0001557: Better wording to describe FD_CLOEXEC.

Notes
(0005632) dannyniu (reporter) 2022-01-23 06:53	I made a few mistakes in the desired action. I'd like to correct: Desired Action: Change the description to > The FD_CLOEXEC file descriptor flag associated with > the new file descriptor shall be set if and only if > the O_CLOEXEC flag is set in oflag Additionally, similar changes for *_CLOFORK and other interfaces for creating file descriptors should probably be made for Issue-8.

(0005633) geoffclare (manager) 2022-01-24 10:35	There is no grammatical ambiguity. The purpose of this text is to require that FD_CLOEXEC is cleared for the fd when O_CLOEXEC is not set in the flags value. It intentionally says nothing about the opposite case. The requirement for FD_CLOEXEC to be set for the fd when O_CLOEXEC is set in the flags value is stated lower down in the part about O_CLOEXEC. The proposed change would add a redundant duplication of this requirement.

(0005651) dannyniu (reporter) 2022-02-02 08:07 edited on: 2022-02-02 08:08	Well, then I guess it's an "if it ain't broken don't fix it" situation. But I guess the least that may be considered is to make the wordings consistent across open(2), pipe2(2), dup3(2), accept4(2). e.g.: At the beginning, say: > The FD_CLOEXEC and FD_CLOFORK file descriptor flags > shall be clear, unless corresponding bits > in the `flag` argument is set. Then in the bullet-list for the *_CLO{EXEC,FORK}, say: > If set, the [said] flag for the new file descriptor shall be set, or: > Atomically set the [said] file descriptor flag on [the file descriptor]. So I guess the project for this bug may need to be changed to Issue 8.

(0005784) geoffclare (manager) 2022-04-07 15:33	Re: Note: 0005651 In order to have consistent wording across open(), pipe2(), dup3(), accept4(), it would be necessary to drastically rearrange the descriptions of the last three, so that the new function with the flags argument is described first and then the old function is said to be equivalent to that function with a flags argument of 0. (Instead of the new function being said to be equivalent to the old function, except ...) This seems like a lot of work for little gain.

Issue History
Date Modified	Username	Field	Change
2022-01-23 03:41	dannyniu	New Issue
2022-01-23 03:41	dannyniu	Name	=> DannyNiu/NJF
2022-01-23 03:41	dannyniu	Organization	=> Individual
2022-01-23 03:41	dannyniu	Section	=> open()
2022-01-23 03:41	dannyniu	Page Number	=> 1408
2022-01-23 03:41	dannyniu	Line Number	=> 46758-46762
2022-01-23 06:53	dannyniu	Note Added: 0005632
2022-01-24 10:35	geoffclare	Note Added: 0005633
2022-02-02 08:07	dannyniu	Note Added: 0005651
2022-02-02 08:08	dannyniu	Note Edited: 0005651
2022-04-07 15:33	geoffclare	Interp Status	=> ---
2022-04-07 15:33	geoffclare	Note Added: 0005784
2022-04-07 15:33	geoffclare	Status	New => Closed
2022-04-07 15:33	geoffclare	Resolution	Open => Rejected
2022-04-28 15:50	Don Cragun	Relationship added	related to 0001577

Aardvark Mark IV