pyassem: Capture line number of first (module) instruction.

And use this line number as first line number of module's code object. This
is quite magic, but achieves (about) the same behavior as CPython has.

Author: pfalcon

compiler/pyassem.py

@@ -48,6 +48,9 @@ def __init__(self):
         # line numbers in the code block. (If it's not set, it will be
         # deduced).
         self.firstline = 0
+        # Line number of first instruction output. Used to deduce .firstline
+        # if it's not set explicitly.
+        self.first_inst_lineno = 0
 
         # Was replaced with ordered_blocks
         #self.blocks = misc.Set()
@@ -117,6 +120,8 @@ def emit(self, *inst):
         if len(inst) == 2 and isinstance(inst[1], Block):
             self.current.addOutEdge(inst[1])
         self.current.emit(inst)
+        if inst[0] == "SET_LINENO" and not self.first_inst_lineno:
+            self.first_inst_lineno = inst[1]
 
     def getBlocksInOrder(self):
         """Return the blocks in reverse postorder
@@ -545,7 +550,7 @@ def _convert_COMPARE_OP(self, arg):
     def makeByteCode(self):
         assert self.stage == CONV
         self.lnotab = lnotab = LineAddrTable()
-        lnotab.setFirstLine(self.firstline or 1)
+        lnotab.setFirstLine(self.firstline or self.first_inst_lineno or 1)
         for t in self.insts:
             opname = t[0]
             if len(t) == 1:
@@ -577,6 +582,11 @@ def newCodeObject(self):
             nlocals = len(self.varnames)
 
         firstline = self.firstline
+        # For module, .firstline is initially not set, and should be first
+        # line with actual bytecode instruction (skipping docstring, optimized
+        # out instructions, etc.)
+        if not firstline:
+            firstline = self.first_inst_lineno
         # If no real instruction, fallback to 1
         if not firstline:
             firstline = 1

compiler/pycodegen.py

@@ -382,6 +382,12 @@ def visitModule(self, node):
         lnf = walk(node.body, self.NameFinder(), verbose=0)
         self.locals.push(lnf.getLocals())
         self.visit(self.skip_docstring(node.body))
+
+        # See if the was a live statement, to later set its line number as
+        # module first line. If not, fall back to first line of 1.
+        if not self.graph.first_inst_lineno:
+            self.graph.first_inst_lineno = 1
+
         self.emit('LOAD_CONST', None)
         self.emit('RETURN_VALUE')