...and should be appearing at a CPAN mirror near you soon.
This CPAN developers release (a.k.a. PDL 2.016 release
candidate 1) includes the code changes expected for the
PDL-2.016 release. Please test with your configuration
and report any issues or problems needed fixing (with
solutions if possible).
Assuming, CPAN Testers reports go well, I expect to
update the release notes, changes, and other readme
type information and the release candidate will be the
basis for PDL-2.016.
As always, documentation fixes, help with the release
notes, non-code clean up is always welcome!
Thanks,
Chris
Here is the git log of work since PDL-2.015:
Update VERSION to 2.015_001 (a.k.a. PDL 2.016 rc1)
Add regular {op}over equivalent names
The operations in Ufunc along a dimension are mostly named
something ending in -over like 'sumover', except for the
cases of routines with names including average, minimum, and
maximum. This was a cute but surprising set of special cases.
Adding this following equivalent routines:
average -> avgover
daverage -> davgover
maximum -> maxover
maximum_ind -> maxover_ind
maximum_n_ind -> maxover_n_ind
minmaximum -> minmaxover
minimum -> minover
minimum_ind -> minover_ind
minimum_n_ind -> minover_n_ind
tweak to badvalue handline in transform.pd
update documentation for Transform, to reflect bad value tracking
add better bad-value detection on the input (output values are bad
if 1/3 or
more of the weighted input values corresponding to that pixel are bad)
add bad value support to the anti-aliased map method
Fix two bugs related to reading RICE-compressed FITS images.
1) The NAXIS* keywords were not replaced with their
ZNAXIS* equivalents, so the NAXIS fields in the header were incorrect.
2) The Z*, TFIELDS, TTYPE*, and TFORM* keywords were deleted from
the wronge header object, so the header of the returned piddle
still had those fields present.
Add a .gitattributes file so GitHub repo language stats are correct.
Do not call pp_done for a module that forgets to call it.
The comment in the code explains the reasoning.
Add pp_done to the end of bad.pd
Add versioning checks to gsl_sf.t.
Obviously the new functions don't work with old gsl, and shouldn't
be tested. Thanks Travis-CI!
Change from eval BLOCK to eval EXPR in t/gsl_sf.t
The module-loading logic needs this, for some reason: probably has
to do with a compile-time vs run-time check.
Load all GSLSF modules in t/gsl_sf.t; add 2 test for Legendre module.
At least by loading the modules we will get a test failure later if
some function we provide bindings for has been deprecated.
Only define deprecated gsl_sf_legendre functions if using GSL<2.0.
Otherwise when use'ing PDL::GSLSF::LEGENDRE with PERL_DL_NONLAZY=1
(as in 'make test') it would complain about undefined symbols (as
it should).
Add functions gsl_sf_legendre_array and gsl_sf_legendre_array_index.
They should work fully at this point. Have not added anything to
the test suite, though.
Deprecate gsl_sf_legendre_Plm_array and gsl_sf_legendre_sphPlm_array.
Just a note int the docs for now.
Add new calling convention for gsl_sf_ellint_D for gsl>=2.0.
In version 2.0 gsl removed the 'n' parameter for gsl_sf_ellint_D,
so we add a new pp_def if the user has gsl>=2.0. Note that this
is a build-time detection, so if the user upgrades gsl they will
need to rebuild PDL to get the new function usage.
allow shortcut PDL::GSL::RNG->new(..)->set_seed(..)
Require ccNcompt return type to be >= long.
Add 2 tests to check for this for byte image inputs.
Add SF/GitHub workflow integration to DEVELOPMENT
primitive.pd: spelling & punctuation fixes, delete trailing whitespace.
Have whichND return piddles of indx type.
This was already the case for which and which_both, so now whichND
returns indx piddles as well. Added 5 tests to t/primitive.t to
deal with (most of?) the edge cases.
Change handling of empty piddles by PDL::info and 'help vars'.
When using the $piddle->info command or 'help vars' or '? vars'
from the perldl shell, both null piddles and empty piddles were
printed as 'PDL->null'. But empties and nulls are not the same:
empties have a type, and a dim list (at least 1 dim is size 0).
This change allows empties to show up in the variable listing as
such, and I added some tests.
Spelling corrections, remove debugging print statment.
Remove extraneous unused variable in t/lm.t
Run dos2unix on Example/Fit/lmfit_example.pl <http://lmfit_example.pl>
Clarify documentation in PDL::Fit::LM. Add small test-suite case.
Mostly this clarifies the usage of the 'sig' argument to lmfit.
'sig' was short for 'sigma', I have changed the former to the latter.
Confusion sowed because in the example the variable was called
$wt ("weight"), when in reality the weights turn out to be
1/(sigma**2).
I added a small test suite in t/lm.t. This uses the data 'silver.dat'
from Gnuplot, which I believe to be compatible with PDL's license.
Comment out debug print statement
allow reshaping of PDLs with dataflow; the dataflow connection gets
severed.
------------------------------------------------------------------------------
What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic
patterns at an interface-level. Reveals which users, apps, and protocols are
consuming the most bandwidth. Provides multi-vendor support for NetFlow,
J-Flow, sFlow and other flows. Make informed decisions using capacity
planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e
_______________________________________________
pdl-devel mailing list
pdl-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/pdl-devel
