Exclude '..1' from up-a-level search (#7128)

* Exclude '..1' from up-a-level search

* slightly simplify (?) regex

* Revert "slightly simplify (?) regex"

This reverts commit 74a54ef.

* new test belying the "simplification" in prev. commit

* another one

* i guess these should all work (?)

* local() avoid scope spillover

* remove the failing test

Author: MichaelChirico

NEWS.md

@@ -111,6 +111,8 @@
 
 18. `fwrite` now allows `dec` to be the same as `sep` for edge cases where only one will be written, e.g. 0-row or 1-column tables. [#7227](https://github.com/Rdatatable/data.table/issues/7227). Thanks @MichaelChirico for the report and @venom1204 for the fix.
 
+19. Ellipsis elements like `..1` are correctly excluded when searching for variables in "up-a-level" syntax inside `[`, [#5460](https://github.com/Rdatatable/data.table/issues/5460). Thanks @ggrothendieck for the report and @MichaelChirico for the fix.
+
 ### NOTES
 
 1. The following in-progress deprecations have proceeded:

R/data.table.R

@@ -264,7 +264,8 @@ replace_dot_alias = function(e) {
   if (!missing(j)) {
     jsub = replace_dot_alias(jsub)
     root = root_name(jsub)
-    av = all.vars(jsub)
+    # exclude '..1' etc. for #5460
+    av = grepv("^[.][.](?:[.]|[0-9]+)$", all.vars(jsub), invert=TRUE)
     all..names = FALSE
     if ((.is_withFALSE_range(jsub, x, root, av)) ||
         (root %chin% c("-","!") && jsub[[2L]] %iscall% '(' && jsub[[2L]][[2L]] %iscall% ':') || ## x[, !(V8:V10)]
@@ -1297,8 +1298,8 @@ replace_dot_alias = function(e) {
   SDenv = new.env(parent=parent.frame())
 
   syms = all.vars(jsub)
-  syms = syms[ startsWith(syms, "..") ]
-  syms = syms[ substr(syms, 3L, 3L) != "." ]  # exclude ellipsis
+  syms = syms[startsWith(syms, "..")]
+  syms = grepv("^[.][.](?:[.]|[0-9]+)$", syms, invert=TRUE) # exclude ellipsis and '..n' ellipsis elements
   for (sym in syms) {
     if (sym %chin% names_x) {
       # if "..x" exists as column name, use column, for backwards compatibility; e.g. package socialmixr in rev dep checks #2779

inst/tests/tests.Rraw

@@ -21568,16 +21568,16 @@ test(2333, as.expression(data.table(a = 1))[["a"]], 1)
 
 # regression test for hexdigits subscript overrun (uint8_t wraps over 255, unsigned overflow is well defined in c)
 local({
-	f = tempfile()
-	on.exit(unlink(f))
-	# the line is likely invalid in current encoding, so disable any translation, #7209
-	# test.data.table() sets options(encoding="UTF-8"), so go the long way around.
-	ff = file(f, encoding = "")
-	tryCatch(
-		writeLines(c('a', rep('0x1.ffffp0', 10000L), `Encoding<-`('0x1.ff\x9fp0', 'bytes'), rep('0x1.ffffp0', 20000L)), ff),
-		finally = close(ff)
-	)
-	test(2334, names(fread(f)), "a")
+  f = tempfile()
+  on.exit(unlink(f))
+  # the line is likely invalid in current encoding, so disable any translation, #7209
+  # test.data.table() sets options(encoding="UTF-8"), so go the long way around.
+  ff = file(f, encoding = "")
+  tryCatch(
+    writeLines(c('a', rep('0x1.ffffp0', 10000L), `Encoding<-`('0x1.ff\x9fp0', 'bytes'), rep('0x1.ffffp0', 20000L)), ff),
+    finally = close(ff)
+  )
+  test(2334, names(fread(f)), "a")
 })
 
 # Tests for new isoyear() helper (complement to isoweek) #7154
@@ -21624,3 +21624,34 @@ local({
   test(2338.9, {fwrite(dd, f, forceDecimal=FALSE); fread(f)}, di)
 })
 
+# '..1' is ...elt(1), not a variable named '1', when doing "up-a-level" search (#5460)
+DT = data.table(a=1.0)
+sqrt_dot_sym = function(...) sqrt(..1)
+test(2339.01, lapply(DT, function(...) sqrt(..1)), list(a=1.0))
+test(2339.02, lapply(DT, sqrt_dot_sym), list(a=1.0))
+test(2339.03, DT[, lapply(.SD, function(...) sqrt(..1))], data.table(a=1.0))
+test(2339.04, DT[, lapply(.SD, sqrt_dot_sym)], data.table(a=1.0))
+sqrt_elt_sym = function(...) sqrt(...elt(1))
+# TODO(R>=3.5.0): run this unconditionally
+if (!inherits(tryCatch(sqrt_elt_sym(1), error=identity), "error")) {
+  test(2339.05, lapply(DT, sqrt_elt_sym), list(a=1.0))
+  test(2339.06, lapply(DT, function(...) sqrt(...elt(1L))), list(a=1.0))
+  test(2339.07, DT[, lapply(.SD, sqrt_elt_sym)], data.table(a=1.0))
+  test(2339.08, DT[, lapply(.SD, function(...) sqrt(...elt(1L)))], data.table(a=1.0))
+}
+# edge case of weird, though technically valid, names:
+#  _not_ '..N' style and distinct from '...'
+#  `123`='a'; DT[, ..123] doesn't work, but that's fine & easy to work around
+DT = data.table(a = 4L)
+local({
+  `.123` = 'a'
+  test(2339.09, DT[, ...123], DT)
+})
+local({
+  ..123 = 'a'
+  test(2339.10, DT[, ....123], DT)
+})
+local({
+  ...123 = 'a'
+  test(2339.11, DT[, .....123], DT)
+})