]> dgit.raspbian.org Git - git-annex.git/log
git-annex.git
10 years agodisplay drop safety proofs in debug mode
Joey Hess [Fri, 9 Oct 2015 17:47:19 +0000 (13:47 -0400)]
display drop safety proofs in debug mode

10 years agofix lockKey to run callback in original Annex monad, not local remote's
Joey Hess [Fri, 9 Oct 2015 17:35:28 +0000 (13:35 -0400)]
fix lockKey to run callback in original Annex monad, not local remote's

10 years agocontent locking during drop working for local git remotes
Joey Hess [Fri, 9 Oct 2015 17:07:03 +0000 (13:07 -0400)]
content locking during drop working for local git remotes

Only ssh remotes lack locking now

10 years agofinish and use lockContent interface
Joey Hess [Fri, 9 Oct 2015 16:36:04 +0000 (12:36 -0400)]
finish and use lockContent interface

10 years agoimprove drop proof code
Joey Hess [Fri, 9 Oct 2015 15:09:46 +0000 (11:09 -0400)]
improve drop proof code

10 years agorefactor
Joey Hess [Fri, 9 Oct 2015 14:30:22 +0000 (10:30 -0400)]
refactor

10 years agoTrustedCopy is good enough to allow dropping
Joey Hess [Thu, 8 Oct 2015 22:32:31 +0000 (18:32 -0400)]
TrustedCopy is good enough to allow dropping

By definition, a trusted repository is trusted to always have its location
tracking log accurate. Thus, it should never be in a position where content
is being dropped from it concurrently, as that would result in the location
tracking log not being accurate.

10 years agotry harder to verify until at least one VerifiedCopyLock is obtained
Joey Hess [Thu, 8 Oct 2015 22:20:36 +0000 (18:20 -0400)]
try harder to verify until at least one VerifiedCopyLock is obtained

This avoids a failure where eg, we start with RecentlyVerifiedCopies
for all remotes, and so didn't do any active verification, which is
required.

Also, dedup the list of VerifiedCopies when checking if we have enough,
in case 2 copies of a UUID slip in.

10 years agorequire 1 locked copy while dropping from local or a remote
Joey Hess [Thu, 8 Oct 2015 22:11:39 +0000 (18:11 -0400)]
require 1 locked copy while dropping from local or a remote

See doc/bugs/concurrent_drop--from_presence_checking_failures.mdwn for
discussion about why 1 locked copy is all we can require, and how this
fixes concurrent dropping bugs.

Note that, since nothing yet generates a VerifiedCopyLock yet, this commit
breaks dropping temporarily.

10 years agosupport invalidating existing VerifiedCopys
Joey Hess [Thu, 8 Oct 2015 21:58:32 +0000 (17:58 -0400)]
support invalidating existing VerifiedCopys

10 years agoadd VerifiedCopy data type
Joey Hess [Thu, 8 Oct 2015 20:55:11 +0000 (16:55 -0400)]
add VerifiedCopy data type

There should be no behavior changes in this commit, it just adds a more
expressive data type and adjusts code that had been passing around a [UUID]
or sometimes a Maybe Remote to instead use [VerifiedCopy].

Although, since some functions were taking two different [UUID] lists,
there's some potential for me to have gotten it horribly wrong.

10 years agoadd removeKey action to Remote
Joey Hess [Thu, 8 Oct 2015 19:01:38 +0000 (15:01 -0400)]
add removeKey action to Remote

Not implemented for any remotes yet; probably the git remote is the only
one that will ever implement it.

10 years agounused import
Joey Hess [Thu, 8 Oct 2015 18:59:34 +0000 (14:59 -0400)]
unused import

10 years agoI think this comment is stale/confusing; remove
Joey Hess [Thu, 8 Oct 2015 18:51:44 +0000 (14:51 -0400)]
I think this comment is stale/confusing; remove

10 years agogit-annex-shell: Added lockcontent command, to prevent dropping of key's content.
Joey Hess [Thu, 8 Oct 2015 18:47:46 +0000 (14:47 -0400)]
git-annex-shell: Added lockcontent command, to prevent dropping of key's content.

10 years agoadd lockContentShared
Joey Hess [Thu, 8 Oct 2015 18:27:37 +0000 (14:27 -0400)]
add lockContentShared

Also, rename lockContent to lockContentExclusive

inAnnexSafe should perhaps be eliminated, and instead use
`lockContentShared inAnnex`. However, I'm waiting on that, as there are
only 2 call sites for inAnnexSafe and it's fiddly.

