Issue 1934: (in)sufficient garbage collection of unused hashed files? - Darcs bug tracker

Title	(in)sufficient garbage collection of unused hashed files?
Priority	bug	Status	wont-fix
Milestone		Resolved in
Superseder		Nosy List	dmitry.kurochkin, kowey, mornfall
Assigned To		Topics	Hashed, Performance

Created on 2010-08-26.19:15:51 by kowey, last changed 2016-04-03.21:41:47 by gh.

Messages
msg12321 (view)	Author: kowey	Date: 2010-08-26.19:15:49
I don't know if it's actually insufficient. Petr says that cleanRepository is not called enough and that it should be called on commit time ideally (darcs record?) [need-action: One discussion point: is this something we really want to happen automatically? Would that preclude some sort of future reflog feature? Or am I in YAGNI violation?] See also issue305
msg12323 (view)	Author: kowey	Date: 2010-08-26.19:26:58
(that said, having darcs-style patches and the whole patch theory thing generally means we can just recover stuff from the patches, right, so why reflog?)
msg12953 (view)	Author: gh	Date: 2010-11-09.12:12:31
See also * pristine files accumulating: http://wiki.darcs.net/Using/GrowingPristineProblem * inventory files accumulating: http://wiki.darcs.net/Using/GrowingInventoriesProblem As explained in the first page, I believe this has to happen in order to avoid race conditions between pushing/getting, see http://bugs.darcs.net/issue687. Maybe we should look at how git does garbage collection in order to see how they handle these race conditions.
msg19133 (view)	Author: gh	Date: 2016-04-03.21:41:45
Well that was sort of fixed in GSoC 2013 (currently `darcs optimize clean` does the GC). But the consensus is that it should not happen automatically.

History
Date	User	Action	Args
2010-08-26 19:15:51	kowey	create
2010-08-26 19:26:59	kowey	set	messages: + msg12323
2010-11-09 12:12:32	gh	set	messages: + msg12953
2016-04-03 21:41:47	gh	set	status: needs-reproduction -> wont-fix messages: + msg19133