cleanup and cosmetics - for a healthy future!! er... what?

diff --git a/elftools/dwarf/compileunit.py b/elftools/dwarf/compileunit.py
index 8852fac0a7bfe06e953b1dc26c04b9f3c9d53a2b..ec04f7de941c67e5ace7c8b9d87770b8221b2df4 100644 (file)
--- a/elftools/dwarf/compileunit.py
+++ b/elftools/dwarf/compileunit.py
@@ -10,6 +10,23 @@ from .die import DIE
 
 
 class CompileUnit(object):
+    """ A DWARF compilation unit (CU). 
+    
+            A normal compilation unit typically represents the text and data
+            contributed to an executable by a single relocatable object file.
+            It may be derived from several source files, 
+            including pre-processed "include files"
+            
+        Serves as a container and context to DIEs that describe objects and code
+        belonging to a compilation unit.
+        
+        CU header entries can be accessed as dict keys from this object, i.e.
+           cu = CompileUnit(...)
+           cu['version']  # version field of the CU header
+        
+        To get the top-level DIE describing the compilation unit, call the 
+        get_top_DIE method.
+    """
     def __init__(self, header, dwarfinfo, structs, cu_die_offset):
         """ header:
                 CU header for this compile unit

diff --git a/elftools/dwarf/die.py b/elftools/dwarf/die.py
index fba5ed29e2b9465ee4222be32ff910b93a0d9106..26e9dbaf986f05fded44b9aae8a8406958d6c8bb 100644 (file)
--- a/elftools/dwarf/die.py
+++ b/elftools/dwarf/die.py
@@ -25,7 +25,11 @@ from ..common.utils import struct_parse, preserve_stream_pos
 #   Raw value as parsed from the section - used for debugging and presentation
 #   (e.g. for a DW_FORM_strp it's the raw string offset into the table)
 #
-AttributeValue = namedtuple('AttributeValue', 'form value raw_value')
+# offset:
+#   Offset of this attribute's value in the stream
+#
+AttributeValue = namedtuple(
+    'AttributeValue', 'form value raw_value offset')
 
 
 class DIE(object):
@@ -73,6 +77,8 @@ class DIE(object):
         """
         return self.tag is None
     
+    #------ PRIVATE ------#
+    
     def _parse_DIE(self):
         """ Parses the DIE info from the section, based on the abbreviation
             table of the CU
@@ -101,10 +107,14 @@ class DIE(object):
         # values from the stream.
         #
         for name, form in abbrev_decl.iter_attr_specs():
-            print '**', self.stream.tell()
+            attr_offset = self.stream.tell()
             raw_value = struct_parse(structs.Dwarf_dw_form[form], self.stream)
             value = self._translate_attr_value(form, raw_value)            
-            self.attributes[name] = AttributeValue(form, value, raw_value)
+            self.attributes[name] = AttributeValue(
+                form=form,
+                value=value,
+                raw_value=raw_value,
+                offset=attr_offset)
         
         self.size = self.stream.tell() - self.offset
 

diff --git a/elftools/dwarf/dwarfinfo.py b/elftools/dwarf/dwarfinfo.py
index 43016439ea5c796a262f122c593feac71376fdb9..7a4bbc03b324b613ce3df5a02d22fb513a2bb306 100644 (file)
--- a/elftools/dwarf/dwarfinfo.py
+++ b/elftools/dwarf/dwarfinfo.py
@@ -51,14 +51,19 @@ class DWARFInfo(object):
         self.little_endian = little_endian
         
         # This is the DWARFStructs the context uses, so it doesn't depend on 
-        # DWARF format and address_size (these are determined per CU) - so we
-        # set them to default values.
+        # DWARF format and address_size (these are determined per CU) - set them
+        # to default values.
         self.structs = DWARFStructs(
             little_endian=self.little_endian,
             dwarf_format=32,
             address_size=4)
         
-        # Populate the list with CUs found in debug_info
+        # Populate the list with CUs found in debug_info. For each CU only its
+        # header is parsed immediately (the abbrev table isn't loaded before
+        # it's being referenced by one of the CU's DIEs). 
+        # Since there usually aren't many CUs in a single object, this
+        # shouldn't present a performance problem.
+        #
         self._CU = self._parse_CUs()
         
         # Cache for abbrev tables: a dict keyed by offset
@@ -121,6 +126,8 @@ class DWARFInfo(object):
             self.stream,
             stream_pos=self.debug_str_loc.offset + offset)
     
+    #------ PRIVATE ------#
+    
     def _parse_CUs(self):
         """ Parse CU entries from debug_info.
         """
@@ -157,7 +164,7 @@ class DWARFInfo(object):
                 cu_structs = DWARFStructs(
                     little_endian=self.little_endian,
                     dwarf_format=dwarf_format,
-                    address_size=8)
+                     address_size=8)
             
             cu_die_offset = self.stream.tell()
             dwarf_assert(

diff --git a/z.py b/z.py
index ea7c762a628d6beeddd1aa27d1da84322c7bd499..5e5e775ce209b57c0efcc79c2a2277f835ce9ab7 100644 (file)
--- a/z.py
+++ b/z.py
@@ -22,12 +22,13 @@ print efile.has_dwarf_info()
 dwarfinfo = efile.get_dwarf_info()
 tt = dwarfinfo.structs.Dwarf_dw_form['DW_FORM_block1'].parse('\x03\x12\x34\x46')
 
-cu = dwarfinfo.get_CU(1)
+cu = dwarfinfo.get_CU(0)
 print 'CU header', cu.header
 topdie = cu.get_top_DIE()
 
 print topdie.size, topdie.tag
-print topdie.attributes
+for attrname, val in topdie.attributes.iteritems():
+    print attrname, val
 
 #~ print dwarfinfo.structs.Dwarf_abbrev_entry.parse('\x13\x01\x01\x03\x50\x04\x00\x00')