10 years agoexport FileMode type
Joey Hess [Thu, 8 Oct 2015 18:26:21 +0000 (14:26 -0400)]
export FileMode type

10 years agoadd tryLockShared
Joey Hess [Thu, 8 Oct 2015 17:40:23 +0000 (13:40 -0400)]
add tryLockShared

10 years agoopen lock file ReadOnly when taking shared lock
Joey Hess [Thu, 8 Oct 2015 17:31:55 +0000 (13:31 -0400)]
open lock file ReadOnly when taking shared lock

It's only necessary to open a file for write when taking an exclusive lock.

10 years agogetting proof-y
Joey Hess [Wed, 7 Oct 2015 19:52:56 +0000 (15:52 -0400)]
getting proof-y

10 years agoDebian: Remove dependency on transformers library, as it is now included in ghc.
Joey Hess [Wed, 7 Oct 2015 16:26:48 +0000 (12:26 -0400)]
Debian: Remove dependency on transformers library, as it is now included in ghc.

Backporters will need to revert this.

10 years agotitles
Joey Hess [Wed, 7 Oct 2015 15:28:07 +0000 (11:28 -0400)]
titles

10 years agoindents
Joey Hess [Wed, 7 Oct 2015 15:25:55 +0000 (11:25 -0400)]
indents

10 years agoalternative solution
Joey Hess [Wed, 7 Oct 2015 15:23:27 +0000 (11:23 -0400)]
alternative solution

10 years agoanalysis of some topologies, and idea to generate moves
Joey Hess [Wed, 7 Oct 2015 15:03:15 +0000 (11:03 -0400)]
analysis of some topologies, and idea to generate moves

10 years agoDebian: Add torrent library to build-depends as it's packaged now, and stop recommend...
Joey Hess [Wed, 7 Oct 2015 13:08:37 +0000 (09:08 -0400)]
Debian: Add torrent library to build-depends as it's packaged now, and stop recommending bittornado | bittorrent.

10 years agobe verbose to debug problem with aws
Joey Hess [Tue, 6 Oct 2015 23:56:25 +0000 (19:56 -0400)]
be verbose to debug problem with aws

10 years agoadd test cases for copy --auto
Joey Hess [Tue, 6 Oct 2015 21:35:29 +0000 (17:35 -0400)]
add test cases for copy --auto

(Only --from for now, should add --to sometime but I'm on low sleep so this
is the mininal test to catch the reversion just fixed, bad me for doing it
backwards and incompletely.)

10 years agocopy --auto was checking the wrong repo's preferred content. (--from was checking...
Joey Hess [Tue, 6 Oct 2015 21:29:44 +0000 (17:29 -0400)]
copy --auto was checking the wrong repo's preferred content. (--from was checking what --to should, and vice-versa.) Fixed this bug, which was introduced in version 5.20150727.

10 years agooops, didn't mean to commit this debug
Joey Hess [Tue, 6 Oct 2015 21:28:20 +0000 (17:28 -0400)]
oops, didn't mean to commit this debug

10 years agoanalysis
Joey Hess [Tue, 6 Oct 2015 21:11:52 +0000 (17:11 -0400)]
analysis

10 years agocomment
Joey Hess [Tue, 6 Oct 2015 21:05:38 +0000 (17:05 -0400)]
comment

10 years agoforce installing older warp; current fails to build
Joey Hess [Sun, 4 Oct 2015 18:51:23 +0000 (14:51 -0400)]
force installing older warp; current fails to build

10 years agosed out the filename from sha1sum output
Joey Hess [Tue, 6 Oct 2015 20:53:24 +0000 (16:53 -0400)]
sed out the filename from sha1sum output

10 years agohopefuly fix sha1sum calling; also refactor and check downloaded program's sha before...
Joey Hess [Tue, 6 Oct 2015 20:44:12 +0000 (16:44 -0400)]
hopefuly fix sha1sum calling; also refactor and check downloaded program's sha before using it

10 years agoconfirmed
Joey Hess [Tue, 6 Oct 2015 20:33:44 +0000 (16:33 -0400)]
confirmed

