* init: When annex.securehashesonly has been set with git-annex config,
copy that value to the annex.securehashesonly git config.
* config --set: As well as setting value in git-annex branch,
set local gitconfig. This is needed especially for
annex.securehashesonly, which is read only from local gitconfig and not
the git-annex branch.
doc/todo/sha1_collision_embedding_in_git-annex_keys.mdwn has the
rationalle for doing it this way. There's no perfect solution; this
seems to be the least-bad one.
This commit was supported by the NSF-funded DataLad project.
Cryptographically secure hashes can be forced to be used in a repository,
by setting annex.securehashesonly. This does not prevent the git repository
from containing files with insecure hashes, but it does prevent the content
of such files from being pulled into .git/annex/objects from another
repository.
We want to make sure that at no point does git-annex accept content into
.git/annex/objects that is hashed with an insecure key. Here's how it
was done:
* .git/annex/objects/xx/yy/KEY/ is kept frozen, so nothing can be
written to it normally
* So every place that writes content must call, thawContent or modifyContent.
We can audit for these, and be sure we've considered all cases.
* The main functions are moveAnnex, and linkToAnnex; these were made to
check annex.securehashesonly, and are the main security boundary
for annex.securehashesonly.
* Most other calls to modifyContent deal with other files in the KEY
directory (inode cache etc). The other ones that mess with the content
are:
- Annex.Direct.toDirectGen, in which content already in the
annex directory is moved to the direct mode file, so not relevant.
- fix and lock, which don't add new content
- Command.ReKey.linkKey, which manually unlocks it to make a
copy.
* All other calls to thawContent appear safe.
Made moveAnnex return a Bool, so checked all callsites and made them
deal with a failure in appropriate ways.
linkToAnnex simply returns LinkAnnexFailed; all callsites already deal
with it failing in appropriate ways.
This commit was sponsored by Riku Voipio.
9c4650358c changed the Read instance for
Key.
I've checked all uses of that instance (by removing it and seeing what
breaks), and they're all limited to the webapp, except one.
That is GitAnnexDistribution's Read instance.
So, 9c4650358c would have broken upgrades
of git-annex from downloads.kitenet.net. Once the .info files there got
updated for a new release, old releases would have failed to parse them
and never upgraded.
To fix this, I found a way to make the .info files that contain
GitAnnexDistribution values be readable by the old version of git-annex.
This commit was sponsored by Ewen McNeill.
Where before the "name" of a key and a backend was a string, this makes
it a concrete data type.
This is groundwork for allowing some varieties of keys to be disabled
in file2key, so git-annex won't use them at all.
Benchmarks ran in my big repo:
old git-annex info:
real 0m3.338s
user 0m3.124s
sys 0m0.244s
new git-annex info:
real 0m3.216s
user 0m3.024s
sys 0m0.220s
new git-annex find:
real 0m7.138s
user 0m6.924s
sys 0m0.252s
old git-annex find:
real 0m7.433s
user 0m7.240s
sys 0m0.232s
Surprising result; I'd have expected it to be slower since it now parses
all the key varieties. But, the parser is very simple and perhaps
sharing KeyVarieties uses less memory or something like that.
This commit was supported by the NSF-funded DataLad project.
This was noticed because it broke the datalad test suite, which pushed
to the remote and then fetched to check if it had received the expected
branches. Auto-init caused the git-annex branch on the remote to
diverge, breaking that test.
https://github.com/datalad/datalad/issues/1319#issuecomment-281649518
The auto-init still happens, it's staged in the journal, and will be
commited by some later git-annex command when it runs. Which is fine,
it's the same as that later command doing the auto-init.
This commit was supported by the NSF-funded DataLad project
sync: When syncing with a local repository located on a crippled
filesystem, run the post-receive hook there, since it wouldn't get run
otherwise. This makes pushing to repos on FAT-formatted removable drives
update them when receive.denyCurrentBranch=updateInstead.
Made Remote.Git export onLocal, which was cleaned up to not have so many
caveats about its use.
This commit was sponsored by Jeff Goeke-Smith on Patreon.
* Added post-recieve hook, which makes updateInstead work with direct
mode and adjusted branches.
* init: Set up the post-receive hook.
This commit was sponsored by Fernando Jimenez on Patreon.
Recent changes had a side effect of displaying errors in the fairly
common case when this push fails. Since the synced/git-annex push
is always forced, those errors are noise, so hide again.
This means 3 separate pushes are done now, where before it only made 2.
A bit more expensive, but ssh connection caching eliminates most of
the costs.
This commit was sponsored by Boyd Stephen Smith Jr. on Patreon.
config group groupwanted numcopies schedule wanted required: Avoid
displaying extraneous messages about repository auto-init, git-annex branch
merging, etc, when being used to get information.
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.
... 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.
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.
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.
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.
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.
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.
Wormhole pairing will start to provide an appid to wormhole on 2021-12-31.
An appid can't be provided now because Debian stable is going to ship a
older version of git-annex that does not provide an appid. Assumption is
that by 2021-12-31, this version of git-annex will be shipped in a Debian
stable release. If that turns out to not be the case, this change will need
to be cherry-picked into the git-annex in Debian stable, or its wormhole
pairing will break.
This commit was sponsored by Thomas Hochstein on Patreon.
.. which can be set to true to make git annex sync default to --content.
This may become the default at some point in the future.
As well as being configuable by git config, it can be configured by
git-annex config to control the default behavior in all clones of a
repository.
Had to add a separate --no-content switch to we can tell if it's been
explicitly set, and should override annex.synccontent. If --content was the
default, this complication would not be necessary.
This commit was sponsored by Jake Vosloo on Patreon.
... to control the default behavior in all clones of a repository.
This includes a new Configurable data type, so the GitConfig type indicates
which values can be configured this way.
The implementation should be quite efficient; the config log is only read
once, and only when a Configurable value has not already been set by
git-config.
Indeed, it would be nice in the future to extend this, so that git-config
is itself only read on demand. Some commands may not need to look at the
git configuration at all.
This commit was sponsored by Trenton Cronholm on Patreon.
This makes it a little bit slower since it has to check file size,
but worth it to fix a potential memory use problem.
This commit was sponsored by Fernando Jimenez on Patreon.
Turns out that Data.List.Utils.split is slow and makes a lot of
allocations. Here's a much simpler single character splitter that behaves
the same (even in wacky corner cases) while running in half the time and
75% the allocations.
As well as being an optimisation, this helps move toward eliminating use of
missingh.
(Data.List.Split.splitOn is nearly as slow as Data.List.Utils.split and
allocates even more.)
I have not benchmarked the effect on git-annex, but would not be surprised
to see some parsing of eg, large streams from git commands run twice as
fast, and possibly in less memory.
This commit was sponsored by Boyd Stephen Smith Jr. on Patreon.
Any config names can be set using this; git-annex commands will only look
at specific ones that make sense and are worth the overhead of querying the
branch.
This might also be useful for storing whatever other config-type stuff the
user might want to shove into the git-annex branch.
This commit was sponsored by Jochen Bartl on Patreon.
Docs say vicfg can configure everything from git-annex branch,
so it ought to configure numcopies.
Note that commenting out existing numcopies does not unset it.
This commit was sponsored by Thom May on Patreon.
Since the user does not know whether it will run su or sudo, indicate
whether the password prompt will be for root or the user's password,
when possible.
I assume that programs like gksu that can prompt for either depending on
system setup will make clear in their prompt what they're asking for.
This way we know that after enable-tor, the tor hidden service is fully
published and working, and so there should be no problems with it at
pairing time.
It has to start up its own temporary listener on the hidden service. It
would be nice to have it start the remotedaemon running, so that extra
step is not needed afterwards. But, there may already be a remotedaemon
running, in communication with the assistant and we don't want to start
another one. I thought about trying to HUP any running remotedaemon, but
Windows does not make it easy to do that. In any case, having the user
start the remotedaemon themselves lets them know it needs to be running
to serve the hidden service.
This commit was sponsored by Boyd Stephen Smith Jr. on Patreon.
weasel explained that apparmor limits on what files tor can read do not
apply to sockets (because they're not files). And apparently the
problems I was seeing with hidden services not being accessible had to
do with onion address propigation and not the location of the socket
file.
remotedaemon looks up the HiddenServicePort in torrc, so if it was
previously configured with the socket in /etc, that will still work.
This commit was sponsored by Denis Dzyubenko on Patreon.
This reverts commit 3037feb1bf.
On second thought, this was an overcomplication of what should be the
lowest-level primitive. Let's build bi-directional links at the pairing
level with eg magic wormhole.
Both the local and remote git repositories get remotes added
pointing at one-another.
Makes pairing twice as easy!
Security: The new LINK command in the protocol can be sent repeatedly,
but only by a peer who has authenticated with us. So, it's entirely safe to
add a link back to that peer, or to some other peer it knows about.
Anything we receive over such a link, the peer could send us over the
current connection.
There is some risk of being flooded with LINKs, and adding too many
remotes. To guard against that, there's a hard cap on the number of remotes
that can be set up this way. This will only be a problem if setting up
large p2p networks that have exceptional interconnectedness.
A new, dedicated authtoken is created when sending LINK.
This also allows, in theory, using a p2p network like tor, to learn about
links on other networks, like telehash.
This commit was sponsored by Bruno BEAUFILS on Patreon.
1 microsecond delay is ugly.. but, maintaining an queue of a list of timestamps
and taking a new one from the queue each time around, or maintaining a timestamp
counter, would probably be slower.
Added to change notification to P2P protocol.
Switched to a TBChan so that a single long-running thread can be
started, and serve perhaps intermittent requests for change
notifications, without buffering all changes in memory.
The P2P runner currently starts up a new thread each times it waits
for a change, but that should allow later reusing a thread. Although
each connection from a peer will still need a new watcher thread to run.
The dependency on stm-chans is more or less free; some stuff in yesod
uses it, so it was already indirectly pulled in when building with the
webapp.
This commit was sponsored by Francois Marier on Patreon.
The attacker could just send a very lot of data, with no \n and it would
all be buffered in memory until the kernel killed git-annex or perhaps OOM
killed some other more valuable process.
This is a low impact security hole, only affecting communication between
local git-annex and git-annex-shell on the remote system. (With either
able to be the attacker). Only those with the right ssh key can do it. And,
there are probably lots of ways to construct git repositories that make git
use a lot of memory in various ways, which would have similar impact as
this attack.
The fix in P2P/IO.hs would have been higher impact, if it had made it to a
released version, since it would have allowed DOSing the tor hidden
service without needing to authenticate.
(The LockContent and NotifyChanges instances may not be really
exploitable; since the line is read and ignored, it probably gets read
lazily and does not end up staying buffered in memory.)
Would have liked to make the Parser parse the file and key pairs, but it
seems that optparse-applicative is unable to handle eg:
many ((,) <$> argument <*> argument)
This commit was sponsored by Thomas Hochstein on Patreon.
* rmurl: Multiple pairs of files and urls can be provided on the
command line.
* rmurl: Added --batch mode.
This commit was sponsored by Trenton Cronholm on Patreon.
* map: Run xdot if it's available in PATH. On OSX, the dot command
does not support graphical display, while xdot does.
* Debian: xdot is a better interactive viewer than dot, so Suggest
xdot, rather than graphviz.
The file matcher needs to be run on the destination file not the tmp
file, in order for filename matches to work properly. However, it also
needs to be able to probe the file for size and mime type.
This is a quick fix to a regression. The double rename is not pretty.
It would be good to either have a way to run the largeFileMatcher
such that it is matching on the final filename but looks at the temp
file, or to make addAnnexedFile not need the temp file in a different
location.
Almost working, but there's a bug in the relaying.
Also, made tor hidden service setup pick a random port, to make it harder
to port scan.
This commit was sponsored by Boyd Stephen Smith Jr. on Patreon.
ghc 8 added backtraces on uncaught errors. This is great, but git-annex was
using error in many places for a error message targeted at the user, in
some known problem case. A backtrace only confuses such a message, so omit it.
Notably, commands like git annex drop that failed due to eg, numcopies,
used to use error, so had a backtrace.
This commit was sponsored by Ethan Aubin.
This makes merging a remote into a freshly created direct mode repository
work the same as it works in indirect mode.
The git-annex branches would get merged in any case by a sync,
since that doesn't use git merge.
This might need to be revisited later to better mirror git's behavior.
This avoids needing to bind to the right port before something else
does.
The socket is in /var/run/user/$uid/ which ought to be writable by only
that uid. At least it is on linux systems using systemd.
For Windows, may need to revisit this and use ports or something.
The first version of tor to support sockets for hidden services
was 0.2.6.3. That is not in Debian stable, but is available in
backports.
This commit was sponsored by andrea rota.
I've long considered the XMPP support in git-annex a wart.
It's nice to remove it.
(This also removes the NetMessager, which was only used for XMPP, and the
daemonstatus's desynced list (likewise).)
Existing XMPP remotes should be ignored by git-annex.
This commit was sponsored by Brock Spratlen on Patreon.
Tor unfortunately does not come out of the box configured to let hidden
services register themselves on the fly via the ControlPort.
And, changing the config to enable the ControlPort and a particular type
of auth for it may break something already using the ControlPort, or
lessen the security of the system.
So, this leaves only one option to us: Add a hidden service to the
torrc. git-annex enable-tor does so, and picks an unused high port for
tor to listen on for connections to the hidden service.
It's up to the caller to somehow pick a local port to listen on
that won't be used by something else. That may be difficult to do..
This commit was sponsored by Jochen Bartl on Patreon.
If a transfer fails for some reason, but some data managed to be sent, the
transfer will be retried. (The assistant already did this.)
Possible impacts:
* More ssh prompts if ssh needs to prompt for a password to connect to a
host, or is prompting about some other problem like a ssh key mismatch.
* More data transfer due to retrying, epecially when a remote does not
support resuming a transfer.
In the worst case, a lot of data will be transferred but it fails before
the end, and then all that data gets transferred again plus one byte more;
repeat until it manages to get the whole file.
Closes https://github.com/datalad/datalad/issues/1020
The use of runWriter in scanUnlockedFiles broke due to this change;
it failed with blocked indefinitely in mvar, because the database write
handle was taken while linkFromAnnex needed to also write to it (to update
the inode cache). So, switched to using a separate runWriter for each call
to addAssociatedFileFast. A little less efficient, but not greatly; the
writes should all still be cached.
In the case where the pointer file is in place, and not the content
of the object, lock's performNew was called with filemodified=True,
which caused it to try to repopulate the object from an unmodified
associated file, of which there were none. So, the content of the object
got thrown away incorrectly. This was the cause (although not the root
cause) of data loss in https://github.com/datalad/datalad/issues/1020
The same problem could also occur when the work tree file is modified,
but the object is not, and lock is called with --force. Added a test case
for this, since it's excercising the same code path and is easier to set up
than the problem above.
Note that this only occurred when the keys database did not have an inode
cache recorded for the annex object. Normally, the annex object would be in
there, but there are of course circumstances where the inode cache is out
of sync with reality, since it's only a cache.
Fixed by checking if the object is unmodified; if so we don't need to
try to repopulate it. This does add an additional checksum to the unlock
path, but it's already checksumming the worktree file in another case,
so it doesn't slow it down overall.
Further investigation found a similar problem occurred when smudge --clean
is called on a file and the inode cache is not populated. cleanOldKeys
deleted the unmodified old object file in this case. This was also
fixed by checking if the object is unmodified.
In general, use of getInodeCaches and sameInodeCache is potentially
dangerous if the inode cache has not gotten populated for some reason.
Better to use isUnmodified. I breifly auited other places that check the
inode cache, and did not see any immediate problems, but it would be easy
to miss this kind of problem.
Fixes a bug introduced with v6 mode that I didn't notice until now.
Probably not many v3 repos left out there, and upgrading them to v6 mode
is not disastrous, only a little premature.
This commit was sponsored by Riku Voipio
* sync: Previously, when run in a branch with a slash in its name,
such as "foo/bar", the sync branch was "synced/bar". That conflicted
with the sync branch used for branch "bar", so has been changed to
"synced/foo/bar".
* adjust: Previously, when adjusting a branch with a slash in its name,
such as "foo/bar", the adjusted branch was "adjusted/bar(unlocked)".
That conflicted with the adjusted branch used for branch "bar",
so has been changed to "adjusted/foo/bar(unlocked)"
* Also, running sync in an adjusted branch did not correctly sync
changes back to the parent branch when it had a slash in its name.
This bug has been fixed.
Eliminate use of Git.Ref.under and Git.Ref.basename; using
Git.Ref.underBase and Git.Ref.base make everything handle deep branches
correctly.
Probably noone was adjusting deep branches, and v6 is still experimental
anyway, so I'm not going to worry about the mess that was left by that bug.
In the case of git-annex sync, using a fixed git-annex with an old unfixed
one will mean they use different sync branches for a deep branch, and so
they may stop syncing until the old one is upgraded. However, that's only
a problem when syncing between repositories without going via a central
bare repository. Added a warning about this to the CHANGELOG, but it's
probably not going to affect many people at all.
This commit was sponsored by Riku Voipio.
Avoid threads emitting json at the same time and scrambling, which was
still possible even with the buffering, just less likely.
Converted json IO actions to JSONChunk data too.
This makes -Jn work with --json and --quiet, where before
setting -Jn disabled those options.
Concurrent json output is currently a mess though since threads output
chunks over top of one-another.
Only done in -J mode because only if there's concurrency can downloading
from two remotes be faster. Without concurrency, it's likely the case that
sequential downloads from the same remote are faster than switching back
and forth between two remotes.
There is some hairy MVar code here, but basically it just keeps
the activeremotes MVar full except when deciding which remote to assign
to a thread.
Also affects gets by sync --content -J
This commit was sponsored by Jochen Bartl.
This was disabled in commit 61ccf95004,
because only the assistant used them, and they were clutter. But, now
--failed also uses them.
Remove the failure log files after successful transfers. Should avoid
most of the clutter problems.
Commit 61ccf95004 mentions a subtle behavior
change, which has now been reverted:
There is one behavior change from this. If glacier is being used, and a
manual git annex get --from glacier fails because the file isn't available
yet, the assistant will no longer later see that failed transfer file and
retry the get.
Note that get --from foo --failed will get things that a previous get --from bar
tried and failed to get, etc. I considered making --failed only retry
transfers from the same remote, but it was easier, and seems more useful,
to not have the same remote requirement.
Noisy due to some refactoring into Types/
metadata --json output format has changed, adding a inner json object
named "fields" which contains only the fields and their values.
This should be easier to parse than the old format, which mixed up
metadata fields with other keys in the json object.
Any consumers of the old format will need to be updated.
This adds a dependency on unordered-containers for parsing MetaData
from JSON, but it's a free dependency; aeson pulls in that library.
Keeping Text.JSON use for now, because it seems a better fit for most of
the commands, which don't use very structured JSON objects, but just output
whatever fields suites them. But this lets Aeson be used when a more
structured data type is available to serialize to JSON.
Show branch:file that is being operated on.
I had to make ActionItem a type and not a type class because
withKeyOptions' passed two different types of values when using the type
class, and I could not get the type checker to accept that.
Added --branch option to copy, drop, fsck, get, metadata, mirror, move, and
whereis commands. This option makes git-annex operate on files that are
included in a specified branch (or other treeish).
The names of the files from the branch that are being operated on are not
displayed yet; only the keys. Displaying the filenames will need changes
to every affected command.
Also, note that --branch can be specified repeatedly. This is not really
documented, but seemed worth supporting, especially since we may later want
the ability to operate on all branches matching a refspec. However, when
operating on two branches that contain the same key, that key will be
operated on twice.
I noticed move --to failing when there was no disk space. The file was sent
to the remote, but it crashed before it could be dropped locally. This
could fix that.
Added guard in Annex.Transfer to prevent this problem at a deeper level.
I'm unhappy ith NoUUID, but having Maybe UUID instead wouldn't help either
if nothing checked that there was a UUID. Since there legitimately need to
be Remotes that do not have a UUID, I can't see a way to fix it at the type
level, short making there be two separate types of Remotes.
Removed the instance LensGpgEncParams RemoteConfig because it encouraged
code that does not take the RemoteGitConfig into account.
RemoteType's setup was changed to take a RemoteGitConfig,
although the only place that is able to provide a non-empty one is
enableremote, when it's changing an existing remote. This led to several
folow-on changes, and got RemoteGitConfig plumbed through.
There's a potential race where the smudge filter is run at the same time an
object is being downloaded. If the download finished after the inAnnex
check, and before the keys db was updated, the associated file would not
get updated with the downloaded content.
I'm not sure this closes the race; it may only narrow the window. Problem
is, the keys database needs to communicate between two different processes.
In the case of the assistant, the transferkeys command is the other
process, and it closes the db handle after getting the file. So, it should
re-open the database and so see the update that the smudge filter has
written to it. But, what if the smudge filter takes a while to update the
database?
The keys database handle needs to be closed after merging, because the
smudge filter, in another process, updates the database. Old cached info
can be read for a while from the open database handle; closing it ensures
that the info written by the smudge filter is available.
This is pretty horribly ad-hoc, and it's especially nasty that the
transferrer closes the database every time.
Could not think of a foolproof way to detect if the old adjusted branch was
just behind the current branch. It's possible that the user amended the
adjusting commit at the head of the adjusted branch, for example.
I decided to bail in this situation, instead of just entering the old
branch, so that if git annex adjust succeeds the user is always in a
*current* adjusted branch, not some old and out of date one.
What could perhaps be done is enter the old branch and then update it. But
that seems too magical; the user may have rebased master or something or
may not want to propigate the changes from the old branch. Best to error
out.
