Introduction and winedbg bt problem

[Kai Brüning]

Good morning,

my name is Kai Brüning. Our company RagTime GmbH develops and markets the desktop publishing package RagTime, which had its origin in the Macintosh world 17 years ago. RagTime is currently available for Windows and Mac OS 9/X. Now we are evaluating whether we can make RagTime work under Wine. So far RagTime starts, but has several problems which make it non-usable.

To better understand these problems I try to get an insight using winedbg. In an earlier installation which has been broken in the meantime I could see backtraces through wine and RagTime code and could set break points in both worlds. After a fresh compile and install of the current wine source tarball under Redhat Linux 9 with no existing Windows installation, the RagTime part of the backtraces is broken. The backtraces show the same bogus function name for all frames.

My questions: If this problem and a cure for it known?
We had a lot of hazzle under Windows with backtraces created inside RagTime concerning the right version of imagehlp.dll. Could this be the problem here, too? That is, might it be helpful to use a native imagehlp.dll?

Best regards,
Kai

[Robert Shearman]

> -----Original Message-----
> From: [e-mail] [mailto:[e-mail]]On
> Behalf Of Kai Brüning
> Sent: 07 November 2003 08:54
> To: [e-mail]
> Subject: Introduction and winedbg bt problem
>
> Good morning,
>
> my name is Kai Brüning. Our company RagTime GmbH develops and
> markets the desktop publishing package RagTime, which had its
> origin in the Macintosh world 17 years ago. RagTime is currently
> available for Windows and Mac OS 9/X. Now we are evaluating
> whether we can make RagTime work under Wine. So far RagTime
> starts, but has several problems which make it non-usable.
>
> To better understand these problems I try to get an insight using
> winedbg. In an earlier installation which has been broken in the
> meantime I could see backtraces through wine and RagTime code and
> could set break points in both worlds. After a fresh compile and
> install of the current wine source tarball under Redhat Linux 9
> with no existing Windows installation, the RagTime part of the
> backtraces is broken. The backtraces show the same bogus function
> name for all frames.
>
> My questions: If this problem and a cure for it known?
> We had a lot of hazzle under Windows with backtraces created
> inside RagTime concerning the right version of imagehlp.dll.
> Could this be the problem here, too? That is, might it be helpful
> to use a native imagehlp.dll?

Winedbg does not use imagehlp.dll at the moment - it has its own symbol
loader. It was working fine for me yesterday on my own program which uses
MSVC6 symbols.
Are you sure that you are running the program in the same directory as the
symbols?

Rob

[Kai Brüning]

Rob wrote:
> Winedbg does not use imagehlp.dll at the moment - it has its own symbol
> loader. It was working fine for me yesterday on my own program which uses
> MSVC6 symbols.
> Are you sure that you are running the program in the same directory as the
> symbols?

Thanks for the tip. The problem was indeed that the current direction has be set to the program (and .pdb) directory. Otherwise winedbg looks for the .pdb file at many locations but not the right one (I learned this from "warn" debug messages).

Now I have the next problem: RagTime uses C++ exceptions quite extensively. Each  exception drops me into winedbg. Continuing with "pass" works nicely, but the frequency makes debugging pretty much infeasable.

Is there any way to tell winedbg to ignore/auto pass C++ exceptions?

Kai

[Eric Pouech]

> Is there any way to tell winedbg to ignore/auto pass C++ exceptions?

no

A+