ISO/IEC JTC1/SC22/WG5 N1760 Responses to CD comments John Reid WG5 has considered all the comments with the CD ballots, see ISO/IEC JTC1/SC22/N4405 (or ISO/IEC JTC1/SC22/WG5 N1740) and responds as follows. Here, documents ISO/IEC JTC1/SC22/WG5 Nxxxx are referenced as Nxxxx and INCITS/PL22.3 documents are referenced as 08-nnn. 1. Canada 1.1 Issues with coindexed objects 1.1.1 08-143r3 Disallow coindexed object as a data-ref in a procedure-designator WG5 has made changes to address this issue, see 08-301r1, 08-309. 1.1.2 08-267 Restriction on coindexed objects with allocatable components WG5 has made changes to address this issue, see 08-308r1. 1.1.3 08-199 Disallow finalizations on coindexed objects WG5 has made changes to address this issue, see 08-308r1. 1.2 Issues with allocatable coarrays These issues have been resolved, see 08-158, 08-162r2. 1.3 Lack of locking mechanism for coarrays A locking mechanism for coarrays has been added, see 08-300. 2. UK 2.1 Extent of revision WG5 could not agree on any significant reduction in the size of Fortran 2008. WG5 discussed the schedule (see N1693) and decided to review the situation at the next meeting (May 2009). 2.2 MPI WG5 decided that the TR on further interoperabililty with C was appropriate for addressing concerns with the MPI interface to Fortran, noting that the TR and the next standard should be published at about the same time. The C Programming Language concepts of void* buffers and of asynchronous modification of a buffer logically following the return from a C function are to be treated. The specific motivation is for the MPI library, but the intention is to apply generally to any such C functions (as might be found in any library providing communications or storage services). 2.3 Coarrays - technical aspects, see N1745 Volatile coarrays are no longer excluded from the rule about referencing a coarray in a segment that is unordered with respect to a segment in which it is defined. For the replacement, involving scalar variables of atomic kind, see 08-297r1. The other comments are addressed in 08-290r2 and 08-304. 2.4 Coarrays - disposition WG5 discussed the advantages and disadvantages of moving the core coarray features to another part of the standard. The majority opinion was in favour of retaining the core coarray features in the standard. 2.5 Description of Coarrays, see N1744 These comments are addressed in 08-290r2. 2.6 Generalized output editing A similar feature has been added, see 08-296r2. 3. Germany The proposed changes to Bessel functions were not accepted. WG5 agrees with the rationale in 08-211r3. An editorial change to the specification of the intrinsic SIZE has been made. 4. USA (see WG5 N1741) 4.1 Editorial and technical repairs WG5 accepts the editorial and technical repairs, except that items 08-199r1, 08-207, 08-214, and 08-216r2 were seen as feature requests. 4.2 08-199r1 Disallow finalization of coindexed objects This is handled by 08-308r1. 4.3 08-207 Allow STORAGE_SIZE in a in an elemental procedure This is handled in 08-299r1 by deleting constraint C1288. The rationale is in 08-302. 4.4 08-214 Prohibit image control in DO CONCURRENT This feature has been added. 4.5 08-216r2 Pointer and target on same image WG5 accepts the technical content of this proposal and asks INCITS/PL22.3 to work further on the edits needed. 4.6 08-143r3 Prohibit coindexed object as a in a This is handled by 08-301r1 and 08-309. 4.7 08-147r1 Restrictions on VALUE are mostly unnecessary This change has been made. 4.8 08-154r1 Change rules for access from submodule to ancestor module by USE association This feature has been added. 4.9 08-165r1 Relax restrictions on ASYNCHRONOUS/VOLATILE This feature has been added. 4.10 08-166r1 Simply contiguous TARGET dummy arguments are guaranteed to be associated This feature has been added. 4.11 08-178 Allow internal procedures in generics This feature has been added. 4.12 08-181 Allow VALUE without INTENT for pure procedure dummy arguments This feature has been added. 4.13 08-251r3 Allow g0.d edit descriptor A similar feature has been added, see 08-296r2. 4.14 08-254r2 New TYPE_IS_EXTENSIBLE intrinsic function This problem has been handled differently by 08-281r2. See also 08-303. 4.15 08-256r2 Add locks to coarray core subset This feature has been added with the changes in 08-300. 4.16 08-266r1 New COMPLEX intrinsic function WG5 did not accept this suggestion.