src - OpenBSD base system

Age	Commit message (Collapse)	Author
2014-07-12	If a constant string needs a name, use a static const array instead	Philip Guenther
	of a pointer or non-const array, as that minimizes the symbols, maximizes the placement into read-only memory, and avoids warnings from gcc -Wformat=2 when they're used as format strings.
2014-07-11	using COPTS is probably better than CFLAGS	Theo de Raadt

2014-07-11	correct readlink termination. from Doug Hogan	Ted Unangst

2014-07-11	clarify a comment about readlink. from Doug Hogan	Ted Unangst

2014-07-10	sprinkle in -Werror-implicit-function-declaration to disable one of the	Ted Unangst
	worst C misfeatures. ok deraadt miod
2014-07-08	sys/user.h can now be substantially gutted.	Theo de Raadt
	ok guenther
2014-07-04	Track whether a process is a zombie or not yet fully built via flags	Philip Guenther
	PS_{ZOMBIE,EMBRYO} on the process instead of peeking into the process's thread data. This eliminates the need for the thread-level SDEAD state. Change kvm_getprocs() (both the sysctl() and kvm backends) to report the "most active" scheduler state for the process's threads. tweaks kettenis@ feedback and ok matthew@
2014-06-19	Fix memory leak in digest_file() on ferror(). OK tedu@ lteo@	Todd C. Miller

2014-06-12	somehow i missed -vW as extensions;	Jason McIntyre
	From: Martin Natano
2014-06-05	Add support for COLUMNS env variable, inspired by FreeBSD but with a dash	Philip Guenther
	of strtonum() from millert@ sprinkled on top. Also, we've always supported TZ for formatting dates, so say so. ok jmc@ millert@
2014-06-04	undo my "no TZ support" notes; guenther says they're wrong	Jason McIntyre
	ok guenther
2014-06-03	-a, -o, and () are marked by posix as obsolete; the finer details are	Jason McIntyre
	left to the user to investigate, on guenther's advice; ok guenther
2014-05-29	document control characters which are extensions;	Jason McIntyre

2014-05-29	flesh out STANDARDS;	Jason McIntyre

2014-05-29	for -a and -g, do not say "as per posix": it adds no value to the user, and	Jason McIntyre
	STANDARDS already notes these flags are compliant; it also stops us looking redundant referring to older posix revs; for -g, also note that output can safely be used by the shell; while here, zap some unneeded Fl i missed in previous;
2014-05-29	some basic formatting fixes;	Jason McIntyre

2014-05-28	onocr suppresses the cr, not outputs it;	Jason McIntyre
	confirmed by/ok millert guenther
2014-05-28	document that posix defaults to -L if no options are given, and tweak previous;	Jason McIntyre

2014-05-28	In his continuing quest to improve our documentation, jmc@ noticed that	Philip Guenther
	we didn't support the -L option required by POSIX, and didn't document the -P option that we did support. Fix that with code and doc bits from NetBSD, but correcting a minor noncompliance in the -L support and applying the jmc@ and sobrado@ magic to the manpage. This commit does not change the default behavior. ok jmc@ millert@ sobrado@
2014-05-27	document SIGALRM, since it's handling is somewhat special, and use	Jason McIntyre
	a standard EXIT STATUS blurb now that we can; ok guenther sobrado millert
2014-05-25	- it seems our notes that posix require rm to behave like rmdir for directories	Jason McIntyre
	is plain wrong: remove it. the note in posix spec is misleading, to be fair - most of the rest of notes in STANDARDS are really historical notes (admittedly caused by the standards i guess) so move them into HISTORY. debatable, yes, but leaves STANDARDS consistent. HISTORY is a hodge podge anyway, by neccesity. - while there, fix a sentence splice in BUGS sobrado agrees with the removal of the rmdir note; millert oks the diff
2014-05-24	Delete pointless casts: free() return value to (void), free()'s	Philip Guenther
	argument to (char *), and malloc/calloc's return value to the type of the variable it's being assigned to. Convert the one calloc() where the zeroing isn't needed to a reallocarray(). ok millert@
2014-05-24	Missed #include <signal.h>	Philip Guenther

