[R-pkg-team] Bug#893895: Bug#893895: dh-r: versioned dependencies are incorrect when epochs are present

Andreas Tille andreas at an3as.eu
Mon May 28 13:11:59 BST 2018


On Sun, May 27, 2018 at 06:56:54PM +0200, Graham Inggs wrote:
> > Could you please be more specific, what exactly is the problem with
> > epochs?
> 
> The R package's DESCRIPTION might have Depends: rsdmx (>= 0.6) which
> dh-r will translate to Depends: r-cran-rsdmx (>= 0.6) in
> debian/control, but that will be satisfied by r-cran-rsdmx
> r-cran-rsdmx/1:0.5-11+dfsg-1 in the archive.

Ahhh, got it now.
 
> > How can I reproduce the problem?
> 
> Either find an R package with a versioned dependency on a package with
> an epoch, or make one by editing an existing package's DESCRIPTION.
> 
> I had a quick look though the R packages [1] (I know that's not all),
> but there was only 1 out of 105 package with an epoch.  I wonder if
> the simplest would be for dh-r to hold a table of the packages that
> have an epoch > 0 as well as the epoch's value, and then simply add
> the epoch where required.  If a package gets an epoch, it only needs
> to be recorded in one place.
> 
> [1] https://qa.debian.org/developer.php?login=r-pkg-team@alioth-lists.debian.net

Here is the complete list:

udd=# select DISTINCT package, regexp_replace(version, ':.*', ':') as epoch  from packages where package like 'r-%' and version like '%:%' ;
      package      | epoch 
-------------------+-------
 r-bioc-altcdfenvs | 1:
 r-cran-deal       | 1:
 r-cran-dichromat  | 1:
 r-cran-epitools   | 1:
 r-cran-maptools   | 1:
 r-cran-rjags      | 1:
 r-cran-rsdmx      | 1:
 r-cran-snow       | 1:
 r-cran-sp         | 1:
 r-cran-spc        | 1:
 r-cran-vcd        | 1:
 r-cran-xtable     | 1:
(12 rows)


That should be a manageable amount for a manual list.  I wonder how
important it would be to do fix this before the transition.  Given that
my Perl knowledge is very rudimentary and I felt forced to do some
changes in dh-r for things I really wanted to have on a very beginner
level I'm not really inspired to do some potentially harmful changes in
the current situation where all r-* packages will be touched without
much testing in advance.

Kind regards

        Andreas.

-- 
http://fam-tille.de



More information about the R-pkg-team mailing list