TDS and/or MAP files in WinDbg?

WINDBG
1

Hi,

is it possible to use TDS (i.e. Borland's Turbo Debugger Symbol) files in WinDbg? If not, is there a way to convert the symbols to PDBs?

Although it is not specifically kernel-related, it's related to WinDbg.

Alternately, would it be possible to write a debugger extension that is able to find (TDS) symbols if you put the respective information into the binary (similarly to the info about the PDB)? And if so, could such an extension be used in order to handle TDS files instead of PDB files and otherwise offer the symbols in a way compatible to any debugger commands using/displaying symbols here and there?

Thanks,

// Oliver

DDKWizard and DDKBUILD: http:</http:>

2

I think that a MAP file would work. You might want to look at sdbgext by Skywing.

Good luck,

mm

Oliver Schneider wrote:

Hi,

is it possible to use TDS (i.e. Borland’s Turbo Debugger Symbol) files in WinDbg? If not, is there a way to convert the symbols to PDBs?

Although it is not specifically kernel-related, it’s related to WinDbg.

Alternately, would it be possible to write a debugger extension that is able to find (TDS) symbols if you put the respective information into the binary (similarly to the info about the PDB)?

And if so, could such an extension be used in order to handle TDS files instead of PDB files and otherwise offer the symbols in a
way compatible to any debugger commands using/displaying symbols

here and there?

Thanks,

// Oliver

3

you can try using luscian wishchik’s (did i spell his name right )
map2dbg utility to convert borlands map to .dbg files which i believe
older windbgs used to decipher

im not sure about latest versions i ve used some old map files some
years earlier this way

or other name i recall is pisti’s dbgmaker utility

i also think oleg starodumov also has some convertors

and of course skywings sdbgexts has a mapfile loader extension

regards

raj_r

On 4/22/08, Oliver Schneider wrote:
> Hi,
>
> is it possible to use TDS (i.e. Borland’s Turbo Debugger Symbol) files in WinDbg? If not, is there a way to convert the symbols to PDBs?
>
> Although it is not specifically kernel-related, it’s related to WinDbg.
>
> Alternately, would it be possible to write a debugger extension that is able to find (TDS) symbols if you put the respective information into the binary (similarly to the info about the PDB)? And if so, could such an extension be used in order to handle TDS files instead of PDB files and otherwise offer the symbols in a way compatible to any debugger commands using/displaying symbols here and there?
>
>
> Thanks,
>
> // Oliver
> –
> —
> DDKWizard and DDKBUILD: http:
>
>
>
> —
> You are currently subscribed to windbg as: xxxxx@gmail.com
> To unsubscribe send a blank email to xxxxx@lists.osr.com
></http:>

4

raj_r wrote:

you can try using luscian wishchik’s (did i spell his name right )
map2dbg utility to convert borlands map to .dbg files which i believe
older windbgs used to decipher

Raj - thanks for that, I’ve been wanting to use WinDbg to debug some
Delphi stuff.

Incidentally, it’s Lucian Wischik - ironic how the same names come up
again and again - I remember him from the Dept of Computer Science at
Cambridge Uni - he used to specialise in type calculi. Don’t know if he
still does.

MH.

5

On 4/22/08, Martin Harvey wrote:

> Raj - thanks for that, I’ve been wanting to use WinDbg to debug some Delphi
> stuff.
>
> Incidentally, it’s Lucian Wischik -

yep thanks googled again before you replied and found out that his
name is spelled lucian
Lucian Jules Wischik

and it seems he is at redmond now working on behavioural types and
contract conformance checking for concurrent programs.

regards
raj

6

Thanks, Martin. Also thanks to the other people who replied.

I think this will do it.

Any idea for the part of storing the name of the debug symbol file in the header? Does that even work with anything but PDB files. MS seems to be very secretive about PDB. However, Sven B. Schneider has described some details about that stuff - with source - in his book about “Undocumented Windows 2000 Secrets”, although the name already suggests that this could be a bit outdated :wink:

// Oliver

PS: Sorry for the laten reply.

