Issue 1735: check => Hash mismatch(es!) Bad index. (2.3.98.2) - Darcs bug tracker

Title	check => Hash mismatch(es!) Bad index. (2.3.98.2)
Priority	bug	Status	resolved
Milestone	2.4.x	Resolved in
Superseder		Nosy List	dagit, darcs-devel, dmitry.kurochkin, kowey, mornfall, tux_rocker
Assigned To	mornfall	Topics	Hashed, Regression

Created on 2010-02-02.17:17:26 by kowey, last changed 2010-06-15.21:31:11 by admin.

Files
File name	Uploaded	Type	Edit	Remove
unnamed	dagit, 2010-02-02.17:56:32	text/html

Messages
msg9931 (view)	Author: kowey	Date: 2010-02-02.17:17:18
This is from issue1731. Both Jason and I are able to reproduce it on MacOS X. $ curl -O http://repos.mornfall.net/darcs/benchmark-repos/ghc-hashed.tgz $ mkdir ghc $ cd ghc $ tar xf ../ghc-hashed.tgz $ darcs check Reinier: I suppose you'll be wanting to hit the brakes on the Darcs 2.4 release? The repository is consistent! Hash mismatch(es)! bindisttest/a/b index: 0000000000000000000000000000000000000000000000000000000000000000 working: 5847c8b50a3191775d91cbaf8029fd1a75725517b0c70dde5bff759fc67ec1f8 bindisttest/a index: 0000000000000000000000000000000000000000000000000000000000000000 working: e9f80a49790bc926a3de5542274ae979753a3aedaef1b9fab0fcb74a21bc7c60 bindisttest index: 3c99d780272456ac9e4ab10056ee3e145700260cefee61921bf89017a15a6535 working: 5a72b14295811905378a648e98da91922bbc69f21c1a209357834fc2e92da3f8 Bad index.
msg9932 (view)	Author: kowey	Date: 2010-02-02.17:26:54
I don't know if this helps any, but if I remove the _darcs/index file and re-run darcs show index, I still get all zeros in the relevant entry. This is despite having the 0.4.5 version of hashed storage. I'll note that this seems rather similar to issue1677
msg9933 (view)	Author: kowey	Date: 2010-02-02.17:52:47
I can reproduce this on Linux with darcs HEAD as well.
msg9934 (view)	Author: dagit	Date: 2010-02-02.17:56:32
On Tue, Feb 2, 2010 at 9:52 AM, Eric Kow <bugs@darcs.net> wrote: > > Eric Kow <kowey@darcs.net> added the comment: > > I can reproduce this on Linux with darcs HEAD as well. > Interesting. Does it still happen if you temporarily remove your global cache and try again? Attachments unnamed
msg9935 (view)	Author: kowey	Date: 2010-02-02.18:44:28
On Tue, Feb 02, 2010 at 17:56:33 +0000, Jason Dagit wrote: > Interesting. Does it still happen if you temporarily remove your global > cache and try again? Yes. But I think Petr has diagnosed the problem. More on this later. -- Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow> PGP Key ID: 08AC04F9
msg10145 (view)	Author: kowey	Date: 2010-03-09.20:18:22
This one got a little lost. I seem to recall we'd discussed this on IRC. I think we just need Petr to confirm that this one was fixed before the 2.4 release
msg10311 (view)	Author: tux_rocker	Date: 2010-03-20.11:24:11
I cannot reproduce this with the current development darcs. mornfall says it's been fixed in a hashed-storage release, but the bug kept creeping up because the index in the tarball was still bad. I'm closing this; if anyone experiences it, he is welcome to re-open it of course.

History
Date	User	Action	Args
2010-02-02 17:17:26	kowey	create
2010-02-02 17:26:56	kowey	set	messages: + msg9932
2010-02-02 17:52:49	kowey	set	messages: + msg9933
2010-02-02 17:56:33	dagit	set	files: + unnamed messages: + msg9934
2010-02-02 18:44:30	kowey	set	messages: + msg9935
2010-03-09 20:18:24	kowey	set	status: unknown -> waiting-for assignedto: mornfall messages: + msg10145
2010-03-20 11:24:14	tux_rocker	set	status: waiting-for -> resolved messages: + msg10311
2010-06-15 21:31:08	admin	set	milestone: 2.4.x
2010-06-15 21:31:11	admin	set	topic: - Target-2.4