src - OpenBSD base system

Age	Commit message (Collapse)	Author
2013-07-08	Swap the program name and it's exit code so the output makes sense again.	Jasper Lievisse Adriaanse
	ok espie@
2013-05-29	I had misgivings about that one, I was right. Missed an obvious testcases,	Marc Espie
	and my crash bunnies ARE a bunch of unreliable nitwits. thanks kili@
2013-05-29	allow installpath to use +=, requested by quite a few people,	Marc Espie
	but did require testing.
2013-05-26	add B: to the list of options actually handled.	Marc Espie

2013-05-13	specialfiles ARE file while destated, so they set lastfile, which is	Marc Espie
	useful for size and digest. BUT samples should never alias specialfiles, so error out right away.
2013-04-16	pass state correctly so we can do error messages.	Marc Espie

2013-02-18	re-apply:	Okan Demirmen
	use localbase for pkg-readmes ok espie@
2013-02-18	revert previous; wrong time.	Okan Demirmen

2013-02-18	use localbase for pkg-readmes	Okan Demirmen
	ok espie@
2013-01-18	much faster -E: try a sneak peek at the 'raw' contents file.	Marc Espie
	if it doesn't contain the last part of the filename, then we don't even need to parse the plist... idea prompted by a chance remark from Vadim.
2013-01-18	move lock down: we don't really need to lock to match the directory	Marc Espie

2013-01-07	only protest midly if we're in exclude mode, so that you can exclude stuff	Marc Espie
	that does not exist, and that you will still see typos.
2012-12-31	simple optimization: if we're already erroring out, there's totally no	Marc Espie
	need to compute those pesky checksums.
2012-12-28	base on landry's remarks, stuff like .libs* and partial* shouldn't ever	Marc Espie
	have manual annotation markers. (the read part is there for legacy, as some stuff is already written back that way. Eventually, no .libs* will have @option manual-install left).
2012-11-06	sort options; ok espie	Jason McIntyre

2012-11-06	work-around python compile issue: don't tie python files for now, so	Marc Espie
	that timestamps don't trigger recompile.
2012-11-06	do interactivity like for the others.	Marc Espie
	force pkglocatedb-- when auto-installing (reported by aja)
2012-10-13	In update verbose mode, when skipping an update candidate due to an	Jeremy Evans
	unmatching pkgpath, if no valid update candidates are found, print that we are skipping the update candidate and all pkgpaths for both the currently installed package and the skipped update candidate. This is designed to help diagnosing issues when ports move locations in the tree, but the necessary @pkgpath entries are not correctly added to the PLIST. Much help and OK espie@
2012-09-06	shorter display for trivial updatesets,	Marc Espie
	foo-0.0->foo-1.0 becomes foo-0.0->1.0
2012-09-01	show more precisely what the problem is.	Marc Espie

2012-07-13	temporary option until I come up with a nicer way: if -DNOINSTALLPATH	Marc Espie
	don't even look in pkg.conf for the installpath.
2012-07-12	minor fix	Marc Espie

2012-07-06	have -i be automatically on for ttys.	Marc Espie
	use -i/-I to force interactive/non interactive mode (-I for not running scripts has been dead for years)
2012-06-14	tweak pkg_info -S to separate pkgnames with @ from libraries with nothing.	Marc Espie
	this goes together with a bsd.port.mk change !!! fixes some ambiguities in out-of-date, for instance. if you're parsing pkg_info -S output manually, this WILL change the way things work
2012-06-11	move PkgConfig.pm where it belongs	Marc Espie
	as requested by matthieu@, ok jasper@, miod@, matthieu@
2012-06-08	activate new pkg_create magic.	Marc Espie
	Note that, in most cases, PFRAG.shared are probably no longer needed...
2012-06-08	silent rename of @comment subdir= to @comment pkgpath=	Marc Espie
	this makes more sense
2012-05-24	put possible fragments in there.	Marc Espie
	fix buglet in parsing libraries
2012-05-15	supplementary magic code, not active yet	Marc Espie

2012-05-15	less hackish test for LIB*VERSION: after building the element,	Marc Espie
	redo a check based on the non-subst version, so we can: - actually parse the library name - check we depend upon the right variable
2012-05-15	allow elements to remove themselves from PackingList, adjust visit	Marc Espie
	accordingly. Make sure obsolete stuff ->add returns undef, unconfuses PkgCreate.
2012-05-07	new method to convert other file names to fullname.	Marc Espie

2012-05-07	well pkg_create itself doesn't care, but a custom ->annotate will get	Marc Espie
	very confused
2012-05-07	in the odd case we don't live under /usr/local, make sure we stop eventually	Marc Espie

2012-05-05	if the lib directory is not localbase->lib, we add directory entries to	Marc Espie
	the .libs* stuff.
2012-05-05	un-specialcase shallow/deep copy: create a normal new plist.	Marc Espie
	(so don't copy FCONTENTS since it's automatically created)
2012-05-01	make fragment handling yet more customizable	Marc Espie

2012-05-01	properly pass extra parameters from State->new to init.	Marc Espie

2012-04-30	allow a subclass to override that specific error	Marc Espie

2012-04-30	make the reading of packing-lists amenable to further annotations, such as	Marc Espie
	those that will be required for update-plist
2012-04-30	move code around a bit more	Marc Espie

2012-04-30	more refactor	Marc Espie

2012-04-30	just a bit of refactoring, make sure option handling is done in state,	Marc Espie
	as should be, so we can reuse it without needing to duplicate all of pkg_create code.
2012-04-28	explain why the base pseudo-handle is a dirty hack and why it's less	Marc Espie
	confusing than the clean approach
2012-04-28	File::Path::rmtree is very stupid and spews out warning messages if there's	Marc Espie
	no cwd. Well, if there's no cwd, we don't care about getting back there, so wrap rmtree to just move to / before calling the real rmtree.
2012-04-28	zap blank spaces	Marc Espie

2012-04-28	work-around the fact that BaseSystem updates have specific dependency	Marc Espie
	handling
2012-04-28	cleaner code: only wrap the call to quirks itself inside an eval.	Marc Espie
	anything else that could fail shouldn't be inside the eval. create an object instead of saying "hey we're cool, let's let other stuff break horribly".
2012-04-28	add a "fake" handle class for base system updates and ForwardDependencies	Marc Espie

2012-04-27	use our light-weight OpenBSD::MkTemp, which is safer, more accurate,	Marc Espie
	and doesn't pull a shitload of other modules with it.