-------- Original-Nachricht --------

Datum: Tue, 22 Apr 2008 09:15:32 -0400
Von: Martin O'Brien
> An: “Kernel Debugging Interest List”
> Betreff: Re:[windbg] TDS and/or MAP files in WinDbg?

> I think that a MAP file would work. You might want to look at sdbgext by
> Skywing.
>
> Good luck,
>
> mm
>
> Oliver Schneider wrote:
> > Hi,
> >
> > is it possible to use TDS (i.e. Borland’s Turbo Debugger Symbol) files
> in WinDbg? If not, is there a way to convert the symbols to PDBs?
> >
> > Although it is not specifically kernel-related, it’s related to WinDbg.
> >
> > Alternately, would it be possible to write a debugger extension that is
> able to find (TDS) symbols if you put the respective information into the
> binary (similarly to the info about the PDB)?
>
> And if so, could such an extension be used in order to handle TDS files
> instead of PDB files and otherwise offer the symbols in a
> way compatible to any debugger commands using/displaying symbols
>
> here and there?
> >
> >
> > Thanks,
> >
> > // Oliver
>
> —
> You are currently subscribed to windbg as: xxxxx@gmxpro.net
> To unsubscribe send a blank email to xxxxx@lists.osr.com

–
—
DDKWizard and DDKBUILD: http:</http:>

7

Oliver:

I’m also sorry for the late reply.

That book is the only source of information I know on the subject. The part on PDB’s is the only part of the book that I have read,
but I found it very helpful and still for the most part accurate. It’s been a little over three years since I’ve done anything with
this, but between this book and more than a little time in spent in a good hex editor, it’s very doable to reverse the whole thing
sufficiently to get the symbol and type information out of it.

As far as the header part, I’m confused. The basic answer is that the debug directory RVA will point you to a CV_NB10 header which
that book describes, the last member of which is the pdb name (\0), but I don’t understand what you’re going to do with this
information. If you mean take a TDS, point the exe header to it, write and extension and then have WinDbg handle the rest - i. e. -
replace the symbol handler, in effect - I seriously doubt that this will work, but I really don’t know. The only thing that I’ve
seen along these lines is under ‘Synthetic Symbols’ in the WinDbg documentation, I think. That’s the best I have; sorry.

I guess what I had in my mind was writing a program to convert a TDS file to a MAP file (I think that there is something out there
that already does this), and then use sdbgext to handle the MAP, which I think would work, not that I’ve ever tried any of this.

Am I missing something?

Good luck,

mm

Oliver Schneider wrote:

Thanks, Martin. Also thanks to the other people who replied.

I think this will do it.

Any idea for the part of storing the name of the debug symbol file in the header? Does that even work with anything but PDB files. MS seems to be very secretive about PDB. However, Sven B. Schneider has described some details about that stuff - with source - in his book about “Undocumented Windows 2000 Secrets”, although the name already suggests that this could be a bit outdated :wink:

// Oliver

PS: Sorry for the laten reply.

-------- Original-Nachricht --------
> Datum: Tue, 22 Apr 2008 09:15:32 -0400
> Von: Martin O'Brien
>> An: “Kernel Debugging Interest List”
>> Betreff: Re:[windbg] TDS and/or MAP files in WinDbg?
>
>> I think that a MAP file would work. You might want to look at sdbgext by
>> Skywing.
>>
>> Good luck,
>>
>> mm
>>
>> Oliver Schneider wrote:
>>> Hi,
>>>
>>> is it possible to use TDS (i.e. Borland’s Turbo Debugger Symbol) files
>> in WinDbg? If not, is there a way to convert the symbols to PDBs?
>>> Although it is not specifically kernel-related, it’s related to WinDbg.
>>>
>>> Alternately, would it be possible to write a debugger extension that is
>> able to find (TDS) symbols if you put the respective information into the
>> binary (similarly to the info about the PDB)?
>>
>> And if so, could such an extension be used in order to handle TDS files
>> instead of PDB files and otherwise offer the symbols in a
>> way compatible to any debugger commands using/displaying symbols
>>
>> here and there?
>>>
>>> Thanks,
>>>
>>> // Oliver
>> —
>> You are currently subscribed to windbg as: xxxxx@gmxpro.net
>> To unsubscribe send a blank email to xxxxx@lists.osr.com
>

8

Martin,

thanks for the reply.

