From suehiro@jrd.dec-j.co.jp Thu Jun 27 09:31:50 1996 Received: from jnet-gw-2.dec-j.co.jp (jnet-gw-2.dec-j.co.jp [202.34.226.25]) by dkuug.dk (8.6.12/8.6.12) with ESMTP id JAA23197 for ; Thu, 27 Jun 1996 09:31:38 +0200 Received: by jnet-gw-2.dec-j.co.jp (8.6.12+win/JNET-GW-951211.1); id QAA05852; Thu, 27 Jun 1996 16:34:12 +0900 Received: from cobra.jrd.dec.com by garfield.jrd.dec.com (8.7.3+usagi/JULT-4.4-gar) id QAA00916; Thu, 27 Jun 1996 16:31:58 +0900 Received: from localhost by cobra.jrd.dec.com (5.65v3.0/JOSF-3.0-cobra) id AA23499; Thu, 27 Jun 1996 16:34:32 +0900 Message-Id: <9606270734.AA23499@cobra.jrd.dec.com> To: sc22wg15@dkuug.dk, posix-dot2@pasc.org Subject: Comments on .2b Japanese proposals Date: Thu, 27 Jun 1996 16:34:32 +0900 From: Yoichi Suehiro This is a response to an action item assigned to Japan at WG15 Copenhagen meeting. We expect that this document can be an input to July IEEE.2 meeting. regards, Yoichi Suehiro ============================================================================= Comments on SC22/WG15 N640 about the proposals from Japan ========================================================== To: IEEE POSIX.2 group From: ITSCJ/SC22/POSIX WG SC22/WG15 members Japan member body reviewed WG15/N640 which described the responses from IEEE.2 group to .2b related proposals by WG15 national bodies. Here are our review comments on the following topics which Japan member body had ever proposed. 1. LC_COLLATE extension for user specified weight names 2. LC_CTYPE extension for locale-specific character mappings ============================================================================= 1. LC_COLLATE extension for user specified weight names First, we'd like to thank IEEE POSIX.2 group for their detail analysis on the proposal. It is true that the proposed syntax extension can be covered by awk script and we've known it. In fact, we wanted to have more than that. Japanese sorting requirements are very complicated and do not match well with the POSIX locale-based collation model. We cannot satisfy the requirements effectively by POSIX locale definition without having several extensions. The proposal for LC_COLLATE syntax extension was one of them. We had not been successful to present complete proposals to Standards bodies so far. It seems that now is the time we withdraw this proposal as well as state-dependent encoding support. It is important that proposals have enough interests from implementors before being standardized. Full support of Japanese sorting requirements cost too high for UNIX vendors to change their current localedef implementation. Therefore, Japan member body decided to accept the proposed resolution in N640 and no longer pursue the proposal. 2. LC_CTYPE extension for locale-specific character mapping We could not find any responses to this proposal. [See attachement below] The proposal was specified in SC22/WG15 N602 [RIN number N158](Japan Action Item Report to WG15, October 1995). The following text is the excerpt from Japan Action Item Report to WG15/RIN 1996-MAY meeting. #9510-10 Rapporteurs to review and respond to the Japanese proposal for # LC_CTYPE extension for locale-specific character mapping in N158. # # This is the proposal mentioned in the response to 9510-05 above. # Two of the ISO C Amendment functions (wctrans, towctrans) # requires this extension to behave in a locale dependent way # on a POSIX system. Japan believes that the proposal is almost # complete and has no objections. [End] ============================================================================== [Attachment: LC_CTYPE extension proposal for locale-specific character mapping] Source: Japan Title: Japanese proposal to POSIX.2b on LC_CTYPE extension for locale-specific character mapping Status: Japanese position Short description: Japan proposes that LC_CTYPE locale definition should be extended to allow locale-specific character mappings to be specified. This extension is necessary to implement wctrans() and towctrans() functions in ISO C amendment on a POSIX conforming system. Text of contribution: ---------------------------------------------------------------------------- [Note: The page numbers refer to the ones of P1003.2/D10.] Sect 2.5 (Locale) PROPOSAL. Page 8-9,12: Problem: The LC_CTYPE (2.5.2.1) locale definition should be enhanced to allow user-specified additional character mapping, similar in the concept to the user-specified additional character class. In the Amendment of ISO C standard, extended character mapping functions (wctrans/towctrans) are specified. The following proposed extension will serve for the machinery to define locale specific character mappings used by the functions. Without having this extension, POSIX conforming systems need to have their own extensions to implement ISO C Amendment specifications. Proposal:[LC_CTYPE extension for specifying character mapping] The proposed extension for character mapping is similar to the extension of character class, which is already specified in .2b draft. New keyword 'charconv' is introduced to define locale-specific character mappings instead of 'charclass' keyword for character class. The way of defining character mapping is not extended with this proposal. The same specification for toupper/tolower mapping can be used for locale-specific character mappings. EXAMPLE: LC_CTYPE # define the names of locale-specific character mappings charconv tojkata;tojhira # tojkata: hiragana => katakana mapping tojkata (,);(,);\ .....definition..... # tojhira: katakana => hiragana mapping tojhira (,);(,);\ .....definition..... END LC_CTYPE [Proposed extension to .2b text] [Page 8] => 2.5.2.1 LC_CTYPE. Add the following keyword items after the item labeled tolower: charconv Define one or more locale-specific character mapping names as strings separated by semicolons. Each named character mapping can then be defined subsequently in the LC_CTYPE definition. A character mapping name shall consist of at least one and at most fourteen bytes of alphanumeric characters from the portable filename character set. The first character of a character mapping name cannot be a digit. The name cannot match any of the LC_CTYPE keywords defined in this standard. charconv-name Define the named locale-specific character mapping. In the POSIX Locale, the locale-specific named character mapping need not exist. If a mapping name is defined by a charconv keyword, but no character mappings are subsequently assigned to it, this is not an error; it shall represent a mapping without any character pairs belonging to it. [Page 12] => 2.5.3.1 Locale Lexical Conventions. Add the following token description: CHARCONV A string of alphanumeric characters from the portable character set, the first of which shall not be a digit, consisting of at least one and at most fourteen bytes, and optionally surrounded by double-quotes. [Page 12] => 2.5.3.2 Locale Grammar. Modify the ctype_keyword and charconv_keyword descriptions as follows: ctype_keyword : charclass_keyword charclass_list EOL | charwidth_keyword charclass_list EOL | defwidth_keyword defwidth_value EOL | charconv_keyword charconv_list EOL | 'charclass' charclass_namelist EOL | 'charconv' charconv_namelist EOL ; charconv_namelist : charconv_namelist ';' CHARCONV | CHARCONV ; charconv_keyword : 'toupper' | 'tolower' | CHARCONV ; ---------------------------------------------------------------------------- [END] =============================================================================