10 years agoMerge branch 'master' of ssh://git-annex.branchable.com
Joey Hess [Tue, 6 Oct 2015 20:29:14 +0000 (16:29 -0400)]
Merge branch 'master' of ssh://git-annex.branchable.com

10 years agoresponse
Joey Hess [Tue, 6 Oct 2015 20:23:04 +0000 (16:23 -0400)]
response

10 years agoAdd some precision about the versions tested and why it is a bug in mind opinion
konubinix [Tue, 6 Oct 2015 20:15:51 +0000 (20:15 +0000)]
Add some precision about the versions tested and why it is a bug in mind opinion

10 years agoresponse
Joey Hess [Tue, 6 Oct 2015 20:08:32 +0000 (16:08 -0400)]
response

10 years agocomment
Joey Hess [Tue, 6 Oct 2015 20:06:39 +0000 (16:06 -0400)]
comment

10 years ago(no commit message)
konubinix [Tue, 6 Oct 2015 20:04:23 +0000 (20:04 +0000)]

10 years agoshow git-annex init w/o description for laptop, where default would suffice. And...
Joey Hess [Tue, 6 Oct 2015 19:59:43 +0000 (15:59 -0400)]
show git-annex init w/o description for laptop, where default would suffice. And for init of usb drive, be clear that it's a description, not a name.

10 years agoresponse
Joey Hess [Tue, 6 Oct 2015 19:42:21 +0000 (15:42 -0400)]
response

10 years agoavoid using print action, which is reserved for debugging
Joey Hess [Tue, 6 Oct 2015 19:26:42 +0000 (15:26 -0400)]
avoid using print action, which is reserved for debugging

10 years agocomment
Joey Hess [Tue, 6 Oct 2015 19:23:04 +0000 (15:23 -0400)]
comment

10 years agohairy problem
Joey Hess [Tue, 6 Oct 2015 18:22:51 +0000 (14:22 -0400)]
hairy problem

10 years agocomment
Joey Hess [Tue, 6 Oct 2015 17:30:58 +0000 (13:30 -0400)]
comment

10 years ago(no commit message)
jec@8edf298ac8f8152bffcd4fdc1a310949d3547fe4 [Tue, 6 Oct 2015 01:37:55 +0000 (01:37 +0000)]

10 years ago(no commit message)
jec@8edf298ac8f8152bffcd4fdc1a310949d3547fe4 [Tue, 6 Oct 2015 01:34:12 +0000 (01:34 +0000)]

10 years agoAdded a comment: account information was emailed
https://me.yahoo.com/a/EbvxpTI_xP9Aod7Mg4cwGhgjrCrdM5s-#7c0f4 [Mon, 5 Oct 2015 23:13:29 +0000 (23:13 +0000)]
Added a comment: account information was emailed

10 years ago(no commit message)
nbren12@2093afead64fc71676a3456db655dc8b1bb15abb [Mon, 5 Oct 2015 23:13:13 +0000 (23:13 +0000)]

10 years agoAdded a comment
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Mon, 5 Oct 2015 19:07:51 +0000 (19:07 +0000)]
Added a comment

10 years agoAdded a comment
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Mon, 5 Oct 2015 18:53:20 +0000 (18:53 +0000)]
Added a comment

10 years ago(no commit message)
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Mon, 5 Oct 2015 18:32:53 +0000 (18:32 +0000)]

10 years ago(no commit message)
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Mon, 5 Oct 2015 18:31:25 +0000 (18:31 +0000)]

10 years agono www doesn't seem to resolve here or on google's public NSs
anarcat [Mon, 5 Oct 2015 17:57:39 +0000 (17:57 +0000)]
no www doesn't seem to resolve here or on google's public NSs

10 years agoAdded a comment: statfs64
jirib@503223f0610c6c66f4e6dc738a5a0b2648c290b1 [Mon, 5 Oct 2015 09:58:10 +0000 (09:58 +0000)]
Added a comment: statfs64

10 years agoAdded a comment
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Mon, 5 Oct 2015 06:28:17 +0000 (06:28 +0000)]
Added a comment

10 years agoAdded a comment
jephte.clain@f0da2444b164953b596ac5e3e38da30b8c04e5e3 [Mon, 5 Oct 2015 05:48:38 +0000 (05:48 +0000)]
Added a comment

10 years agoMinor grammar fix in my comment
Øyvind A. Holm [Sun, 4 Oct 2015 23:33:12 +0000 (01:33 +0200)]
Minor grammar fix in my comment

10 years agoAdded a comment: Yay, man!
sunny256 [Sun, 4 Oct 2015 23:28:43 +0000 (23:28 +0000)]
Added a comment: Yay, man!

