[Bug 241081] Review Request: R-widgetTools-1.12.0-2 - Tools to support the construction of tcltk widgets

[bugzilla at redhat.com]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug report.

Summary: Review Request: R-widgetTools-1.12.0-2 - Tools to support the construction of tcltk widgets


https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=241081





------- Additional Comments From tibbs at math.uh.edu  2007-07-02 21:17 EST -------
OK, now we're getting somewhere.  Here are the remaining issues that I see:

This packager is still missing a dependency on R.
You should own the directory /usr/share/R/library/widgetTools. The easiest way
to do this without recursively pulling in everything below is to have
  %dir %{_datadir}/R/library/%{packname}
in your %files section.
I think the latex directory is documentation as well.

These are all minor.  However, here's a tough one: The reference to %{_libdir}
in %post means that this isn't actually a noarch package.  I built it in x86_64
rawhide, and being noarch it should install fine on i386 rawhide, but it doesn't:

/var/tmp/rpm-tmp.97210: line 1: /usr/lib64/R/doc/html/search/index.txt: No such
file or directory
error: %post(R-widgetTools-1.12.0-7.fc8.noarch) scriptlet failed, exit status 1

This is bad; either you need to figure out the directory at runtime or this
can't be a noarch package.  And ignoring that, don't your scriptlets as is
ignore the arch-specific R modules that might be installed by looking only in
_datadir?

I'm beginning to wonder whether it's at all worth it to have noarch R packages.

I know this is tougher than it needs to be because we have no R packaging
guidelines and I'm just sort of guessing as to what R actually needs.

-- 
Configure bugmail: https://bugzilla.redhat.com/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug, or are watching the QA contact.