If you mean take a TDS, point the exe header to it, write
and extension and then have WinDbg handle the rest - i. e. -
replace the symbol handler, in effect - I seriously doubt that this will
work, but I really don't know.
That was indeed what I had in mind.

But since this seems to be a dead end, I am going to look into the other option of creating a PDB instead.

The only thing that I've
seen along these lines is under 'Synthetic Symbols' in the WinDbg
documentation, I think. That's the best I have; sorry.
Thanks.

I guess what I had in my mind was writing a program to convert a TDS file
to a MAP file (I think that there is something out there
that already does this), and then use sdbgext to handle the MAP, which I
think would work, not that I've ever tried any of this.
Actually BCB and friends are perfectly capable to create MAP files, so this won't be the biggest problem in this case.

For the immediate needs I am going to use sdbgext, but I'll look into it to see whether there is a way to create PDBs and also modify the header in order to use the PDB. Preferably as a build-step from within BDB/BDS etc with the so-called ToolsAPI.

Thanks again for the detailed reply, it definitely got me kickstarted in the right direction.

// Oliver

DDKWizard and DDKBUILD: http:</http:>

9

Glad to help.

Good luck,

mm

Oliver Schneider wrote:

Martin,

thanks for the reply.

> If you mean take a TDS, point the exe header to it, write
> and extension and then have WinDbg handle the rest - i. e. -
> replace the symbol handler, in effect - I seriously doubt that this will
> work, but I really don’t know.
That was indeed what I had in mind.

But since this seems to be a dead end, I am going to look into the other option of creating a PDB instead.

> The only thing that I’ve
> seen along these lines is under ‘Synthetic Symbols’ in the WinDbg
> documentation, I think. That’s the best I have; sorry.
Thanks.

> I guess what I had in my mind was writing a program to convert a TDS file
> to a MAP file (I think that there is something out there
> that already does this), and then use sdbgext to handle the MAP, which I
> think would work, not that I’ve ever tried any of this.
Actually BCB and friends are perfectly capable to create MAP files, so this won’t be the biggest problem in this case.

For the immediate needs I am going to use sdbgext, but I’ll look into it to see whether there is a way to create PDBs and also modify the header in order to use the PDB. Preferably as a build-step from within BDB/BDS etc with the so-called ToolsAPI.

Thanks again for the detailed reply, it definitely got me kickstarted in the right direction.

// Oliver

10

On 4/23/08, Oliver Schneider wrote:
> Thanks, Martin. Also thanks to the other people who replied.
>
> I think this will do it.
>
> Any idea for the part of storing the name of the debug symbol file in the header? Does that even work with anything but PDB files. MS seems to be very secretive about PDB. However, Sven B. Schneider has described some details about that stuff - with source - in his book about “Undocumented Windows 2000 Secrets”, although the name already suggests that this could be a bit outdated :wink:
>

i dont know if you have take a look at map2dbg by lucian
and pistis (istvan hajnals ) debugmaker

the former converts maps to .dbg files without problems and latest windbg does
decipher it without problems

and debugmaker from pisti can add arbitrary DEBUG_DATA_DIRECTORY
HEADER information and add .dbg info to files that are stripped

just to confitm

i wrote a single liner .c

D:\ms-dbg>type helldbg.c
#include <stdio.h>

int main (void)

