replace most Pos objects/ptrs with indexes into a position table

Pos objects are somewhat wasteful as they duplicate the origin file name and
input type for each object. on files that produce more than one Pos when parsed
this a sizeable waste of memory (one pointer per Pos). the same goes for
ptr<Pos> on 64 bit machines: parsing enough source to require 8 bytes to locate
a position would need at least 8GB of input and 64GB of expression memory. it's
not likely that we'll hit that any time soon, so we can use a uint32_t index to
locate positions instead.

1 files changed, 3 insertions, 3 deletions

diff --git a/src/libexpr/attr-path.cc b/src/libexpr/attr-path.cc
index c6e3a9c92..1c12dfbe2 100644
--- a/src/libexpr/attr-path.cc
+++ b/src/libexpr/attr-path.cc
@@ -41,13 +41,13 @@ std::vector<Symbol> parseAttrPath(EvalState & state, std::string_view s)
 }
 
 
-std::pair<Value *, Pos> findAlongAttrPath(EvalState & state, const std::string & attrPath,
+std::pair<Value *, PosIdx> findAlongAttrPath(EvalState & state, const std::string & attrPath,
     Bindings & autoArgs, Value & vIn)
 {
     Strings tokens = parseAttrPath(attrPath);
 
     Value * v = &vIn;
-    Pos pos = noPos;
+    PosIdx pos = noPos;
 
     for (auto & attr : tokens) {
 
@@ -83,7 +83,7 @@ std::pair<Value *, Pos> findAlongAttrPath(EvalState & state, const std::string &
                 throw AttrPathNotFound(suggestions, "attribute '%1%' in selection path '%2%' not found", attr, attrPath);
             }
             v = &*a->value;
-            pos = *a->pos;
+            pos = a->pos;
         }
 
         else {