Added cedta.pkgEvalsUserCode and only any eval() is affected not othe…

…r code in the package itself (e.g. knitr::kable). Closes #809.

DESCRIPTION

@@ -5,7 +5,7 @@ Author: M Dowle, T Short, S Lianoglou, A Srinivasan with contributions from R Sa
 Maintainer: Matt Dowle <mdowle@mdowle.plus.com>
 Depends: R (>= 2.14.0)
 Imports: methods, chron
-Suggests: ggplot2 (>= 0.9.0), plyr, reshape, reshape2, testthat (>= 0.4), hexbin, fastmatch, nlme, xts, bit64, gdata, GenomicRanges, caret
+Suggests: ggplot2 (>= 0.9.0), plyr, reshape, reshape2, testthat (>= 0.4), hexbin, fastmatch, nlme, xts, bit64, gdata, GenomicRanges, caret, knitr
 Description: Fast aggregation of large data (e.g. 100GB in RAM), fast ordered joins, fast add/modify/delete of columns by group using no copies at all, list columns and a fast file reader (fread). Offers a natural and flexible syntax, for faster development.
 License: GPL (>= 2)
 URL: https://github.com/Rdatatable/data.table/wiki

R/cedta.R

@@ -1,26 +1,38 @@
 
-cedta.override = c("gWidgetsWWW","statET","FastRWeb","slidify","rmarkdown","knitr")
-# These packages tend to be ones that run user code in their own environment and thus do not
-# themselves Depend or Import data.table.
+cedta.override = NULL  # If no need arises, will deprecate.
+
+cedta.pkgEvalsUserCode = c("gWidgetsWWW","statET","FastRWeb","slidify","rmarkdown","knitr")
+# These packages run user code in their own environment and thus do not
+# themselves Depend or Import data.table. knitr's eval is passed envir=globalenv() so doesn't
+# need to be listed here currently, but we include it in case it decides to change that.
+# The others create their own environment to eval in.
+# The packages might use data.frame syntax elsewhere in their package code so we just want
+# the eval (only) of user code to be data.table-aware.
+
 # If a new package needs to be added to this vector, a user may add to it using :
-#   assignInNamespace("cedta.override", c(data.table:::cedta.override,"<nsname>"), "data.table")
+#   assignInNamespace("cedta.pkgEvalsUserCode", c(data.table:::cedta.pkgEvalsUserCode,"<nsname>"), "data.table")
 # But please let us know so we can add the package to this vector in the package upstream, so other
 # users don't have to tread the same path. Then you can remove your assignInNamepace() call.
-# Or, packages like those above can set a variable in their namespace
+
+# Packages may also set a variable in their namespace :
 #   .datatable.aware = TRUE
-# which achieves the same thing.  Either way.
+# which makes them data.table-aware optionally and possibly variably.
 # http://stackoverflow.com/a/13131555/403310
 
 cedta = function(n=2L) {
     # Calling Environment Data Table Aware
     te = topenv(parent.frame(n))
-    if (!isNamespace(te)) return(TRUE)  # e.g. DT queries in .GlobalEnv
+    if (!isNamespace(te)) {
+        # e.g. DT queries at the prompt (.GlobalEnv) and knitr's eval(,envir=globalenv()) but not DF[...] inside knitr::kable v1.6
+        return(TRUE)
+    }
     nsname = getNamespaceName(te)
     ans = nsname == "data.table" ||
         "data.table" %chin% names(getNamespaceImports(te)) ||
         "data.table" %chin% tryCatch(get(".Depends",paste("package",nsname,sep=":"),inherits=FALSE),error=function(e)NULL) ||
         (nsname == "utils" && exists("debugger.look",parent.frame(n+1L))) || 
         (nsname == "base"  && all(c("FUN", "X") %in% ls(parent.frame(n)))  ) || # lapply
+        (nsname %chin% cedta.pkgEvalsUserCode && any(sapply(sys.calls(), "[[", 1L)=="eval")) ||
         nsname %chin% cedta.override ||
         identical(TRUE, tryCatch(get(".datatable.aware",asNamespace(nsname),inherits=FALSE),error=function(e)NULL))
     if (!ans && getOption("datatable.verbose"))

R/data.table.R

@@ -80,12 +80,13 @@ print.data.table = function(x,
 {
     if (.global$print != "" &&
         length(SYS<-last(sys.calls()))>=2 &&
-        typeof(SYS[[2]]) == "list" &&   # auto-printing; e.g. not explicit print.  When explicit print this is not list; e.g. symbol or language
+        typeof(SYS[[2]]) %chin% c("list","promise") &&  # auto-printing; e.g. not explicit print. "promise" to future proof
         address(x) == .global$print ) {
         #  := in [.data.table sets print=address(x) to suppress next autoprint at the console. See FAQ 2.22 and README item in v1.9.5
         # Other options investigated (could revisit): Cstack_info(), .Last.value gets set first before autoprint, history(), sys.status(),
         #   topenv(), inspecting next statement in caller, using clock() at C level to timeout suppression after some number of cycles
-        # The issue is distinguishing "> DT" (after a previous := in a function) from "> DT[,foo:=1]". To print.data.table() there is no difference.
+        # The issue is distinguishing "> DT" (after a previous := in a function) from "> DT[,foo:=1]". To print.data.table() there
+        # is no difference.
         .global$print = ""
         return(invisible())
 

README.md

@@ -44,10 +44,12 @@
   12. Compatibility with big endian machines (e.g., SPARC and PowerPC) is restored. Most Windows, Linux and Mac systems are little endian; type `.Platform$endian` to confirm. Thanks to Gerhard Nachtmann for reporting and the [QEMU project](http://qemu.org/) for their PowerPC emulator.
 