{
printf(“hello debug\n”);
return 1;
}

D:\ms-dbg>
compiled it using bcc32 commandline tools

bcc32 -v -ls helldbg.c

ran it

D:\ms-dbg>helldbg.exe
hello debug

map2dbgged it

map2dbg helldbg.exe

it converted 573 symbols and wrote a .dbg file

ran windbg on it

did a x helldbg!* on this exe
it can find the symbols

0040a10f helldbg!_TLS_index =
0040a113 helldbg!_TLS_index4 =
0040a117 helldbg!_hInstance =
0040a11c helldbg! CPPdebugHook_segment =
0040a11c helldbg! CPPdebugHook =
0040a120 helldbg!_PackageInfoTable =
0040a128 helldbg!d2_1 =

tried disassembling helldbg!main

0:000> uf helldbg!main
helldbg!c2_0:
00401150 55 push ebp
00401151 8bec mov ebp,esp
00401153 6828a14000 push offset helldbg!d2_1 (0040a128)
00401158 e80f270000 call helldbg!printf (0040386c)
0040115d 59 pop ecx
0040115e b801000000 mov eax,1
00401163 5d pop ebp
00401164 c3 ret

0:000> da helldbg!d2_1
0040a128 “hello debug.”

found msvc equivalent CRTMain to be equally good for bcc32 exes in windbg

0:000> ln 401000
(00401000) helldbg!c1_0 | (00401059) helldbg!_GetExceptDLLinfo
Exact matches:
helldbg!_acrtused =
helldbg!c1_0 =
0:000> uf helldbg!c1_0
helldbg!c1_0:
00401000 eb10 jmp helldbg!c1_0+0x12 (00401012)

helldbg!c1_0+0x12:
00401012 a10fa14000 mov eax,dword ptr [helldbg!_TLS_index (0040a10f)]
00401017 c1e002 shl eax,2
0040101a a313a14000 mov dword ptr [helldbg!_TLS_index4 (0040a113)],eax
0040101f 52 push edx
00401020 6a00 push 0
00401022 e8d9880000 call helldbg!_CurrExcContext+0x9e (00409900)
00401027 8bd0 mov edx,eax
00401029 e88e100000 call helldbg!c240_0 (004020bc)
0040102e 5a pop edx
0040102f e824040000 call helldbg!__CRTL_MEM_UseBorMM (00401458)
00401034 e887100000 call helldbg!__CRTL_VCLLIB_Linkage (004020c0)
00401039 6a00 push 0
0040103b e8981c0000 call helldbg!_ExceptInit (00402cd8)

all in all quiet good

also if you are compiling there is .bpl that you can registee using
install components in bcb which can automagically create .dbg files
directly
is what the readme says

also boost ppl have updated this map2dbg for bcb 6 as well it seems
from a glance in google

regards

raj_r</stdio.h>

11

I’m surprised that WinDbg still supports dbg files, which I think were last seen in W2K (maybe NT4; I don’t know)). I suppose
that’s an upside to software from MS that is free, because it seems like if this were anything sold by MS, this feature would have
been history, probably before the support for the new way was known to work in all cases. In any case, I don’t really know what TDS
even is, other than a Borland debug format, so I’m confused how this helps him, as I recall that he wants WinDbg to natively
support TDS, which seems unlikely to me? I’m just curious.

Thanks,

mm

raj_r wrote:

