Bug 490723

Package does not build for me - missing BuildRequires?

** R
** inst
** preparing package for lazy loading
Error in loadNamespace(i[[1]], c(lib.loc, .libPaths())) :
  there is no package called 'Matrix'
Calls: <Anonymous> ... namespaceImportFrom -> asNamespace -> loadNamespace
Execution halted
ERROR: lazy loading failed for package 'IRanges'
** Removing '/builddir/build/BUILDROOT/R-IRanges-1.0.14-1.fc10.x86_64/usr/lib64\
/R/library/IRanges'
RPM build errors:
error: Bad exit status from /var/tmp/rpm-tmp.9f4Ede (%install)

Indeed R-Matrix is needed for build, I will update the spec in that way.

SPEC:
http://pingou.fedorapeople.org/RPMs/R-IRanges.spec
SRPM:
http://pingou.fedorapeople.org/RPMs/R-IRanges-1.0.14-2.fc10.src.rpm

Created attachment 336194 [details]
Review result

I attach my review. In summary it looks quite good.

> ! The package license corresponds to the license as mentioned in the
>  DESCRIPTION file and in the website, however several source files
>  contain the statement:
> 
> * This file is copyright 2002 Jim Kent, but license is hereby
> * granted for all use - public, private or commercial. */
> 
>  which is not the same as "Artistic licence 2.0". Needs some
>  consultation with upstream.

I will ask upstream for clarification

> ! Specfile is written in legible English and uses macro consitently,
>  however:
> 
>  The Source and URL fields uses the macro %{BioC} with is not defined
>  anywhere. The urls point to the right location if this macro is
>  assumed to be an empty string, so techncally they are correct, but
>  it is a source of confusion. Suggestion - remove the macros.

The macro is defined (line 2 of the file), however the %define should be changed to %global according to the newly approved guidelines.

>  The comment that says "#i368 arch" should probaly read something
>  like "#architecture dependent package", because that I think is what
>  you really mean.

That comment is generated by R2spec, it can be changed/ignored/removed
(I will change R2spec to this)

> ! Package compiles, but there are warnings that should be fixed:
> 
> IntervalTree.c:48: warning: missing braces around initializer
> memalloc.c:293: warning: format '%d' expects type 'int', but argument 2 has type 'size_t'
> memalloc.c:293: warning: format '%d' expects type 'int', but argument 3 has type 'size_t'
> 
>  The "... may be used uninitialized" warnings might be harmless (or
>  not, ...)

> I have no idea if these warnings are harmful or not:
> 
> Creating a generic for "cbind" in package  "IRanges"
>    (the supplied definition differs from and overrides the implicit generic in package "base": Signatures differ:  (...), (deparse.level))
> Creating a generic for "rbind" in package  "IRanges"
>    (the supplied definition differs from and overrides the implicit generic in package "base": Signatures differ:  (...), (deparse.level))

I will point these warning to upstream.

> ! %check is present and check pass, but
> 
>  Comparing 'runalltests.Rout' to 'runalltests.Rout.save' ...11,12c11
> < 	 rbind,
> < 	 sapply 
> ---
> > 	 rbind 
> 
>  It test still says "OK", but it looks "strange".

This is just the output of the test run

> ! No package owns the package's main directory /usr/lib64/R/library/IRanges

Oups, I will change this

>  Ideally both main and devel should own the main directory in order
>  to avoid orphaned directories after package removal, but at least
>  main must own it.

I do not agree. Since -devel has the main package has a requirement, only the main need to own the directory.

> ! Since the -devel package does not include a .pc file it should not
>  require pkgconfig (or is there some other reason for this?)

Will be fixed

* Sun Mar 22 2009 pingou <pingou -AT- pingoured.fr> 1.0.14-3
- The main package owns the directory
- Remove pkgconfig as R for the devel package
- Define becomes global

SPEC:
http://pingou.fedorapeople.org/RPMs/R-IRanges.spec
SRPM:
http://pingou.fedorapeople.org/RPMs/R-IRanges-1.0.14-3.fc10.src.rpm

I got a reply to the mail I sent:
-----------------------------------
> We also see some warnings while compiling it:
>  IntervalTree.c:48: warning: missing braces around initializer
>  memalloc.c:293: warning: format '%d' expects type 'int', but argument 2
> has type 'size_t'
>  memalloc.c:293: warning: format '%d' expects type 'int', but argument 3
> has type 'size_t'
>

These are probably harmless but easy to rectify.


>
>  Creating a generic for "cbind" in package  "IRanges"
>    (the supplied definition differs from and overrides the implicit generic
> in package "base": Signatures differ:  (...), (deparse.level))
>  Creating a generic for "rbind" in package  "IRanges"
>    (the supplied definition differs from and overrides the implicit generic
> in package "base": Signatures differ:  (...), (deparse.level))
>

These are intentional and may be ignored.
-------------------------------------

Nothing regarding the license yet though

Answer concerning the license:

> We will add a new field to the DESCRIPTION file, like ExtraLicenses or
> something that will indicate the C files that are under freely licensed for
> all uses by Jim Kent, in a way that is compatible with the Artistic license.

Updated:
Spec URL: http://pingou.fedorapeople.org/RPMs/R-IRanges.spec
SRPM URL: http://pingou.fedorapeople.org/RPMs/R-IRanges-1.1.55-1.fc10.src.rpm

Package approved.

Thanks for the review :)

New Package CVS Request
=======================
Package Name: R-IRanges
Short Description: Low-level containers for storing set of integer ranges
Owners: pingou
Branches: F-9 F-10
InitialCC:

cvs done.

R-IRanges-1.1.55-1.fc10 has been submitted as an update for Fedora 10.
http://admin.fedoraproject.org/updates/R-IRanges-1.1.55-1.fc10

R-IRanges-1.1.55-1.fc9 has been submitted as an update for Fedora 9.
http://admin.fedoraproject.org/updates/R-IRanges-1.1.55-1.fc9

R-IRanges-1.1.55-1.fc10 has been pushed to the Fedora 10 stable repository.  If problems still persist, please make note of it in this bug report.

R-IRanges-1.1.55-1.fc9 has been pushed to the Fedora 9 stable repository.  If problems still persist, please make note of it in this bug report.