Fix 2 bugs

Bug 1 windows files are not resolved correctly TLDR.. fuck windows
Bug 2 log files could have basic blocks with no parent module and they
get module number 65535. I would ignore them for now.

cutterDRcovPlugin/drcov.py

@@ -1,7 +1,6 @@
 import re
-from pathlib import Path
 import struct
-
+from .extras import file_name
 
 MIN_DRCOV_FILE_SIZE = 20
 DRCOV_VERSION = 2
@@ -52,9 +51,9 @@ def parse_module_entry(f, version):
     #XXX now put commas and spaces in the file path and this gets fucked up
     entry = re.split(",\s+", entry)
     if version == 2:
-        return {"start": int(entry[1], 16), "name": Path(entry[-1]).name}
+        return {"start": int(entry[1], 16), "name": file_name(entry[-1])}
     else:
-        return {"start": int(entry[2], 16), "name": Path(entry[-1]).name}
+        return {"start": int(entry[2], 16), "name": file_name(entry[-1])}
 
 def read_module_list(f):
     modules = []    
@@ -79,6 +78,11 @@ def read_bb_list(f, module_count):
         # size of struct is 64 bit
         bb = f.read(8)
         offset, size, mod_num = struct_unpack(bb)
+        if mod_num > module_count:
+            # we have a case where dynamocov failed to capture which modules
+            # does this basic block belongs to
+            print("Warning: we have unknown module number:", mod_num)
+            continue
         bblist[mod_num][offset] = size
     return bblist
 

cutterDRcovPlugin/extras.py

@@ -1,5 +1,6 @@
 from PySide2.QtWidgets import QTableWidgetItem
 from PySide2.QtCore import Qt
+import ntpath
 
 class PercentWidgetItem(QTableWidgetItem):
     def __lt__(self, other):
@@ -25,3 +26,19 @@ def centered_text(x):
 def hexPad(num, pad):
     return "{0:#0{1}x}".format(num,pad + 2)
 
+
+# https://stackoverflow.com/questions/8384737/extract-file-name-from-path-no-matter-what-the-os-path-format
+# cuz windows sucks :( .. hard
+def file_name(path):
+   # There's one caveat: Linux filenames may contain backslashes. So on linux,
+    # r'a/b\c' always refers to the file b\c in the a folder, while on Windows,
+    # it always refers to the c file in the b subfolder of the a folder. So when
+    # both forward and backward slashes are used in a path, you need to know the
+    # associated platform to be able to interpret it correctly. In practice it's
+    # usually safe to assume it's a windows path since backslashes are seldom
+    # used in Linux filenames, but keep this in mind when you code so you don't
+    # create accidental security holes.
+
+    head, tail = ntpath.split(path)
+    return tail or ntpath.basename(head)
+