On 4/23/08, Oliver Schneider wrote:
>> Thanks, Martin. Also thanks to the other people who replied.
>>
>> I think this will do it.
>>
>> Any idea for the part of storing the name of the debug symbol file in the header? Does that even work with anything but PDB files. MS seems to be very secretive about PDB. However, Sven B. Schneider has described some details about that stuff - with source - in his book about “Undocumented Windows 2000 Secrets”, although the name already suggests that this could be a bit outdated :wink:
>>
>
> i dont know if you have take a look at map2dbg by lucian
> and pistis (istvan hajnals ) debugmaker
>
> the former converts maps to .dbg files without problems and latest windbg does
> decipher it without problems
>
> and debugmaker from pisti can add arbitrary DEBUG_DATA_DIRECTORY
> HEADER information and add .dbg info to files that are stripped
>
> just to confitm
>
> i wrote a single liner .c
>
> D:\ms-dbg>type helldbg.c
> #include <stdio.h>
>
> int main (void)
>
> {
> printf(“hello debug\n”);
> return 1;
> }
>
>
> D:\ms-dbg>
> compiled it using bcc32 commandline tools
>
> bcc32 -v -ls helldbg.c
>
> ran it
>
> D:\ms-dbg>helldbg.exe
> hello debug
>
> map2dbgged it
>
>
> map2dbg helldbg.exe
>
> it converted 573 symbols and wrote a .dbg file
>
> ran windbg on it
>
> did a x helldbg!* on this exe
> it can find the symbols
>
> 0040a10f helldbg!_TLS_index =
> 0040a113 helldbg!_TLS_index4 =
> 0040a117 helldbg!_hInstance =
> 0040a11c helldbg! CPPdebugHook_segment =
> 0040a11c helldbg! CPPdebugHook =
> 0040a120 helldbg!_PackageInfoTable =
> 0040a128 helldbg!d2_1 =
>
> tried disassembling helldbg!main
>
> 0:000> uf helldbg!main
> helldbg!c2_0:
> 00401150 55 push ebp
> 00401151 8bec mov ebp,esp
> 00401153 6828a14000 push offset helldbg!d2_1 (0040a128)
> 00401158 e80f270000 call helldbg!printf (0040386c)
> 0040115d 59 pop ecx
> 0040115e b801000000 mov eax,1
> 00401163 5d pop ebp
> 00401164 c3 ret
>
>
> 0:000> da helldbg!d2_1
> 0040a128 “hello debug.”
>
>
> found msvc equivalent CRTMain to be equally good for bcc32 exes in windbg
>
>
> 0:000> ln 401000
> (00401000) helldbg!c1_0 | (00401059) helldbg!_GetExceptDLLinfo
> Exact matches:
> helldbg!_acrtused =
> helldbg!c1_0 =
> 0:000> uf helldbg!c1_0
> helldbg!c1_0:
> 00401000 eb10 jmp helldbg!c1_0+0x12 (00401012)
>
> helldbg!c1_0+0x12:
> 00401012 a10fa14000 mov eax,dword ptr [helldbg!_TLS_index (0040a10f)]
> 00401017 c1e002 shl eax,2
> 0040101a a313a14000 mov dword ptr [helldbg!_TLS_index4 (0040a113)],eax
> 0040101f 52 push edx
> 00401020 6a00 push 0
> 00401022 e8d9880000 call helldbg!_CurrExcContext+0x9e (00409900)
> 00401027 8bd0 mov edx,eax
> 00401029 e88e100000 call helldbg!c240_0 (004020bc)
> 0040102e 5a pop edx
> 0040102f e824040000 call helldbg!__CRTL_MEM_UseBorMM (00401458)
> 00401034 e887100000 call helldbg!__CRTL_VCLLIB_Linkage (004020c0)
> 00401039 6a00 push 0
> 0040103b e8981c0000 call helldbg!_ExceptInit (00402cd8)
>
>
> all in all quiet good
>
> also if you are compiling there is .bpl that you can registee using
> install components in bcb which can automagically create .dbg files
> directly
> is what the readme says
>
>
>
> also boost ppl have updated this map2dbg for bcb 6 as well it seems
> from a glance in google
>
> regards
>
> raj_r
></stdio.h>

12

On 4/28/08, Martin O’Brien wrote:
> I’m surprised that WinDbg still supports dbg files, which I think were last
> seen in W2K (maybe NT4; I don’t know)). I suppose that’s an upside to
> software from MS that is free, because it seems like if this were anything
> sold by MS, this feature would have been history, probably before the
> support for the new way was known to work in all cases.

