Minutes, June 12, 2001

Issue discussion:

010302.1    DW_AT_description attribute

    The DW_AT_description attribute provides additional compiler
    or language specific descriptions for an object. The intent
    is that a consumer (i.e., debugger) would print this as part
    of or along with the name of the object.

    Adopted, with the second paragraph of the proposed text as non-
    normative commentary.

010302.2    Byte size and bit offset as object attributes

    This proposal would have permitted size and offset attributes on
    any object, not just on base types. It would also permit the
    dynamic specification of the values for these attributes. The
    committee generally felt that permitting size and offset on any
    object was too broad an application and it was not clear that
    this was needed. Allowing dynamic values for DW_AT_byte_size and
    DW_AT_bit_offset was adopted.

    Partially adopted: Change 5.1, 5.5, 5.6.1, 5.6.6, (and any other
    appropriate sections) to permit expressions as well as constant
    values for DW_AT_byte_size and DW_AT_bit_offset

Other discussion:

    IEEE-ISTO status: Have had further discussions with ISTO about
    becomming a forum. Still have not received participant agreement
    and will follow up. ISTO will provide web site and other services,
    prepare and distribute PR about Dwarf 2 standard, and maintain
    participant and/or user list. Distributed updated version of
    DWARF2 Operating Procedures document (available at
    http://www.eagercon.com/dwarf/DWARF_operating_procedures.pdf)
    for additional comment. (Please do read and comment.)

    Copyright: We expect to replace the existing copyright page
    with copyright information as required by IEEE-ISTO. Will ask
    ISTO for copyright info and doc style comments.
   
Document review:

    Five committee members reviewed and commented on the documents.
    Special thanks to Ron Brender for flying from the East Coast
    to participate in the editorial review. Thanks to the committee
    members who spent the afternoon in this review: Dave Anderson,
    Ron Brender, Felix Burton, Michael Eager, and David Weatherford.

    A number of editorial changes were noted. Please submit any
    other changes, whether major or minor, to the Editor, Ron Brender.

    The description and/or semantics of the stack machine used to
    evaluate Dwarf expressions is less than completely clear. In
    particular, the semantics of DW_OP_piece in 2.4.1.6 and the
    semantics of register encodings in 2.5.1 leave much to the
    imagination. Pointer to member type (DW_TAG_ptr_to_member) may
    not be adequate to describe implementations in adequate detail.
    It isn't clear whether we wish to revise these portions of the
    document at this time.

    Some C++ keywords are not represented in Dwarf. A proposal to
    add a TAG for "mutable" and any others which are missing will be
    prepared.