10 years agodevblog
Joey Hess [Sun, 4 Oct 2015 20:12:41 +0000 (16:12 -0400)]
devblog

10 years agoMerge branch 'winprocfix'
Joey Hess [Sun, 4 Oct 2015 19:46:25 +0000 (15:46 -0400)]
Merge branch 'winprocfix'

10 years agoresponse
Joey Hess [Sun, 4 Oct 2015 19:34:30 +0000 (15:34 -0400)]
response

10 years agoresponse
Joey Hess [Sun, 4 Oct 2015 19:32:54 +0000 (15:32 -0400)]
response

10 years agocomment
Joey Hess [Sun, 4 Oct 2015 19:28:12 +0000 (15:28 -0400)]
comment

10 years agoreview
Joey Hess [Sun, 4 Oct 2015 19:26:20 +0000 (15:26 -0400)]
review

10 years agoremove workaround
Joey Hess [Sun, 4 Oct 2015 19:04:52 +0000 (15:04 -0400)]
remove workaround

10 years agoMerge branch 'master' of ssh://git-annex.branchable.com
Joey Hess [Sun, 4 Oct 2015 18:56:23 +0000 (14:56 -0400)]
Merge branch 'master' of ssh://git-annex.branchable.com

10 years agoforce installing older warp; current fails to build
Joey Hess [Sun, 4 Oct 2015 18:51:23 +0000 (14:51 -0400)]
force installing older warp; current fails to build

10 years agoAdded a comment
dah [Sun, 4 Oct 2015 18:25:29 +0000 (18:25 +0000)]
Added a comment

10 years agoAdded a comment
https://www.google.com/accounts/o8/id?id=AItOawnmF_9CAtfqdZkC4e-_dCX-rK5bqh4RWkw [Sun, 4 Oct 2015 17:58:48 +0000 (17:58 +0000)]
Added a comment

10 years agoAdded a comment: no statfs64 on OpenBSD
jirib@503223f0610c6c66f4e6dc738a5a0b2648c290b1 [Sun, 4 Oct 2015 13:47:30 +0000 (13:47 +0000)]
Added a comment: no statfs64 on OpenBSD

10 years ago(no commit message)
jirib@503223f0610c6c66f4e6dc738a5a0b2648c290b1 [Sun, 4 Oct 2015 13:37:53 +0000 (13:37 +0000)]

10 years ago(no commit message)
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Sun, 4 Oct 2015 11:09:51 +0000 (11:09 +0000)]

10 years ago(no commit message)
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Sun, 4 Oct 2015 11:09:02 +0000 (11:09 +0000)]

10 years ago(no commit message)
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Sun, 4 Oct 2015 11:08:07 +0000 (11:08 +0000)]

10 years agoAdded a comment
https://www.google.com/accounts/o8/id?id=AItOawnmF_9CAtfqdZkC4e-_dCX-rK5bqh4RWkw [Sun, 4 Oct 2015 07:49:02 +0000 (07:49 +0000)]
Added a comment

10 years agoAdded a comment
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Sat, 3 Oct 2015 21:29:29 +0000 (21:29 +0000)]
Added a comment

10 years agoAdded a comment
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Sat, 3 Oct 2015 21:27:46 +0000 (21:27 +0000)]
Added a comment

10 years agoAdded a comment
encryptio [Sat, 3 Oct 2015 21:09:31 +0000 (21:09 +0000)]
Added a comment

10 years agoAdded a comment: Need advice on using gitlab with the assistant
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Sat, 3 Oct 2015 21:03:46 +0000 (21:03 +0000)]
Added a comment: Need advice on using gitlab with the assistant

10 years ago(no commit message)
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Sat, 3 Oct 2015 20:11:53 +0000 (20:11 +0000)]

10 years agoAdded a comment
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Sat, 3 Oct 2015 19:18:24 +0000 (19:18 +0000)]
Added a comment

10 years agoAdded a comment
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Sat, 3 Oct 2015 19:03:18 +0000 (19:03 +0000)]
Added a comment

10 years ago(no commit message)
ovidiu@66ace8a8d99ce938b0538ffa0f26d30db02a9626 [Sat, 3 Oct 2015 18:36:49 +0000 (18:36 +0000)]

10 years agoAdded a comment
jephte.clain@f0da2444b164953b596ac5e3e38da30b8c04e5e3 [Sat, 3 Oct 2015 09:11:02 +0000 (09:11 +0000)]
Added a comment

10 years agoPorted disk free space checking code to work on Solaris.
Joey Hess [Fri, 2 Oct 2015 20:29:29 +0000 (16:29 -0400)]
Ported disk free space checking code to work on Solaris.

