J3/04-291

Subject: Interop subgroup comments
From:    Interop subgroup / Bill Long
Date:    3-May-2004
------------------------------------

Comments on N1583.

UK Comments, C5, edit for [279:34]

--- Subgroup agrees with this edit.



UK Comments, C5, edit for [398:8]

--- Because a particular binding label, as a global entity, cannot be
    attached to two different derived types, the difficulty this edit
    was attempting to remedy cannot occur. In addition, there are
    possible cases where the user may want to not expose some
    component names in a bind(c) type to a Fortran user.

[398:8] {Strike the edit proposed in N1583.}



UK Comments, C8, edit for [77:21]

--- The edit as currently written adds text to constraint C540. This
    should not be the case. The beginning of the edit is modified to
    avoid a back reference to the constraint. Replace the edit for
    [77:21] with:

[77:21+] "If the value of the <scalar-char-initialization-expr> after
discarding leading and trailing blanks has nonzero length, it shall be
valid as an identifier on the companion processor."

UK Comments, C8, edit for [265:12]

--- Subgroup agrees with this edit.


UK Comments, C8, edit for [279:28]

--- Debate in WG5 plenary session on the issue of whether PRIVATE
    procedures are allowed to have binding labels lead to straw
    votes. The conclusion was to leave the standard unchanged on this
    issue.  Therefore, the proposed edit at [279:28] should be
    deleted.

[279:28] {Strike the edit proposed in N1583.}


UK Comments, C8, edits for [403:4-6] and [403:8]

--- These edits contain references to procedures, but are in the
    section on variables and common blocks. This was a typo in each
    comment. The edits should be repaced by:

[403:4-6] Replace "If a variable ... ignored." by "If a variable or
common block has the BIND attribute with the NAME= specifier and the
value of its expression after discarding leading and trailing blanks
has nonzero length, the variable or common block has this as its
binding label. The case of letters in the binding label is
significant."

[403:8] Add "Otherwise the variable or common block has no binding
label." at the end of the paragraph.



UK Comment, C8, edit for [403:35]

--- This edit relates to default binding labels for procedures. The
    current FCD omits mention of procedure pointers as exceptions
    which needs correction.  The UK comment also included private
    procedures on the grounds that binding lalels would not be allowed
    for provate procedures.  The second part of the UK edit is
    removed, reflecting the WG5 straw vote.  The edit should be
    changed to:

[403:35] Change "not a dummy procedure" to "not a dummy procedure or
a procedure pointer".



UK Comments, C8, edits for [403:32-34], and [403:36]

--- Subgroup agrees with these edits.

----------------------------------------

US and UK Comments on N1587, [8:17-20] in the TR, N1581.

--- Subgroup agrees that binding labels should not be allowed for
    procedures in a submodule, but proposes an alternate wording and
    move the constraint to [403:36+] in the FCD.

Replace TR:[8:17-20] with:

TR:[10:3+] [After the second paragraph of 15.4.1 insert the following
constraint]

C1506 A procedure defined in a submodule shall not have a binding
label unless its interface is declared in the ancestor module.

-----------------------------------------