Red Hat Bugzilla – Bug 183947
Review Request: R-waveslim - R package for wavelets
Last modified: 2007-11-30 17:11:26 EST
Spec Name or Url: http://www.fc.up.pt/pessoas/jamatos/fedora-extras/R-waveslim.spec
SRPM Name or Url: http://www.fc.up.pt/pessoas/jamatos/fedora-extras/R-waveslim-1.5-1.src.rpm
Basic wavelet routines for time series (1D), image (2D)
and array (3D) analysis. The code provided here is based on
wavelet methodology developed in Percival and Walden (2000);
Gencay, Selcuk and Whitcher (2001); the dual-tree complex wavelet
transform (CWT) from Kingsbury (1999, 2001) as implemented by
Selesnick; and Hilbert wavelet pairs (Selesnick 2001, 2002).
Not a review; just a few comments from a newbie. These are based on the other
comments that I've received and a quick scan...
%build missing an explicit make statement. Something like:
And where is it the Makefile? ;-)
Seriously this rule does not apply to programming languages that have their
own way to install packages. R is a free implementation of the S programming
If you look to most python packages you will notice that that there is not any
make there as well, they use distutils. The same should apply to perl packages
and I would expect to ruby as well.
Thank you for your input but it does not apply in this case. :-)
I like the package; contratulations on working up a specfile generator for R
packages. Everything looks pretty good. rpmlint says:
W: R-waveslim invalid-license GPL version 2 or newer
W: R-waveslim no-documentation
We have no standards for R package naming, but R-packagename is reasonable.
The specfile is clean and understandable.
The source matches upstream.
Package dependencies are sane.
rpmlint doesn't like the License: tag. No package in extras yet specifies the
acceptable GPL version; I would suggest just using GPL. When GPLv3 is out I
suppose we'll have to be more specific.
rpmlint doesn't find any documentation. Of course this package includes a pile
of documentation, but it's all incorporated into R's internal documentation
system. The base R package doesn't list any of the internal documentation as
%doc so I don't think it would be reasonable to do so here. This shouldn't be a
blocker, but if someone takes issue with it then the base R package needs some
work as well.
The package fails to build in mock:
* Installing *source* package 'waveslim' ...
gfortran -fPIC -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions
-fstack-protector --param=ssp-buffer-size=4 -m32
-march=i386 -mtune=pentium4 -fasynchronous-unwind-tables -c bell-p-w.f -o
make: gfortran: Command not found
make: *** [bell-p-w.o] Error 127
ERROR: compilation failed for package 'waveslim'
error: Bad exit status from /var/tmp/rpm-tmp.71138 (%install)
I believe you need a RuildRequires: gcc-gfortran.
According to the discussion had in packagers mailing list I have:
Rename License to GPL
Added DESCRIPTION in %doc
Added gcc-gfortran in BR
I have modified cran2rpmspec to follow these guidelines (the last two at
least) as well.
Looks good. The package now builds in mock and rpmlint is now silent.
The only question I have is the following from %check:
* checking if this is a source package ... WARNING
Subdirectory 'waveslim/src' contains object files.
WARNING: There was 1 warning, see
No idea what that means, but perhaps it's something for a future cleanup.
(In reply to comment #5)
> Looks good. The package now builds in mock and rpmlint is now silent.
> The only question I have is the following from %check:
> * checking if this is a source package ... WARNING
> Subdirectory 'waveslim/src' contains object files.
I guess that it is because of the file waveslim/src/so_locations that is a
:st = .text 0x5ffe0000, 0x00010000:\
:st = .data 0x5fff0000, 0x00010000:
> WARNING: There was 1 warning, see
> for details
> No idea what that means, but perhaps it's something for a future cleanup.
Probably, or to report it upstream. :-)
Build on target fedora-development-extras succeeded.