Content-type: multipart/alternative; Boundary="1__=09BBF98FDFDF3DD38f9e8a93df938690918c09BBF98FDFDF3DD3" --1__=09BBF98FDFDF3DD38f9e8a93df938690918c09BBF98FDFDF3DD3 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: quoted-printable The Linux RPM version already builds both a debug and non-debug version= . I usually, but not always, post the debug version on SF. Maybe we should consider making that a requirement for all versions and platforms? Thanks, W. David Ashley IBM Systems and Technology Group Lab Services Open Object Rexx Team Mobile Phone: 512-289-7506 = "Rick McGuire" = = To Sent by: "Open Object Rexx Developer Mail= ing oorexx-devel-boun List" = ces@lists.sourcef = = cc = 07/18/2007 08:50 Subj= ect AM Re: [Oorexx-devel] Debugging = release version-only failures. = = Please respond to = Open Object Rexx = Developer Mailing = List = = = = Ok, one more thing I meant to mention....we might want to explore addin= g an option to build a release version with debug symbols to the build proce= ss. This could be useful in the future. Rick On 7/17/07, Mark Miesfeld wrote: On 7/17/07, Rick McGuire wrote: > > There's an interesting bug happening with the abs method that's dri= ving me a > bit nuts on what's going on. Here's a simple program that can reproduce the > problem: > > /* Testing Abs() function versus ~abs method */ > test =3D -43.5009 > postestabsmeth =3D test~abs > > This only fails with the release build, which is making it difficul= t to > debug. I added some printfs to the RexxNumberString::abs() method,= and it > appears to complete without error and return. The crash is trying = to > reference storage location 0xffffffff, so there's probably a loop > underflowing somewhere. If I can just figure out where it's dying,= I can > probably figure out what's going on. Any suggestions on how to att= ack this? One thing I was thinking of was gflags combined with the application verifier. But, I don't think it will help here. I just used it and it did not show anything useful. I had also remembered something about a user mode stack trace databas= e with gflags and I was thinking that you could unwind the stack from the crash until you reached the first function in the rexx address space. Then use the map file to figure out the function from the address. But, I can't find any good documentation on how works. I see how to enable it, but I don't see where the trace is being put. gflags is part of the debugging tools for Windows. Here is a link to= the download page: http://www.microsoft.com/whdc/devtools/debugging/default.mspx Install Debugging Tools for Windows 32-bit Version Download page for the latest 32-bit packages. However, it also it not clear to me that any of it will work on a non-debug version of the image. I swore I had seen something that referred to using the tools on non-debug images, but I don't see anything like that now. -- Mark Miesfeld ---------------------------------------------------------------------= ---- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Oorexx-devel mailing list Oorexx-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/oorexx-devel -----------------------------------------------------------------------= -- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Oorexx-devel mailing list Oorexx-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/oorexx-devel = --1__=09BBF98FDFDF3DD38f9e8a93df938690918c09BBF98FDFDF3DD3 Content-type: text/html; charset=US-ASCII Content-Disposition: inline Content-transfer-encoding: quoted-printable

The Linux RPM version already builds both a debug and non-debug vers= ion. I usually, but not always, post the debug version on SF. Maybe we = should consider making that a requirement for all versions and platform= s?

Thanks,
W. David Ashley
IBM Systems and Technology Group Lab Services
Open Object Rexx Team
Mobile Phone: 512-289-7506


3D"Inactive"Rick McG= uire" ---07/18/2007 08:51:19 AM---"Rick McGuire" <obj= ect.rexx@gmail.com>

=
          "Rick McGuire" <object.rexx@gmail.= com>
          Sent by: oorexx-devel-bounces@lists.sourceforge.net

          07/18/2007 08:50 AM
          Please respond to
          Open Object Rexx Developer Mailing List <oorexx-devel@lists.sourcefo= rge.net>

=
3D=
To
3D""
"Open Object Rexx Developer Mailing List" &l= t;oorexx-devel@lists.sourceforge.net>
3D=
cc
3D""
3D=
Subject
3D""
Re: [Oorexx-devel] Debugging release version-only fail= ures.
3D=3D""

Ok, one more thing I meant to mention....we might want= to explore adding an option to build a release version with debug symb= ols to the build process. This could be useful in the future.

Rick

On 7/17/07, Mark Miesfeld <miesfeld@gmail.com> wrote:

    On 7/17/07, Rick McGuire <object= .rexx@gmail.com> wrote:
    >
    > There's an interesting bug happening with the abs method that's dr= iving me a
    > bit nuts on what's going on. Here's a simple program that can rep= roduce the
    > problem:
    >
    > /* Testing Abs() function versus ~abs method */
    > test =3D -43.5009
    > postestabsmeth =3D test~abs
    >
    > This only fails with the release build, which is making it difficu= lt to
    > debug. I added some printfs to the RexxNumberString::abs() method= , and it
    > appears to complete without error and return. The crash is trying= to
    > reference storage location 0xffffffff, so there's probably a loop =
    > underflowing somewhere. If I can just figure out where it's dying= , I can
    > probably figure out what's going on. Any suggestions on how to at= tack this?

    One thing I was thinking of was gflags combined with the application verifier. But, I don't think it will help here. I just used it and it did not show anything useful.

    I had also remembered something about a user mode stack trace database<= br> with gflags and I was thinking that you could unwind the stack from the crash until you reached the first function in the rexx address
    space. Then use the map file to figure out the function from the
    address. But, I can't find any good documentation on how works. I
    = see how to enable it, but I don't see where the trace is being put. <grin>

    gflags is part of the debugging tools for Windows. Here is a link to the download page:


    http://www.microsof= t.com/whdc/devtools/debugging/default.mspx

    Install Debugging Tools for Windows 32-bit Version
    Download page for the latest 32-bit packages.

    However, it also it not clear to me that any of it will work on a
    non-debug version of the image. I swore I had seen something that
    = referred to using the tools on non-debug images, but I don't see
    anything like that now.

    --
    Mark Miesfeld

    -----------------------------------------------------------------------= --
    This SF.net email is sponsored by DB2 Express
    Download DB2 Express C - the FREE version of DB2 express and take
    control of your XML. No limits. Just data. Click to get it now.
    <= u>
    http://sourceforge.net/powerbar/db2/<= /u>
    _______________________________________________
    Oorexx-devel mailing list
    <= br> Oorexx-devel@lists.sourceforge.net
    https://lists.sourcefor= ge.net/lists/listinfo/oorexx-devel
-------------------------------------------------------------------= ------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://source= forge.net/powerbar/db2/___________________________________= ____________
Oorexx-devel mailing list
Oorexx-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/oorexx-devel

= --1__=09BBF98FDFDF3DD38f9e8a93df938690918c09BBF98FDFDF3DD3--