Restore reshape2 redirection (with warning)

NEWS.md

@@ -244,7 +244,7 @@
 
 7. Added a note to `?frank` clarifying that ranking is being done according to C sorting (i.e., like `forder`), [#2328](https://github.com/Rdatatable/data.table/issues/2328). Thanks to @cguill95 for the request.
 
-8. Historically, `dcast` and `melt` were built as enhancements to `reshape2`'s own `dcast`/`melt`. We removed dependency on `reshape2` in v1.9.6 but maintained some backward compatibility. As that package has been deprecated since December 2017, we have now formally completed the split from `reshape2` by removing some last vestiges, [#3549](https://github.com/Rdatatable/data.table/issues/3549). We thank the `reshape2` authors for their original inspiration for these functions.
+8. Historically, `dcast` and `melt` were built as enhancements to `reshape2`'s own `dcast`/`melt`. We removed dependency on `reshape2` in v1.9.6 but maintained some backward compatibility. As that package has been deprecated since December 2017, we will begin to formally complete the split from `reshape2` by removing some last vestiges. In particular we now warn when redirecting to `reshape2` methods and will later error before ultimately completing the split; see [#3549](https://github.com/Rdatatable/data.table/issues/3549) and [#3633](https://github.com/Rdatatable/data.table/issues/3633). We thank the `reshape2` authors for their original inspiration for these functions, and @ProfFancyPants for testing and reporting regressions in dev which have been fixed before release.
 
 9. `DT[col]` where `col` is a column containing row numbers of itself to select, now suggests the correct syntax (`DT[(col)]` or `DT[DT$col]`), [#697](https://github.com/Rdatatable/data.table/issues/697). This expands the message introduced in [#1884](https://github.com/Rdatatable/data.table/issues/1884) for the case where `col` is type `logical` and `DT[col==TRUE]` is suggested.
 

R/fcast.R

@@ -8,9 +8,21 @@ guess = function(x) {
   return(var)
 }
 
-dcast = function(data, formula, fun.aggregate = NULL, ..., margins = NULL,
-                  subset = NULL, fill = NULL, value.var = guess(data)) {
-  UseMethod("dcast", data)
+dcast <- function(
+  data, formula, fun.aggregate = NULL, ..., margins = NULL,
+  subset = NULL, fill = NULL, value.var = guess(data)
+) {
+  if (is.data.table(data)) UseMethod("dcast", data)
+  # nocov start
+  else {
+    data_name = deparse(substitute(data))
+    ns = tryCatch(getNamespace("reshape2"), error=function(e)
+      stop("The dcast generic in data.table has been passed a ",class(data)[1L],", but data.table::dcast currently only has a method for data.tables. Please confirm your input is a data.table, with setDT(", data_name, ") or as.data.table(", data_name, "). If you intend to use a reshape2::dcast, try installing that package first, but do note that reshape2 is deprecated and you should be migrating your code away from using it."))
+    warning("The dcast generic in data.table has been passed a ", class(data)[1L], " and will attempt to redirect to the reshape2::dcast; please note that reshape2 is deprecated, and this redirection is now deprecated as well. Please do this redirection yourself like reshape2::dcast(", data_name, "). In the next version, this warning will become an error.")
+    ns$dcast(data, formula, fun.aggregate = fun.aggregate, ..., margins = margins,
+             subset = subset, fill = fill, value.var = value.var)
+  }
+  # nocov end
 }
 
 check_formula = function(formula, varnames, valnames) {

R/fmelt.R

@@ -1,9 +1,22 @@
 # reshape2 dependency was originally abandoned because (1) we wanted to be in control
 #   of the R version dependency and (2) reshape2::dcast is not generic.
-#   Anyway, reshape2 package is deprecated since December 2017.
+#   reshape2 package is deprecated since December 2017, so we'll deprecate our
+#   redirection as well
 
-melt = function(data, ..., na.rm = FALSE, value.name = "value") {
-  UseMethod("melt", data)
+melt <- function(data, ..., na.rm = FALSE, value.name = "value") {
+  if (is.data.table(data)) {
+    UseMethod("melt", data)
+    # if data is not data.table and reshape2 is installed, this won't dispatch to reshape2's method;
+    # CRAN package edarf and others fail without the else branch
+  # nocov start
+  } else {
+    data_name = deparse(substitute(data))
+    ns = tryCatch(getNamespace("reshape2"), error=function(e)
+      stop("The melt generic in data.table has been passed a ",class(data)[1L],", but data.table::melt currently only has a method for data.tables. Please confirm your input is a data.table, with setDT(", data_name, ") or as.data.table(", data_name, "). If you intend to use a method from reshape2, try installing that package first, but do note that reshape2 is deprecated and you should be migrating your code away from using it."))
+    warning("The melt generic in data.table has been passed a ", class(data)[1L], " and will attempt to redirect to the relevant reshape2 method; please note that reshape2 is deprecated, and this redirection is now deprecated as well. To continue using melt methods from reshape2 while both libraries are attached, e.g. melt.list, you can prepend the namespace like reshape2::melt(", data_name, "). In the next version, this warning will become an error.")
+    ns$melt(data, ..., na.rm=na.rm, value.name=value.name)
+  }
+  # nocov end
 }
 
 patterns = function(..., cols=character(0L)) {