.dbg files were last probably seen in w2k but i recall seeing .dbg
files in xp (maybe plain xp and not xp sp1 or xp sp2 too for some
kernelcomponents (may be hal or bootvid.dll not sure and im really not
sure if i saw them in xp (my symbol cache atm doesnt hold .dbg files
they are rsds format pdbs mostly )

In any case, I
> don’t really know what TDS even is, other than a Borland debug format, so
> I’m confused how this helps him, as I recall that he wants WinDbg to
> natively support TDS, which seems unlikely to me? I’m just curious.
>

windbg natively will not support tds i think ( ive never gone down
that path would be more appropriate answer i think) so yes it seems
its mostly unlikely

but tds file is 99% plain ascii text and borland had released a debug
hook library(bordbg.dll ) by which one can browse through .tds files
and .tds files have source level debugging information

but borlands linker map files have all symbols except source level information
(you can find tds2map which claims to embed source level info as well
into map files though i havent used it in real sense apart from
experimenting)

and by converting tha map to .dbg files one can use windbg
effortlessly on borland compiled files
and thats what i said in my prior posts

i have never tried mucking around with tds files directly in windbg

regards

raj_r

13

Thanks, raj_r.

mm

raj_r wrote:

On 4/28/08, Martin O’Brien wrote:
>> I’m surprised that WinDbg still supports dbg files, which I think were last
>> seen in W2K (maybe NT4; I don’t know)). I suppose that’s an upside to
>> software from MS that is free, because it seems like if this were anything
>> sold by MS, this feature would have been history, probably before the
>> support for the new way was known to work in all cases.
>
> …dbg files were last probably seen in w2k but i recall seeing .dbg
> files in xp (maybe plain xp and not xp sp1 or xp sp2 too for some
> kernelcomponents (may be hal or bootvid.dll not sure and im really not
> sure if i saw them in xp (my symbol cache atm doesnt hold .dbg files
> they are rsds format pdbs mostly )
>
> In any case, I
>> don’t really know what TDS even is, other than a Borland debug format, so
>> I’m confused how this helps him, as I recall that he wants WinDbg to
>> natively support TDS, which seems unlikely to me? I’m just curious.
>>
>
>
> windbg natively will not support tds i think ( ive never gone down
> that path would be more appropriate answer i think) so yes it seems
> its mostly unlikely
>
> but tds file is 99% plain ascii text and borland had released a debug
> hook library(bordbg.dll ) by which one can browse through .tds files
> and .tds files have source level debugging information
>
> but borlands linker map files have all symbols except source level information
> (you can find tds2map which claims to embed source level info as well
> into map files though i havent used it in real sense apart from
> experimenting)
>
> and by converting tha map to .dbg files one can use windbg
> effortlessly on borland compiled files
> and thats what i said in my prior posts
>
> i have never tried mucking around with tds files directly in windbg
>
> regards
>
> raj_r
>

14

Thanks indeed, raj_r. It has shed some further light on the issue.

The fact that WinDbg doesn’t have any means of supporting TDS “natively” won’t hurt, if I am able to figure out the TDS and the PDB format. As was mentioned before, Schreibers book gives a good clue, although it may be partially outdated.

// Oliver

-------- Original-Nachricht --------

Datum: Tue, 29 Apr 2008 23:10:56 -0400
Von: Martin O'Brien
> An: “Kernel Debugging Interest List”
> Betreff: Re:[windbg] TDS and/or MAP files in WinDbg?