2014-05-24	Make signal setup clearer via helper function, eliminating a gap in	Philip Guenther
	ignoring signals when they were already ignored ok millert@
2014-05-24	delete restricted ed. no one will ever use this. never documented or enabled.	Daniel Dickman
	ok millert@
2014-05-23	Archives written with a non-standard blocksize currently show up as	Philip Guenther
	truncated reads. Until better detection of that case can is implemented, back out that part of rev 1.45 problem noted by sthen@
2014-05-23	Make the signal handler safe: block signals when updating data-structures	Philip Guenther
	that are walked by routines called from the signal handler and use dprintf() instead fprintf() in ar_close(). ok millert@
2014-05-23	Use errc/warnc to simplify code.	Philip Guenther
	ok jca@ krw@
2014-05-21	Use errc/warnc to simplify code.	Philip Guenther
	ok jca@ krw@
2014-05-21	document which keywords are posix conformant;	Jason McIntyre
	ok guenther sobrado
2014-05-21	Exit with non-zero status if a read is truncated, or if a compression	Philip Guenther
	program was used but it didn't exit successfully. Original diff by mpf@
2014-05-20	As suggested by guenther@ and millert@, replace seek+[read\|write] with	Kenneth R Westerback
	p[read\|write]. Makes the code much clearer by eliminating extra error checking and verbiage. No intentional functional change. Tweaks by and ok guenther@
2014-05-18	use reallocarray	Marc Espie
	okay miod@, millert@
2014-05-17	better description for -A and -a, with help from millert;	Jason McIntyre
	ok millert sobrado guenther
2014-05-16	more STANDARDS fleshing; ok sobrado millert	Jason McIntyre

2014-05-16	Make the df calculations of available space the same as that done in	Kenneth R Westerback
	ffs_statfs(). In 1998, with /usr/src/sys/ufs/ffs/ffs_vfsops.c r1.16, mickey@ improved the calculation of available blocks in ffs_statfs(). Yesterday guenther@ noticed that this fix had not been applied to the same calculation being done by df(8) when run against raw devices. Which meant different values were displayed depending on which device/file system you gave to df. ok otto@
2014-05-15	more for STANDARDS: unsupported lists (noted by guenther) and env vars;	Jason McIntyre

2014-05-15	flesh out STANDARDS; ok guenther	Jason McIntyre

2014-05-07	use reallocarray. ok guenther	Ted Unangst

2014-05-07	Remove the undocumented and ineffective -F option (force hard link	Ingo Schwarze
	to directory). This doesn't reduce functionality but merely results in a better error message when trying to use the option, and it simplifies the code. The -f option first appeared in AT&T Version 7 UNIX (1979), and Keith Bostic renamed it to -F for 4.3BSD-Reno in 1990 because it conflicted with System V and POSIX. Meanwhile, NetBSD, FreeBSD, and DragonFly removed it, too. From Tristan Le Guern <tleguern at bouledef dot eu>. OK guenther@ krw@
2014-05-07	convert calloc to use reallocarray instead. ok millert	Ted Unangst

2014-05-06	reallocarray	Ted Unangst

2014-05-06	spell NULL as NULL	Ted Unangst

2014-04-24	calloc is better. from Peter Malone	Ted Unangst

2014-04-24	calloc is better. from Peter Malone	Ted Unangst

2014-04-20	Remove more commentary about recently removed des support	Theo de Raadt
	from Daniel Dickman
2014-04-18	tweak;	Jason McIntyre

2014-04-15	note some unsupported features, and clean up some macros - in particular,	Jason McIntyre
	whack a dose of unneccessary Em; ok millert
2014-04-15	flesh out STANDARDS; ok millert sobrado	Jason McIntyre

2014-04-14	more x removal;	Jason McIntyre