0000974: several issues with awk's ARGC/ARGV

ID	Project	Category	View Status	Date Submitted	Last Update

0000974	1003.1(2013)/Issue7+TC1	Shell and Utilities	public	2015-08-10 11:09	2024-06-11 09:02

Reporter	stephane	Assigned To
Priority	normal	Severity	Editorial	Type	Error
Status	Closed	Resolution	Accepted As Marked

Name	Stephane Chazelas
Organization
User Reference
Section	awk
Page Number	2460
Line Number	78773-78782
Interp Status	Approved
Final Accepted Text	See 0000974:0003231


Summary	0000974: several issues with awk's ARGC/ARGV
Description	The description of ARGC/ARGV in the awk specification implies that it contains the list of arguments (to the exclusion of progname) while all awk implementations that I tried have the basename of the argv[0] that awk received in ARGV[0]. It also doesn't specify what should happen if one sets ARGC to 0 or 1, or what the effect of assigning ARGV[0] may be. Or of assigning all of ARGV[1]..ARGV[ARGC-1] to null. It doesn't say what's the consequence of doing: c=ARGC; ARGC=0; ARGC=c (whether setting ARGC=0 would truncate the ARGV array).
Desired Action	Specify that ARGV[0] is the basename(argv[0]) and is otherwise not used and assigning a value to it has no side effect (it is not used for error reporting by awk for instance in any implementation I tried). Specifify that setting ARGC to 0 or 1 causes awk to read input from stdin if done from the BEGIN statement. But otherwise setting ARGC to a number that would exclude the currently processed file (as in `awk '{ARGC=1; print}' file`) is unspecified (some awk stop processing after the current record, some at the end of the current file). Specify that setting or all of ARGV[1]..ARGV[ARGC-1] to null where ARGC > 1 in the BEGIN statement is unspecified (mawk and gawk read from stdin, but heirloom's just exits at the end of the BEGIN statement). Specify that modifying the value of ARGC doesn't affect the content of ARGV.
Tags	tc3-2008

geoffclare 2015-08-10 13:53 manager bugnote:0002787	All certified UNIX systems put the full pathname from argv[0] in ARGV[0]. So does GNU awk if you set POSIXLY_CORRECT in the environment: $ /usr/bin/gawk 'BEGIN { print ARGV[0] }' /dev/null /usr/bin/gawk $ (unset POSIXLY_CORRECT; /usr/bin/gawk 'BEGIN { print ARGV[0] }' /dev/null) gawk

stephane 2015-08-10 16:02 reporter bugnote:0002788	We may also want to clarify the expected outcome for things like: awk -v ARGC=1 '{print}' file1 file2 (should it read from stdin or file1 and file2?) or awk '{print}' ARGC=3 or awk 'BEGIN{ARGV[1] = "ARGC=3"};{print}' file (should it read from stdin?)

