Skip to content

completeSubclasses() methods bug (PR#4051)

2 messages · Colin A. Smith, John Chambers

#
Full_Name: Colin A. Smith
Version: 1.8.0
OS: Mac OS X 10.2.6
Submission from: (NULL) (128.102.184.81)


annaffy 1.0.1 (in BioC CVS) fails to load. annaffy 1.0 (on the BioC web site)
has the same problem.

It looks like the load is failing because of a bug in completeSubclasses() in
r-devel. It calls setIs() and doesn't specify an environment via the "where"
argument. setIs() "where" defaults to topEnv() and fails because "aafGO" doesn't
exist in that environment and "list" is sealed.

completeSubclasses should be specifying my package for "where" instead of using
the default.
Error in setIs(class2, obji@superClass, extensionObject = obji, doComplete =
FALSE) : 
        Cannot create a setIs relation when neither of the classes ("aafGO" and
"list") is local and modifiable in this package
Error in setClass("aafGO", "aafList", prototype = list(), where = where) : 
        Error in contained classes ("aafList") for class "aafGO"; class
definition removed from "annaffy"
Error in library(annaffy) : .First.lib failed
#
Yes, thanks.  I have a corrected version checked out.  It has some other
changes--as soon as I make a cleaned-up version with this fix only, I'll
commit it.

John
colin@colinsmith.org wrote: