git-annex.git
9 years agoAdded a comment
CandyAngel [Thu, 16 Feb 2017 14:10:56 +0000 (14:10 +0000)]
Added a comment

9 years ago(no commit message)
JohnKozak [Thu, 16 Feb 2017 13:56:30 +0000 (13:56 +0000)]

9 years agoAdded a comment
andy [Thu, 16 Feb 2017 06:34:50 +0000 (06:34 +0000)]
Added a comment

9 years ago(no commit message)
yarikoptic [Thu, 16 Feb 2017 03:49:46 +0000 (03:49 +0000)]

9 years agodevblog
Joey Hess [Wed, 15 Feb 2017 20:44:10 +0000 (16:44 -0400)]
devblog

9 years agosync: Improve integration with receive.denyCurrentBranch=updateInstead
Joey Hess [Wed, 15 Feb 2017 20:13:30 +0000 (16:13 -0400)]
sync: Improve integration with receive.denyCurrentBranch=updateInstead

By displaying error messages from the remote then it fails to update
its checked out branch.

Error messages in the default receive.denyCurrentBranch are still
suppressed, which matches user expectations.

This commit was sponsored by Nick Daly on Patreon.

9 years agosimpler more generic processTranscript'
Joey Hess [Wed, 15 Feb 2017 20:00:59 +0000 (16:00 -0400)]
simpler more generic processTranscript'

This allows using functions that generate CreateProcess and passing the
result to processTranscript', which is more flexible, and also simpler
than the old interface.

This commit was sponsored by Riku Voipio.

9 years agoRun ssh with -n whenever input is not being piped into it
Joey Hess [Wed, 15 Feb 2017 19:08:46 +0000 (15:08 -0400)]
Run ssh with -n whenever input is not being piped into it

... to avoid it consuming stdin that it shouldn't.

This fixes git-annex-checkpresentkey --batch remote, which didn't output
results for all keys passed into it.

Other git-annex commands that communicate with a remote over ssh may also
have been consuming stdin that they shouldn't have, which could have
impacted using them in eg, shell scripts. For example, a shell script
reading files from stdin and passing them to git annex drop would be
impacted by this bug, whenever git annex drop ran git-annex-shell
checkpresent, it would consume part/all of the stdin that the shell script
was supposed to consume.

Fixed by adding a ConsumeStdin parameter to Annex.Ssh.sshOptions, which
is used throughout git-annex to run ssh (in order for ssh connection
caching to work). Every call site was checked to see if it used
CreatePipe for stdin, and if not was marked NoConsumeStdin.

9 years agoadd the new armel-ancient build
Joey Hess [Wed, 15 Feb 2017 18:43:19 +0000 (14:43 -0400)]
add the new armel-ancient build

9 years agoMerge branch 'master' of ssh://git-annex.branchable.com
Joey Hess [Wed, 15 Feb 2017 17:56:01 +0000 (13:56 -0400)]
Merge branch 'master' of ssh://git-annex.branchable.com

9 years agoforgot to compile last commit; fix mistakes
Joey Hess [Wed, 15 Feb 2017 17:55:06 +0000 (13:55 -0400)]
forgot to compile last commit; fix mistakes

9 years agoAdded a comment
lasitus [Wed, 15 Feb 2017 16:01:39 +0000 (16:01 +0000)]
Added a comment

9 years ago(no commit message)
CandyAngel [Wed, 15 Feb 2017 11:13:25 +0000 (11:13 +0000)]

9 years agoAdded a comment
lasitus [Wed, 15 Feb 2017 02:17:25 +0000 (02:17 +0000)]
Added a comment

9 years agoremoved
lasitus [Wed, 15 Feb 2017 02:14:05 +0000 (02:14 +0000)]
removed

9 years agoAdded a comment
lasitus [Wed, 15 Feb 2017 02:13:25 +0000 (02:13 +0000)]
Added a comment

9 years agoMerge branch 'master' of ssh://git-annex.branchable.com
Joey Hess [Tue, 14 Feb 2017 21:42:25 +0000 (17:42 -0400)]
Merge branch 'master' of ssh://git-annex.branchable.com

9 years agothoughts
Joey Hess [Tue, 14 Feb 2017 21:39:37 +0000 (17:39 -0400)]
thoughts

9 years agoAdded a comment
db48x [Tue, 14 Feb 2017 21:10:13 +0000 (21:10 +0000)]
Added a comment

9 years agoAdded a comment
https://me.yahoo.com/a/80VlVB0Bx9TaNOXIj3OCQ8eimAtIOhqjUQ--#1e80e [Tue, 14 Feb 2017 20:21:41 +0000 (20:21 +0000)]
Added a comment

9 years agosync, merge: Fail when the current branch has no commits yet, instead of not merging...
Joey Hess [Tue, 14 Feb 2017 19:58:05 +0000 (15:58 -0400)]
sync, merge: Fail when the current branch has no commits yet, instead of not merging in anything from remotes and appearing to succeed.

At first I wanted to make it go ahead and merge into the newborn branch,
so made it use Git.Branch.currentUnsafe to get the current branch. But that
failed:

fatal: ambiguous argument 'refs/heads/master..refs/heads/synced/master':
unknown revision or path not in the working tree.

A whole nother code path to handle merging into newborn branches seemed
excessive, so went with displaying a warning and propigating failure
status.

This commit was sponsored by Brock Spratlen on Patreon.

9 years agoAdded a comment
lasitus [Tue, 14 Feb 2017 19:08:38 +0000 (19:08 +0000)]
Added a comment

9 years agoadd news item for git-annex 6.20170214
Joey Hess [Tue, 14 Feb 2017 18:56:57 +0000 (14:56 -0400)]
add news item for git-annex 6.20170214

9 years agoreleasing package git-annex version 6.20170214
Joey Hess [Tue, 14 Feb 2017 18:56:11 +0000 (14:56 -0400)]
releasing package git-annex version 6.20170214

9 years agocomment
Joey Hess [Tue, 14 Feb 2017 18:54:53 +0000 (14:54 -0400)]
comment

9 years agoMerge branch 'master' of ssh://git-annex.branchable.com
Joey Hess [Tue, 14 Feb 2017 18:49:23 +0000 (14:49 -0400)]
Merge branch 'master' of ssh://git-annex.branchable.com

9 years agoAdded a comment
lasitus [Tue, 14 Feb 2017 18:46:22 +0000 (18:46 +0000)]
Added a comment

9 years agoresponse
Joey Hess [Tue, 14 Feb 2017 18:21:58 +0000 (14:21 -0400)]
response

9 years ago(no commit message)
http://xgm.de/oid/ [Tue, 14 Feb 2017 18:14:24 +0000 (18:14 +0000)]

9 years agoAdded a comment
http://xgm.de/oid/ [Tue, 14 Feb 2017 18:03:27 +0000 (18:03 +0000)]
Added a comment

9 years agoMerge branch 'master' of ssh://git-annex.branchable.com
Joey Hess [Tue, 14 Feb 2017 15:45:38 +0000 (11:45 -0400)]
Merge branch 'master' of ssh://git-annex.branchable.com

9 years agoavoid OSX build failure when there are no rpaths to remove
Joey Hess [Tue, 14 Feb 2017 15:44:52 +0000 (11:44 -0400)]
avoid OSX build failure when there are no rpaths to remove

9 years agoremoved
https://me.yahoo.com/a/80VlVB0Bx9TaNOXIj3OCQ8eimAtIOhqjUQ--#1e80e [Tue, 14 Feb 2017 00:05:09 +0000 (00:05 +0000)]
removed

9 years agoAdded a comment: Retrieving initremote parameters for new user
https://me.yahoo.com/a/80VlVB0Bx9TaNOXIj3OCQ8eimAtIOhqjUQ--#1e80e [Tue, 14 Feb 2017 00:04:50 +0000 (00:04 +0000)]
Added a comment: Retrieving initremote parameters for new user

9 years agoAdded a comment: Retrieving initremote parameters for new user
https://me.yahoo.com/a/80VlVB0Bx9TaNOXIj3OCQ8eimAtIOhqjUQ--#1e80e [Tue, 14 Feb 2017 00:03:35 +0000 (00:03 +0000)]
Added a comment: Retrieving initremote parameters for new user

9 years agodevblog
Joey Hess [Mon, 13 Feb 2017 21:42:03 +0000 (17:42 -0400)]
devblog

9 years agoWork around sqlite's incorrect handling of umask when creating databases.
Joey Hess [Mon, 13 Feb 2017 21:30:28 +0000 (17:30 -0400)]
Work around sqlite's incorrect handling of umask when creating databases.

Refactored some common code into initDb.

This only deals with the problem when creating new databases. If a repo
got bad permissions into it, it's up to the user to deal with it.

This commit was sponsored by Ole-Morten Duesund on Patreon.

9 years agoresponse
Joey Hess [Mon, 13 Feb 2017 20:19:39 +0000 (16:19 -0400)]
response

9 years agocomment
Joey Hess [Mon, 13 Feb 2017 20:11:58 +0000 (16:11 -0400)]
comment

9 years agoremove recent addition of receive.denyNonFastForwards to walkthrough
Joey Hess [Mon, 13 Feb 2017 20:02:53 +0000 (16:02 -0400)]
remove recent addition of receive.denyNonFastForwards to walkthrough

That is a dangerous setting, and not needed.

I tried the walkthrough as it was without it, and syncing in the USB
remote worked ok. There was a problem merging from origin/master, since
that branch didn't exist yet, but that didn't prevent sync from working
at all.

To avoid even that problem, reordered the walkthrough, so files get
committed to the repo before the remote gets set up.

This commit was sponsored by Ole-Morten Duesund on Patreon.

9 years agofix commit id
Joey Hess [Mon, 13 Feb 2017 19:55:32 +0000 (15:55 -0400)]
fix commit id

9 years agoinclude problem PairData in error message
Joey Hess [Mon, 13 Feb 2017 19:51:13 +0000 (15:51 -0400)]
include problem PairData in error message

Users occasionally report this error firing, and I can't see why,
so include the rejected PairData in the error message.

This is safe even if it contains evil escape characters, because showing
it displays them in escaped form.

This commit was sponsored by Bruno BEAUFILS on Patreon.

9 years agocomment
Joey Hess [Mon, 13 Feb 2017 19:39:46 +0000 (15:39 -0400)]
comment

9 years agoS3: Fix check of uuid file stored in bucket, which was not working.
Joey Hess [Mon, 13 Feb 2017 19:35:24 +0000 (15:35 -0400)]
S3: Fix check of uuid file stored in bucket, which was not working.

The check was broken in two ways.. First, nowhere did it error out when
checkUUIDFile found a different UUID already in the file. Instead,
it overwrote the uuid file.

And, checkUUIDFile's implementation was for some reason always failing with
a ConnectionClosed exception. Apparently something to do with using two
different runResourceT's and a response getting GCed inbetween. I'm pretty
sure that used to work, but changed to a more obviously correct
implementation.

This commit was sponsored by Peter Hogg on Patreon.

9 years agocomment
Joey Hess [Mon, 13 Feb 2017 18:53:55 +0000 (14:53 -0400)]
comment

9 years agocomment
Joey Hess [Mon, 13 Feb 2017 18:43:31 +0000 (14:43 -0400)]
comment

9 years agoMerge branch 'master' of ssh://git-annex.branchable.com
Joey Hess [Mon, 13 Feb 2017 16:46:15 +0000 (12:46 -0400)]
Merge branch 'master' of ssh://git-annex.branchable.com

9 years agofix OSX build
Joey Hess [Mon, 13 Feb 2017 16:41:28 +0000 (12:41 -0400)]
fix OSX build

9 years ago(no commit message)
lasitus [Mon, 13 Feb 2017 03:08:14 +0000 (03:08 +0000)]

9 years agocorrect spelling mistakes
Edward Betts [Sat, 11 Feb 2017 09:38:49 +0000 (09:38 +0000)]
correct spelling mistakes

9 years ago(no commit message)
http://xgm.de/oid/ [Sun, 12 Feb 2017 16:28:03 +0000 (16:28 +0000)]

9 years agocorrect spelling mistake
edward [Sat, 11 Feb 2017 09:16:26 +0000 (09:16 +0000)]
correct spelling mistake

9 years agocorrect spelling mistake
edward [Sat, 11 Feb 2017 09:14:27 +0000 (09:14 +0000)]
correct spelling mistake

9 years agoImprove pid locking code to work on filesystems that don't support hard links.
Joey Hess [Fri, 10 Feb 2017 19:21:58 +0000 (15:21 -0400)]
Improve pid locking code to work on filesystems that don't support hard links.

Probing for hard link support in the pid locking code is redundant since
git-annex init already probes that. But, it didn't seem worth threading
that data through; the pid locking code runs at most once per git-annex
process, and only on unusual filesystems. Optimising a single hard link
and unlink isn't worth it.

This commit was sponsored by Francois Marier on Patreon.

9 years agocomment
Joey Hess [Fri, 10 Feb 2017 18:51:50 +0000 (14:51 -0400)]
comment

9 years agocomment
Joey Hess [Fri, 10 Feb 2017 18:04:55 +0000 (14:04 -0400)]
comment

9 years agoAdded git template directory to Linux standalone tarball and OSX app bundle.
Joey Hess [Fri, 10 Feb 2017 17:55:54 +0000 (13:55 -0400)]
Added git template directory to Linux standalone tarball and OSX app bundle.

Git does not provide a switch to find out where this directory is, and
while the git-init man page says it will always be in
/usr/share/git-core/templates, that's not the case on OSX with git
installed from homebrew. So, I used a hack taking the --man-path and
constructing a path from that. Works on both Debian and OSX at least.

9 years agoMerge branch 'master' of ssh://git-annex.branchable.com
Joey Hess [Fri, 10 Feb 2017 17:51:53 +0000 (13:51 -0400)]
Merge branch 'master' of ssh://git-annex.branchable.com

9 years ago(no commit message)
yarikoptic [Fri, 10 Feb 2017 02:33:49 +0000 (02:33 +0000)]

9 years agoAdded a comment
yarikoptic [Fri, 10 Feb 2017 01:38:33 +0000 (01:38 +0000)]
Added a comment

9 years agoclose
Joey Hess [Thu, 9 Feb 2017 22:59:45 +0000 (18:59 -0400)]
close

9 years agoupdate
Joey Hess [Thu, 9 Feb 2017 22:58:13 +0000 (18:58 -0400)]
update

9 years agoAdded a comment: RequestTimeoutException when uploading to glacier
victorsavu3@d059e0a74888923e70a555ce8ab75e8ba00b6eb7 [Thu, 9 Feb 2017 20:45:42 +0000 (20:45 +0000)]
Added a comment: RequestTimeoutException when uploading to glacier

9 years agocomment
Joey Hess [Thu, 9 Feb 2017 20:13:02 +0000 (16:13 -0400)]
comment

9 years agoimport --reinject-duplicates
Joey Hess [Thu, 9 Feb 2017 19:40:44 +0000 (15:40 -0400)]
import --reinject-duplicates

This is the same as running git annex reinject --known, followed by
git-annex import. The advantage to having it in one command is that it
only has to hash each file once; the two commands have to
hash the imported files a second time.

This commit was sponsored by Shane-o on Patreon.

9 years agoMerge branch 'master' of ssh://git-annex.branchable.com
Joey Hess [Thu, 9 Feb 2017 19:33:42 +0000 (15:33 -0400)]
Merge branch 'master' of ssh://git-annex.branchable.com

9 years agoMake import --deduplicate and --skip-duplicates only hash once, not twice
Joey Hess [Thu, 9 Feb 2017 19:32:22 +0000 (15:32 -0400)]
Make import --deduplicate and --skip-duplicates only hash once, not twice

import: --deduplicate and --skip-duplicates were implemented inneficiently;
they unncessarily hashed each file twice. They have been improved to only
hash once.

The new approach is to lock down (minimally) and hash files, and then
reuse that information when importing them.

This was rather tricky, especially in detecting changes to files while
they are being imported.

The output of import changed slightly. While before it silently skipped
over files with eg --skip-duplicates, now it shows each file as it starts
to act on it. Since every file is hashed first thing, it would otherwise
not be clear what file import is chewing on. (Actually, it wasn't clear
before when any of the duplicates switches were used.)

This commit was sponsored by Alexander Thompson on Patreon.

9 years agoAdded a comment
benjamin.poldrack@d09ccff6d42dd20277610b59867cf7462927b8e3 [Thu, 9 Feb 2017 09:05:34 +0000 (09:05 +0000)]
Added a comment

9 years agoNew Bug Report
dharrison@27b3fc3aa4a83cdc03f8e82ab4a8d4f6455e603d [Wed, 8 Feb 2017 01:42:30 +0000 (01:42 +0000)]
New Bug Report

9 years agoremoved
dharrison@27b3fc3aa4a83cdc03f8e82ab4a8d4f6455e603d [Wed, 8 Feb 2017 01:32:52 +0000 (01:32 +0000)]
removed

9 years agoAdded a comment: MacOSX and Debian cannot decrypt eachother's S3 credentials
dharrison@27b3fc3aa4a83cdc03f8e82ab4a8d4f6455e603d [Wed, 8 Feb 2017 01:26:09 +0000 (01:26 +0000)]
Added a comment: MacOSX and Debian cannot decrypt eachother's S3 credentials

9 years agoAdded a comment
CandyAngel [Tue, 7 Feb 2017 22:51:15 +0000 (22:51 +0000)]
Added a comment

9 years agoresponse
Joey Hess [Tue, 7 Feb 2017 22:07:47 +0000 (18:07 -0400)]
response

9 years agoimport: Changed how --deduplicate, --skip-duplicates, and --clean-duplicates determin...
Joey Hess [Tue, 7 Feb 2017 21:35:51 +0000 (17:35 -0400)]
import: Changed how --deduplicate, --skip-duplicates, and --clean-duplicates determine if a file is a duplicate

Before, only content known to be present somewhere was considered a
duplicate. Now, any content that has been annexed before will be considered
a duplicate, even if all annexed copies of the data have been lost.

Note that --clean-duplicates and --deduplicate still check numcopies,
so won't delete duplicate files unless there's an annexed copy.

This makes import use the same method as reinject --known.

The man page already said that duplicate meant "its content is either
present in the local repository already, or git-annex knows of another
repository that contains it, or it was present in the annex before but has
been removed now". So, this is really only bringing the implementation into
line with the man page.

This commit was sponsored by Jochen Bartl on Patreon.

9 years agoresponse
Joey Hess [Tue, 7 Feb 2017 21:13:48 +0000 (17:13 -0400)]
response

9 years agoclarify description of --skip-duplicates behavior
Joey Hess [Tue, 7 Feb 2017 20:39:29 +0000 (16:39 -0400)]
clarify description of --skip-duplicates behavior

9 years agoclose as dup
Joey Hess [Tue, 7 Feb 2017 19:21:55 +0000 (15:21 -0400)]
close as dup

9 years agorespond and close
Joey Hess [Tue, 7 Feb 2017 19:18:45 +0000 (15:18 -0400)]
respond and close

9 years agoinitremote: When a uuid= parameter is passed, use the specified UUID for the new...
Joey Hess [Tue, 7 Feb 2017 19:10:41 +0000 (15:10 -0400)]
initremote: When a uuid= parameter is passed, use the specified UUID for the new special remote, instead of generating a UUID.

This can be useful in some situations, eg when the same data can be
accessed via two different special remote backends.

