortunately,
at some point Apple has
at 12:45 PM, Derek Lamb de...@boulder.swri.edu wrote:
Ignorance, maybe?
On the other hand--An application that installs kernel extensions,
startup items, system-wide preference panes, or any other system-wide
resources cannot be installed [as a bundle] by dragging the application,
since those
The one downside will be short text ads at the bottom of every list email. But
other than that I think it is smart to move them.
While we're at it, does it also make sense to import the pre-2005 archives? We
would have in one place complete archives from pdl-porters back to 1997
February and
ditto on both counts. I have a Virtual Box set up that could do this, if I
could (take the time to) figure out the packaging part of it.
On Oct 25, 2014, at 12:21 PM, Chris Marshall devel.chm...@gmail.com wrote:
That would be great. I'm pretty sure I never got one from you. Thanks,
Chris.
FWIW, generating documentation on the release manager's machine has been
basically the method for generating the docs at pdl.perl.org
http://pdl.perl.org/ for some time: once Chris releases a new version, I
build it, compiling documentation for all the optional-but-included modules,
and upload
At the very beginning of Basic/Ops/ops.pd there are currently two MOD macros to
define the modulus operator. BU_MOD works on byte and ushort piddles, whereas
MOD works on all other types. In both of those, but particularly worrisome in
MOD, there are C typecasts to (int). I think this is the
Hi Karl,
On a relatively clean virtual machine that just got upgraded to Yosemite,
install goes smoothly but when starting the pdl shell I get the following:
snip
ReadLines, NiceSlice, MultiLines enabled
Undefined subroutine Term::ReadLine::Gnu::XS::ornaments called at
this version:
https://dl.dropboxusercontent.com/u/2148080/SciPDL-v2.6-Yosemite.zip
hopefully that will fix the issue.
PDL-porters: please send requests for additional things, e.g. TriD, PLplot...
Karl
On 8 Jan 2015, at 10:16 am, Derek Lamb de...@boulder.swri.edu wrote:
Hi Karl,
Yes
I have not but will try later today. Yes, there was some problem with the
migration of the wiki from mediawiki to the SF allura wiki platform that I
was not able to address. SF wasn't too helpful either. More later.
Derek
On Jan 7, 2015, at 9:45 AM, Zakariyya Mughal zaki.mug...@gmail.com
Derek
That's a bit odd, as you say the subroutine IS defined in the XS.pm file
loaded.
I am wondering if this is triggered by something funny in your ~/.perldlrc
file? Does
perl -e 'use PDL'
work?
Karl
On 6 Jan 2015, at 10:52 am, Derek Lamb de...@boulder.swri.edu wrote:
Hi
82 matches
Mail list logo