fixed nim-lang/nimsuggest#48 type aware sug (#16814)

* suggesting identifiers accounts context over scope (distance)
* key takeaway: context fit is prioritized over a heuristics like scope

nimsuggest/tests/tsug_template.nim

@@ -6,7 +6,7 @@ tmp#[!]#
 discard """
 $nimsuggest --tester $file
 >sug $1
-sug;;skMacro;;tsug_template.tmpb;;macro (){.noSideEffect, gcsafe, locks: 0.};;$file;;2;;6;;"";;0;;Prefix
-sug;;skConverter;;tsug_template.tmpc;;converter ();;$file;;3;;10;;"";;0;;Prefix
-sug;;skTemplate;;tsug_template.tmpa;;template ();;$file;;1;;9;;"";;0;;Prefix
+sug;;skMacro;;tsug_template.tmpb;;macro (){.noSideEffect, gcsafe, locks: 0.};;$file;;2;;6;;"";;100;;Prefix
+sug;;skConverter;;tsug_template.tmpc;;converter ();;$file;;3;;10;;"";;100;;Prefix
+sug;;skTemplate;;tsug_template.tmpa;;template ();;$file;;1;;9;;"";;100;;Prefix
 """

nimsuggest/tests/tsug_typedecl.nim

@@ -0,0 +1,26 @@
+# suggestions for type declarations
+
+from system import string, int, bool
+
+type
+  super = int
+  someType = bool
+
+let str = "hello"
+
+proc main() =
+  let a: s#[!]#
+
+# This output show seq, even though that's not imported. This is due to the
+# entire symbol table, regardless of import visibility is currently being
+# scanned. This is hardly ideal, but changing it with the current level of test
+# coverage is unwise as it might break more than it fixes.
+
+discard """
+$nimsuggest --tester $file
+>sug $1
+sug;;skType;;tsug_typedecl.someType;;someType;;*nimsuggest/tests/tsug_typedecl.nim;;7;;2;;"";;100;;Prefix
+sug;;skType;;tsug_typedecl.super;;super;;*nimsuggest/tests/tsug_typedecl.nim;;6;;2;;"";;100;;Prefix
+sug;;skType;;system.string;;string;;*lib/system.nim;;*;;*;;*;;100;;Prefix
+sug;;skType;;system.seq;;seq;;*lib/system.nim;;*;;*;;*;;100;;Prefix
+"""