> Thanks, raj_r.
>
> mm
>
>
> raj_r wrote:
> > On 4/28/08, Martin O’Brien wrote:
> >> I’m surprised that WinDbg still supports dbg files, which I think were
> last
> >> seen in W2K (maybe NT4; I don’t know)). I suppose that’s an upside to
> >> software from MS that is free, because it seems like if this were
> anything
> >> sold by MS, this feature would have been history, probably before the
> >> support for the new way was known to work in all cases.
> >
> > …dbg files were last probably seen in w2k but i recall seeing .dbg
> > files in xp (maybe plain xp and not xp sp1 or xp sp2 too for some
> > kernelcomponents (may be hal or bootvid.dll not sure and im really not
> > sure if i saw them in xp (my symbol cache atm doesnt hold .dbg files
> > they are rsds format pdbs mostly )
> >
> > In any case, I
> >> don’t really know what TDS even is, other than a Borland debug format,
> so
> >> I’m confused how this helps him, as I recall that he wants WinDbg to
> >> natively support TDS, which seems unlikely to me? I’m just curious.
> >>
> >
> >
> > windbg natively will not support tds i think ( ive never gone down
> > that path would be more appropriate answer i think) so yes it seems
> > its mostly unlikely
> >
> > but tds file is 99% plain ascii text and borland had released a debug
> > hook library(bordbg.dll ) by which one can browse through .tds files
> > and .tds files have source level debugging information
> >
> > but borlands linker map files have all symbols except source level
> information
> > (you can find tds2map which claims to embed source level info as well
> > into map files though i havent used it in real sense apart from
> > experimenting)
> >
> > and by converting tha map to .dbg files one can use windbg
> > effortlessly on borland compiled files
> > and thats what i said in my prior posts
> >
> > i have never tried mucking around with tds files directly in windbg
> >
> > regards
> >
> > raj_r
> >
>
> —
> You are currently subscribed to windbg as: xxxxx@gmxpro.net
> To unsubscribe send a blank email to xxxxx@lists.osr.com

–
—
DDKWizard and DDKBUILD: http:</http:>

15

On 4/30/08, Oliver Schneider wrote:
> Thanks indeed, raj_r. It has shed some further light on the issue.
>
> The fact that WinDbg doesn’t have any means of supporting TDS “natively” won’t hurt, if I am able to figure out the TDS and the PDB format. As was mentioned before, Schreibers book gives a good clue, although it may be partially outdated.

oh btw talking about sven’s book i think he released an updated pdbx
pdbexploder
for xp pds files (the new rsds format ) search for his e book relaease
site rawol something and scour around downloads you will land it

also in matt pietreks personal site (wheaty.net) you can get hold of
a precompiled and updated msdn mag article’s called dbghelpdemo which
can decipher and decompile ?? the latest pdbs to its stream the code
is wrapped around the ms-dia (debug information accessor sdk’s com
classes

also jeremy gordon (goasm authour) has an article on the format of
lates rsds pdbs

regards

raj_r

16

Excellent information, raj_r. Thanks again.

mm

raj_r wrote:

On 4/30/08, Oliver Schneider wrote:
>> Thanks indeed, raj_r. It has shed some further light on the issue.
>>
>> The fact that WinDbg doesn’t have any means of supporting TDS “natively” won’t hurt, if I am able to figure out the TDS and the PDB format. As was mentioned before, Schreibers book gives a good clue, although it may be partially outdated.
>
>
> oh btw talking about sven’s book i think he released an updated pdbx
> pdbexploder
> for xp pds files (the new rsds format ) search for his e book relaease
> site rawol something and scour around downloads you will land it
>
> also in matt pietreks personal site (wheaty.net) you can get hold of
> a precompiled and updated msdn mag article’s called dbghelpdemo which
> can decipher and decompile ?? the latest pdbs to its stream the code
> is wrapped around the ms-dia (debug information accessor sdk’s com
> classes
>
> also jeremy gordon (goasm authour) has an article on the format of
> lates rsds pdbs
>
> regards
>
> raj_r
>

17

On 4/30/08, Oliver Schneider wrote:
Thanks indeed, raj_r. It has shed some further light on the issue.

On 4/30/08, Martin O’Brien wrote:
> Excellent information, raj_r. Thanks again.

glad to be of help

regards

raj_r