JTC1/SC22/WG15 - RCT N070 September 15, 1995 Proposed Disposition of Comments ISO/IEC JTC1/SC22/WG15 ISO/IEC Project JTC1.22.14515-1 ISO/IEC Document 14515-1 CD TITLE: Information Technology - Test Methods for Measuring Conformance to POSIX; Part 1: System Interfaces ==================== UK Comments ======================================== The UK vote comment on CD 14515-1: POSIX - Test Methods for Measuring Conformance to POSIX, Part 1 System Interfaces. YES vote, with the following comments: The standard forming the foundation for this work, ISO/IEC 9945-1:1990 is five years old and is in the process of being amended; the IEEE standard from which this work is derived was published in 1992. The UK would recommend that this work, CD 14515-1, be progressed as it stands, unchanged, to reinforce current industry standards and practice. The UK would recommend that the following points be forwarded to the US development body for consideration at the next revision of the 2003.1 work: . Improve unspecific wording, eg: 1.4.7.2 Recommendations: (lines 434 - 437) ...PCTS.1 should not rely solely on the return value from that compiler or translator , but should also examine any messages produced during test execution. For what? As it stands this is a no-op; some action should also be suggested/recommended/ required. 1.4.12.1.1 Requirements: ...the signal handling routine behaves correctly... What does this mean? Shall abort? Shall behave as specified by the appropriate definition? Shall ignore the call? 1.4.16.1 Requirements: The PCTS.1 shall supply all files... Do not create them on the target system with the file creating utility. Does this mean exactly what it says, or should it also deny use of all file creation techniques available on the target system? Table 2-1 Typographical Conventions: While is a valid example of the lexography of an 'include' file in a C program source, it is not a "C language header" 3.2 Process Termination: atexit() is defined in the ISO C standard (Section 7.10.4.2). The wording at lines 653 and 698 should be revised to reflect this. ==================== End of UK Comments ============================ Per the UK request, these recommendations will be forwarded to the US Development group for consideration in the future revision and/or amendment processes. As the UK, in some cases, is looking for clarification of the standard we wish to remind the UK that this can be handled via the interpretation process within IEEE PASC. ==================== Sweden Comments =============================== The comment with the negative ballot from Sweden reads: "Standardization of test methods should be avoided. The greatest disadvantage is, that the test method in practice may supersede the actual standard, making that standard useless. In any case there are two parallel documents, that need to be read (and maintained) There is also a lot of work needed to prepare this test standard, which is generally quite parallel to that of the original standard. In this proposed standard, there are large similarities and duplication between the POSIX.1 standard, and the test method standard. Two examples: . The ermos (chap 2.4) have all ermos listed in both standards . In the environment (chap 2.6) only two of the 12 environment variables are listed in the test standard, thus making those variables more "important" than the others, without any such preference being visible to the POSIX standard. Technical report may be an acceptable way of obtaining a uniform testing. Then it is quite clear, that the standard is the main document, and that the test document is secondary to the actual standard." ==================== End of comment from Sweden ======================= This ballot is procedural in nature and therefore no action will be forthcoming from the WG15. The October 1994 JTC1 plenary approved a JTC1 Policy Statement on Conformity Assessment that makes it quite clear that test methods are regarded in JTC1 as being appropriate for International Standardization. Thus, the comment is out of line with JTC1 policy. If Sweden wishes to take this matter further the proper place for such general policy discussion is the JTC1 Special Working Group on Conformity Assessment, in which Sweden has always consistently supported such standardization. ISO/IEC Project JTC1.22.14515-1, was accepted by the SC22 as a project to produce a Test Methods Standard. WG15 would point out to all JTC1 member bodies this standard has the intent of defining a means of measuring conformance to ISO/IEC 9945-1:1990, and defers to ISO/IEC 9945-1:1990 in any question of interpretation of the technical specifications of the POSIX System Application Program Interface. Within the scope of the document it so states this. Therefore we believe it is clear that ISO/IEC 9945-1:1990 is the main document in regards to specifying the POSIX System Application Program Interface.