🚀 go-pugleaf

In article <PINKAS.92Mar24181505@caraway.intel.com>
pinkas@caraway.intel.com (Israel Pinkas) writes:

| I have heard rumors that Solaris will use ELF binaries but that the
| debugging information will be STABS based.  (STABS is the COFF format.
| SunOS for the Sun386i used COFF.)

STABS is the BSD debug format, not COFF.  STABS encodes the debug
information using the name field as a string that the debugger has to
parse, a type field and two integer fields, one of which can be
relocated (line numbers are the only stab that uses the second field).
STABS are intermixed with other entries in the symbol table.  Anything
that the compiler and debugger can agree upon can be passed through.
Since there is no central stabs authority, you can have problems when
the compiler and debugger disagree.

COFF uses a binary encoding in it's own section in the object format,
based on the data structures in the original Ritchie C compiler, which
compresses the basic type information into 16 bits (4 bit basic type,
6 fields of 2 bits each to represent pointer/array/function), with
extra words following to provide things like array sizes.  COFF is
extermely crippling when it comes to describe non-C languages.

DWARF is also a binary encoding used by most of the systems using the
ELF object file format (except for MIPS systems which use the ECOFF
debug format).  It is much more flexible than COFF in the languages it
can support.  The USL programming language SIG is putting the revised
format out for public review soon (maybe it's already out on the
streets).
--
Michael Meissner	email: meissner@osf.org		phone: 617-621-8861
Open Software Foundation, 11 Cambridge Center, Cambridge, MA, 02142

You are in a twisty little passage of standards, all conflicting.