   13. `DT[, LHS := RHS]` with RHS is of the form `eval(parse(text = foo[1]))` referring to columns in `DT` is now handled properly. Closes [#880](https://github.com/Rdatatable/data.table/issues/880). Thanks to tyner.
+
+  14. `knitr::kable()` works again without needing to upgrade from knitr v1.6 to v1.7. Packages which evaluate user code and don't wish to import data.table need to be added to `data.table:::cedta.pkgEvalsUserCode` and now only the `eval` part is made data.table-aware (the rest of such package's code is left data.table-unaware). `data.table:::cedta.override` is now empty and will be deprecated if no need for it arises.
 
-  14. `subset` handles extracting duplicate columns in consistency with data.table's rule - if a column name is duplicated, then accessing that column using column number should return that column, whereas accessing by column name (due to ambiguity) will always extract the first column. Closes [#891](https://github.com/Rdatatable/data.table/issues/891). Thanks to @jjzz.
+  15. `subset` handles extracting duplicate columns in consistency with data.table's rule - if a column name is duplicated, then accessing that column using column number should return that column, whereas accessing by column name (due to ambiguity) will always extract the first column. Closes [#891](https://github.com/Rdatatable/data.table/issues/891). Thanks to @jjzz.
 
-  15. `rbindlist` handles combining levels of data.tables with both ordered and unordered factor columns properly. Closes [#899](https://github.com/Rdatatable/data.table/issues/899). Thanks to @ChristK.
+  16. `rbindlist` handles combining levels of data.tables with both ordered and unordered factor columns properly. Closes [#899](https://github.com/Rdatatable/data.table/issues/899). Thanks to @ChristK.
 
 #### NOTES
 

inst/tests/tests.Rraw

@@ -14,6 +14,7 @@ if (!exists("test.data.table",.GlobalEnv,inherits=FALSE)) {
     require(gdata)
     require(GenomicRanges)
     require(caret)
+    require(knitr)
     # reshape2 ahead of reshape ...
     try(detach(package:reshape2),silent=TRUE)
     try(detach(package:reshape),silent=TRUE)
@@ -5434,6 +5435,13 @@ for (i in seq_along(DT)) {
         test(test_no, ans1, ans2)
         0L
     })
+
+# That data.table-unaware code in packages like knitr still work
+if ("package:knitr" %in% search()) {
+    DT = data.table(x=1, y=2)
+    test(1395, kable(DT), output="x.*y.*1.*2")  # kable in knitr v1.6 calls DF[...] syntax
+} else {
+    cat("Test 1395 not run. If required call library(knitr) first.\n")
 }
 
 ##########################
@@ -5465,14 +5473,14 @@ for (i in seq_along(DT)) {
 ##########################
 options(warn=0)
 options(oldbwb) # set at top of this file
-plat = paste(", endian=",.Platform$endian,", ","sizeof(long double)==",.Machine$sizeof.longdouble,sep="")
+plat = paste("endian=",.Platform$endian,", sizeof(long double)==",.Machine$sizeof.longdouble,sep="")
 if (nfail > 0) {
     if (nfail>1) {s1="s";s2="s: "} else {s1="";s2=" "}
     cat("\r")
-    stop(nfail," error",s1," out of ",ntest, " (lastID=",lastnum, plat, ") in inst/tests/tests.Rraw on ",date(),". Search tests.Rraw for test number",s2,paste(whichfail,collapse=", "),".")
+    stop(nfail," error",s1," out of ",ntest, " (lastID=",lastnum,", ",plat, ") in inst/tests/tests.Rraw on ",date(),". Search tests.Rraw for test number",s2,paste(whichfail,collapse=", "),".")
     # important to stop() here, so that 'R CMD check' fails
 }
-cat("\rAll ",ntest," tests (lastID=",lastnum, plat, ") in inst/tests/tests.Rraw completed ok in ",timetaken(started.at)," on ",date(),"\n",sep="")
+cat("\rAll ",ntest," tests (lastID=",lastnum,") in inst/tests/tests.Rraw completed ok in ",timetaken(started.at)," on ",date()," (",plat,")\n",sep="")
 # Reporting lastnum rather than ntest makes it easier to check user has the latest version, assuming
 # each release or patch has extra tests.
 # date() is included so we can tell when CRAN checks were run (in particular if they have been rerun since