.
Last update: 1997-05-20
9945-2-95 _____________________________________________________________________________ Topic: od file operand Relevant Sections: 4.45.4 Defect Report: ----------------------- Date: Fri, 24 Feb 95 00:52:07 GMT From: Don.Cragun@Eng.Sun.COM (Don Cragun) Dear Standards Board, I would like to request a formal interpretation on the following issue concerning the od utility in POSIX.2. Standard: ISO/IEC 9945-2:1993 Topic: od file operand Relevant Sections: 4.45.4 In section 4.45.4 of the description of the od utility's file operand (P371, L7342-7346), it says: "file A pathname of a file to be read. If no file operands are specified, the standard input shall be used. The results are unspecified it the first character of file is a plus sign (+) or the first character of the first file operand is numeric, unless at least one of the -A, -j, -N, or -t options is specified." Although the rationale doesn't say this, it seems obvious that the intent of the last sentence in this description was to allow implementations to provide an obsolescent synopsis form, corresponding to historic practice, along the lines of: od [-bcCDdFfOoSsvXx] [filename] [[+]offset[.][b]] Unfortunately, one common command form: od -c file 10. is not allowed because the offset operand is identified by a numeric as the first character of the "second" file operand. Although the wording in the standard would allow the command: od -c 10. to treat 10 as a decimal offset (rather than as a filename), historic practice (in both System V and BSD implementations) required this to be specified as: od -c +10. To allow implementations to actually provide the historic forms as extensions, the phrase "first character of the first file operand" on P371, L7344-7345 should have been "first character of the second file operand". Was this wording intended to prevent implementations from supporting historic behavior, or was this an editorial mistake? Interpretation response: ------------------------- The standard states the behavior for the file operand of the od cmd and conforming implementations must conform to this. However, concerns have been raised about this which are being referred to the sponsor. Rationale: None Forwarded to Interpretations group: 26 Feb 95 Proposed resolution circulated: May 16th Comments due: June 15th Finalised: June 16th 1995