On Solaris, using f_bsize provided a value that is apparently much larger
than the real block size. The solaris docs for statvfs say
f_bsize is the "preferred" file system block size, and I guess the
filesystem prefers larger blocks, but uses smaller ones or something.
The docs also say that f_frsize is the "fundamental" block size.

Switched to using f_frsize on Linux and kFreeBSD too, since I guess
f_bsize could in theory vary the same way there too. Assuming that Solaris
is not violating the posix spec, I guess the linux man page for statvfs
is not as well written and I misunderstood it.

10 years agodeal with more backward-compatible breaking renamings in conduit
Joey Hess [Fri, 2 Oct 2015 19:18:54 +0000 (15:18 -0400)]
deal with more backward-compatible breaking renamings in conduit

This is the kind of annoying thing that makes me not want to use a library.
conduitManagerSettings was a perfectly fine name and could have been kept
forever.

10 years agoother 80% of avoding verification when hard linking to objects in shared repo
Joey Hess [Fri, 2 Oct 2015 17:56:42 +0000 (13:56 -0400)]
other 80% of avoding verification when hard linking to objects in shared repo

In c6632ee5c8e66c26ef18317f56ae02bae1e7e280, it actually only handled
uploading objects to a shared repository. To avoid verification when
downloading objects from a shared repository, was a lot harder.

On the plus side, if the process of downloading a file from a remote
is able to verify its content on the side, the remote can indicate this
now, and avoid the extra post-download verification.

As of yet, I don't have any remotes (except Git) using this ability.
Some more work would be needed to support it in special remotes.

It would make sense for tahoe to implicitly verify things downloaded from it;
as long as you trust your tahoe server (which typically runs locally),
there's cryptographic integrity. OTOH, despite bup being based on shas,
a bup repo under an attacker's control could have the git ref used for an
object changed, and so a bup repo shouldn't implicitly verify. Indeed,
tahoe seems unique in being trustworthy enough to implicitly verify.

10 years agodisabling verification also disables size verification
Joey Hess [Fri, 2 Oct 2015 16:38:02 +0000 (12:38 -0400)]
disabling verification also disables size verification

It's not expensive to do size verification, but let's be consistent and
turn it off too.

10 years agoavoid verification when hard linking to objects in shared repository
Joey Hess [Fri, 2 Oct 2015 16:36:03 +0000 (12:36 -0400)]
avoid verification when hard linking to objects in shared repository

Such a repository is implicitly trusted, so there's no point.

10 years agoresponse
Joey Hess [Fri, 2 Oct 2015 16:19:02 +0000 (12:19 -0400)]
response

10 years agoAllow building with S3 disabled again.
Joey Hess [Fri, 2 Oct 2015 16:12:02 +0000 (12:12 -0400)]
Allow building with S3 disabled again.

10 years agoAdded a comment: not easy to access history
Agenta [Fri, 2 Oct 2015 13:16:30 +0000 (13:16 +0000)]
Added a comment: not easy to access history

10 years agoAdded a comment: iiuc assistant depends on s3
jirib@503223f0610c6c66f4e6dc738a5a0b2648c290b1 [Fri, 2 Oct 2015 11:11:13 +0000 (11:11 +0000)]
Added a comment: iiuc assistant depends on s3

10 years ago(no commit message)
jirib@503223f0610c6c66f4e6dc738a5a0b2648c290b1 [Fri, 2 Oct 2015 10:20:48 +0000 (10:20 +0000)]

10 years ago(no commit message)
https://www.google.com/accounts/o8/id?id=AItOawnmF_9CAtfqdZkC4e-_dCX-rK5bqh4RWkw [Fri, 2 Oct 2015 09:25:02 +0000 (09:25 +0000)]

10 years agoAdded a comment
spwhitton [Fri, 2 Oct 2015 04:15:48 +0000 (04:15 +0000)]
Added a comment