[Bug c++/11134] DWARF Problems with const int globals

[dberlin@dberlin.org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11134



------- Additional Comments From dberlin@dberlin.org  2003-06-09 14:29 -------
Subject: Re:  DWARF Problems with const int globals

He *must* be referring to DWARF2 already.
DW_DLV_NO_ENTRY is an error value returned by the libdwarf dwarf2 library
SGI wrote (and Intel rewrote, and Ulrich Drepper is writing for elfutils).


On Mon, 9 Jun 2003, drow@mvista.com wrote:

> PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11134
>
>
>
> ------- Additional Comments From drow@mvista.com  2003-06-09 14:21 -------
> Subject: Re:  New: DWARF Problems with const int globals
>
> This isn't really an answer to your bug report, but it counts as a
> workaround...
>
> On Mon, Jun 09, 2003 at 02:15:57PM -0000, drkent@lanl.gov wrote:
> > PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.
> >
> > http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11134
> >
> >            Summary: DWARF Problems with const int globals
> >            Product: gcc
> >            Version: 3.3
> >             Status: UNCONFIRMED
> >           Severity: normal
> >           Priority: P2
> >          Component: c++
> >         AssignedTo: unassigned@gcc.gnu.org
> >         ReportedBy: drkent@lanl.gov
> >                 CC: gcc-bugs@gcc.gnu.org
> >
> > I am developing a dynamic instrumentation tool that uses DWARF and have
> > encountered a few problems that could be from the dwarf information generated by
> > g++  3.3.
> >
> > There are a few global variables defined as "const int".  When the tool parses
> > the binary, it attempts to locate these variables and then gives a
> > DW_DLV_NO_ENTRY error which means that the attribute is not where the dwarf
> > information says it is.
> >
> > I have avoided these problems by changing the global const ints to #define
> > statments, and the code runs properly.
> >
> > Please contact me with any questions.
>
> If you are developing this tool, please, please, please use DWARF-2
> instead!  The DWARF writer is on its last legs, especially in C++, and
> the format is inferior.
>