Even better solution to issue #30.

R/RowFactor.R

@@ -57,9 +57,11 @@ RowFactor <- function(x, name = deparse(expr), levelnames=levels(x),
             else insert <- nopagebreak
             catname <- paste(insert, levelnames[i], sep="")
     	    test <- i  # Work around a bug in R 2.12.x!
-    	    test <- call("labelSubset",
+
+    	    # labelSubset is local; can't use call() here; see issue #30
+    	    test <- as.call(list(labelSubset,
     	                 subset = call("==", call("as.integer", call("as.factor", expr)), i),
-    	                 label = deparse(expr))
+    	                 label = deparse(expr)))
             term <- call("*", call("Heading", makeName(catname)), 
                               test)
             if (i == 1)

R/tabular.R

@@ -629,21 +629,10 @@ tabular.formula <- function(table, data=NULL, n, suppressLabels=0, ...) {
     if (missing(n) && inherits(data, "data.frame"))
     	n <- nrow(data)
 
-    # We need access to labelSubset() (and perhaps other functions in future)
-    # when evaluating a table expression (issue #30), but we don't want
-    # to mask the user's copy.
-
-    parent <- if (is.environment(data)) data else environment(table)
-    if (!exists("labelSubset", envir = parent)) {
-      withTableFns <- new.env(parent = parent)
-      withTableFns$labelSubset <- labelSubset
-    } else
-      withTableFns <- parent
-
     if (is.null(data))
-    	data <- withTableFns
+    	data <- environment(table)
     else if (is.list(data))
-    	data <- list2env(data, parent = withTableFns)
+    	data <- list2env(data, parent = environment(table))
     else if (!is.environment(data))
     	stop("'data' must be a dataframe, list or environment")