-1

I'm doing dump debugging, using PYKD, and therefore I'm using the PYKD dbgCommand() for getting information on objects.

The problem is: the dbgCommand() results need to be parsed in order to be used, as you can see in following example:

source code : result = dbgCommand(("dt -c CStringArray m_nSize " + pointer_format) % (ptr)).split(' : ')
example     : dt -c CStringArray m_nSize 0x03966ce8
example output : 
  <application>!CStringArray
  +0x008 m_nSize 0n16

I'm only interested in the size itself (0n16), and I can parse the result in order to get this, but as I'm having lots of objects (some 100,000) this becomes very time-consuming, therfore I'd like to limit the result as much as possible.

Is there a way (using other display options, using other commands than dt, using native visualisers if needed) to get following situation:

dt <options> CStringArray m_nSize 0x03966ce8
0n16 // only that, nothing else

In the meanwhile, I've already arrived one step further, using dd command, as you can see:

0:000> dd 0x03966ce8+0x008 L1 // for a CStringArray, m_nSize is at memory address +0x008
                              // L1 means: limit the amount of answers to one byte
03966cf0  00000010            // the result only contains one line.

Now I only need to find a way not to see the memory address anymore.

Dominique
  • 16,450
  • 15
  • 56
  • 112

2 Answers2

2

Why don't you want to use typedVar class from pykd?

Try:

print( typedVar('CStringArray', address).m_nSize )
ussrhero
  • 606
  • 4
  • 5
0
0:000> dt -c foo m_nsize
Local var @ 0x2dfdb8 Type CStringArray
+0x008 m_nSize 0n5
0:000> .printf "%x\n" , @@c++(foo.m_nSize)
5
0:000>
blabb
  • 8,674
  • 1
  • 18
  • 27