rhansen 2016-05-19 16:31 manager bugnote:0003231 Last edited: 2016-05-19 16:43	Interpretation response ------------------------ The standard does not speak to this issue, and as such no conformance distinction can be made between alternative implementations based on this. This is being referred to the sponsor. Rationale: ------------- None. Notes to the Editor (not part of this interpretation): ------------------------------------------------------- On page 2460 line 78773-78779 section awk, change: ARGC The number of elements in the ARGV array. ARGV An array of command line arguments, excluding options and the program argument, numbered from zero to ARGC-1. The arguments in ARGV can be modified or added to; ARGC can be altered. As each input file ends, awk shall treat the next non-null element of ARGV, up to the current value of ARGC-1, inclusive, as the name of the next input file. Thus, setting an element of ARGV to null means that it shall not be treated as an input file. The name '-' indicates the standard input. If an argument matches the format of an assignment operand, this argument shall be treated as an assignment rather than a file argument. to: ARGC A number determining when the iteration described for ARGV stops. When an awk program starts, ARGC shall be initialized to the number of elements in the ARGV array. ARGC can be updated by the awk program and by assignment operands. If ARGC is set to a value less than 1, the behavior is unspecified. It is unspecified whether alterations to ARGC can be made using the -v option. ARGV An array containing, initially, the command name (see [xref to 2.9.1]) used to invoke awk in <tt>ARGV[0]</tt> and the command line arguments, if any, excluding options and the program operand, in <tt>ARGV[1]</tt> through <tt>ARGV[ARGC−1]</tt>. The elements in ARGV can be assigned new values or deleted, and new elements can be added. It is unspecified whether this alteration to ARGV can be made using the -v option. After processing the BEGIN actions, if any, awk begins interating over the elements of ARGV, processing them as if they were argument operands. It shall behave as if the implementation maintains an internal counter that is initialized to 1 and increments by 1 at the end of each iteration. For each iteration, the following shall occur: If the internal counter is greater than or equal to the current value of ARGC and no file operands have been processed, awk shall set FILENAME to '-' and process standard input as if it was given as a file operand. The internal counter shall not be incremented at the end of this iteration. Otherwise, if the internal counter is greater than or equal to the current value of ARGC, the iterations shall stop and processing of the END actions, if any, shall begin. Any ARGV elements with index values greater than or equal to ARGC shall not be processed as argument operands. Otherwise, if the element <tt>ARGV[</tt>internal counter value<tt>]</tt> does not exist, it is unspecified whether that element is created. No other action shall be taken. Otherwise, if <tt>ARGV[</tt>internal counter value<tt>]</tt> is a null string, no action is taken. Otherwise, if <tt>ARGV[</tt>internal counter value<tt>]</tt> matches the format of an assignment operand (see OPERANDS), awk shall proccess the assignment. Otherwise, <tt>ARGV[</tt>internal counter value<tt>]</tt> shall be treated as a file operand, FILENAME shall be set to that value, and the named file, or standard input if the value is '-', shall be processed as an input file. Since only non-null elements are processed, setting an element of ARGV to the null string or deleting it means that it shall not be treated as an argument operand. On page 2460, lines 78798-78800 section awk EXTENDED DESCRIPTION, change the definition of the FILENAME variable: A pathname of the current input file. Inside a BEGIN action the value is undefined. Inside an END action the value shall be the name of the last input file processed. to: The pathname used to open the current input file, or '-' if the file is standard input. Inside a BEGIN action FILENAME shall be unset. Inside an END action the value shall be the name of the last input file processed. If an application changes the value of FILENAME, the results are unspecified. On page 2453 line 78479 section awk, change: If no file operands are specified, or if a file operand is '-', the standard input shall be used. to: If no file operands or their equivalents, achieved by modifying the awk variables ARGV and ARGC, are specified, or if a file operand is '-', the standard input shall be used. On page 2453 line 78502 section awk, change: If there are no file arguments, assignments shall be executed before processing the standard input. to: If there are no file arguments or their equivalents, achieved by modifying the awk variables ARGV and ARGC, assignments shall be executed before processing the standard input. On page 2453 line 78506 section awk, change: The standard input shall be used only if no file operands are specified, or if a file operand is '-', or if a progfile option-argument is '-'; see the INPUT FILES section. to: The standard input shall be used only if no file operands or their equivalents, achieved by modifying the awk variables ARGV and ARGC, are specified; or if a file operand, or its equivalent, is '-'; or if a progfile option-argument is '-'; see the INPUT FILES section. On page 2453 line 78514 section awk, change: Standard input in the absence of any file operands to: Standard input in the absence of any file operands, or their equivalents On page 2478 after line 79587 section awk APPLICATION USAGE, add a new paragraph: To specify a file operand naming a file with a name containing an <equals-sign>, users can use "./" as the first two characters of a relative file pathname that starts with an <underscore> or an alphabetic character to keep the file operand from being interpreted as an assignment operand. Similarly, "./-" can be used to access a file named '-' in the current directory rather than use standard input.

ajosey 2016-12-15 18:11 manager bugnote:0003523	Interpretation proposed: 15 Dec 2016

ajosey 2017-01-18 15:24 manager bugnote:0003551	Interpretation Approved: 18 Jan 2017

Date Modified	Username	Field	Change
2015-08-10 11:09	stephane	New Issue
2015-08-10 11:09	stephane	Name	=> Stephane Chazelas
2015-08-10 11:09	stephane	Section	=> awk
2015-08-10 11:09	stephane	Page Number	=> 2460
2015-08-10 11:09	stephane	Line Number	=> 78773-78782
2015-08-10 13:53	geoffclare	Note Added: 0002787
2015-08-10 16:02	stephane	Note Added: 0002788
2016-05-19 16:31	rhansen	Note Added: 0003231
2016-05-19 16:35	rhansen	Note Edited: 0003231
2016-05-19 16:38	rhansen	Note Edited: 0003231
2016-05-19 16:39	rhansen	Note Edited: 0003231
2016-05-19 16:43	rhansen	Note Edited: 0003231
2016-05-19 16:44	rhansen	Interp Status	=> Pending
2016-05-19 16:44	rhansen	Final Accepted Text	=> See 0000974:0003231
2016-05-19 16:44	rhansen	Status	New => Interpretation Required
2016-05-19 16:44	rhansen	Resolution	Open => Accepted As Marked
2016-05-19 16:44	rhansen	Tag Attached: tc3-2008
2016-12-15 18:11	ajosey	Interp Status	Pending => Proposed
2016-12-15 18:11	ajosey	Note Added: 0003523
2017-01-18 15:24	ajosey	Interp Status	Proposed => Approved
2017-01-18 15:24	ajosey	Note Added: 0003551
2019-10-21 09:22	geoffclare	Status	Interpretation Required => Applied
2021-05-06 15:10	nick	Relationship added	related to 0001451
2024-06-11 09:02	agadmin	Status	Applied => Closed

View Issue Details

Relationships

Activities

Issue History