9 years agoadd SetupStage parameter to RemoteType.setup
Joey Hess [Tue, 7 Feb 2017 18:35:58 +0000 (14:35 -0400)]
add SetupStage parameter to RemoteType.setup

Most remotes have an idempotent setup that can be reused for
enableremote, but in a few cases, it needs to tell which, and whether
a UUID was provided to setup was used.

This is groundwork for making initremote be able to provide a UUID.
It should not change any behavior.

Note that it would be nice to make the UUID always be provided to setup,
and make setup not need to generate and return a UUID. What prevented
this simplification is Remote.Git.gitSetup, which needs to reuse the
UUID of the git remote when setting it up, and so has to return that
UUID.

This commit was sponsored by Thom May on Patreon.

9 years agocomment
Joey Hess [Tue, 7 Feb 2017 17:51:23 +0000 (13:51 -0400)]
comment

9 years agoassistant: Make --autostart --foreground wait for the children it starts.
Joey Hess [Tue, 7 Feb 2017 17:31:45 +0000 (13:31 -0400)]
assistant: Make --autostart --foreground wait for the children it starts.

Before, the --foreground was ignored when autostarting.

This commit was sponsored by Denis Dzyubenko on Patreon.

9 years agocomment
Joey Hess [Tue, 7 Feb 2017 17:11:02 +0000 (13:11 -0400)]
comment

9 years agocomment
Joey Hess [Tue, 7 Feb 2017 17:08:33 +0000 (13:08 -0400)]
comment

9 years agoFix build with aws 0.16. Thanks, aristidb.
Joey Hess [Tue, 7 Feb 2017 17:01:57 +0000 (13:01 -0400)]
Fix build with aws 0.16. Thanks, aristidb.

9 years agocomment
Joey Hess [Tue, 7 Feb 2017 16:56:13 +0000 (12:56 -0400)]
comment

9 years agocomment
Joey Hess [Tue, 7 Feb 2017 16:48:12 +0000 (12:48 -0400)]
comment

9 years agocomment
Joey Hess [Tue, 7 Feb 2017 16:38:31 +0000 (12:38 -0400)]
comment

9 years agoAdded a comment
yarikoptic [Mon, 6 Feb 2017 20:46:29 +0000 (20:46 +0000)]
Added a comment

9 years agoinitial whining
yarikoptic [Mon, 6 Feb 2017 19:25:47 +0000 (19:25 +0000)]
initial whining

9 years agoinitial thoughts
yarikoptic [Mon, 6 Feb 2017 15:32:45 +0000 (15:32 +0000)]
initial thoughts

9 years agoinitial thoughts
yarikoptic [Mon, 6 Feb 2017 00:49:28 +0000 (00:49 +0000)]
initial thoughts

9 years ago(no commit message)
http://xgm.de/oid/ [Sun, 5 Feb 2017 18:17:08 +0000 (18:17 +0000)]

9 years agoAdded a comment
kristianrumberg@85de814abe9398ab5573b77bb8a7a80ceaf3b9e1 [Sun, 5 Feb 2017 17:02:10 +0000 (17:02 +0000)]
Added a comment

9 years agoAdded feature request for local state variables in special remote protocol
trenta3 [Sun, 5 Feb 2017 15:52:55 +0000 (15:52 +0000)]
Added feature request for local state variables in special remote protocol

9 years agoAdded a comment
aristidb@4551cc1215222a56e271a796b92908680f3941b5 [Sun, 5 Feb 2017 12:49:22 +0000 (12:49 +0000)]
Added a comment

9 years agoAdded a comment
aristidb@4551cc1215222a56e271a796b92908680f3941b5 [Sun, 5 Feb 2017 12:29:27 +0000 (12:29 +0000)]
Added a comment

9 years ago(no commit message)
aristidb@4551cc1215222a56e271a796b92908680f3941b5 [Sun, 5 Feb 2017 12:26:10 +0000 (12:26 +0000)]

9 years ago(no commit message)
ilovezfs [Sun, 5 Feb 2017 12:21:37 +0000 (12:21 +0000)]

9 years ago(no commit message)
http://xgm.de/oid/ [Sun, 5 Feb 2017 10:37:00 +0000 (10:37 +0000)]