This adds the overhead of a copy when serializing and deserializing keys.
I have not benchmarked much, but runtimes seem barely changed at all by that.
When a lot of keys are in memory, it improves memory use.
And, it prevents keys sometimes getting PINNED in memory and failing to GC,
which is a problem ByteString has sometimes. In particular, git-annex sync
from a borg special remote had that problem and this improved its memory
use by a large amount.
Sponsored-by: Shae Erisson on Patreon
Necessarily threw out the old protocol, so if an old git-annex assistant
is running, and starts a transferkeys from the new git-annex, it would
fail. But, that seems unlikely; the assistant starts up transferkeys
processes and then keeps them running. Still, may need to test that
scenario.
The new protocol is simple read/show and looks like this:
TransferRequest Download (Right "origin") (Key {keyName = "f8f8766a836fb6120abf4d5328ce8761404e437529e997aaa0363bdd4fecd7bb", keyVariety = SHA2Key (HashSize 256) (HasExt True), keySize = Just 30, keyMtime = Nothing, keyChunkSize = Nothing, keyChunkNum = Nothing}) (AssociatedFile (Just "foo"))
TransferOutput (ProgressMeter (Just 30) (MeterState {meterBytesProcessed = BytesProcessed 0, meterTimeStamp = 1.6070268727892535e9}) (MeterState {meterBytesProcessed = BytesProcessed 30, meterTimeStamp = 1.6070268728043e9}))
TransferOutput (OutputMessage "(checksum...) ")
TransferResult True
Granted, this is not optimally fast, but it seems good enough, and is
probably nearly as fast as the old protocol anyhow.
emitSerializedOutput for ProgressMeter is not yet implemented. It needs
to somehow start or update a progress meter. There may need to be a new
message that allocates a progress meter, and then have ProgressMeter
update it.
This commit was sponsored by Ethan Aubin
This is groundwork for external backends, but also makes sense to keep
this information with the rest of a Backend's implementation.
Also, removed isVerifiable. I noticed that the same information is
encoded by whether a Backend implements verifyKeyContent or not.
Finally builds (oh the agoncy of making it build), but still very
unmergable, only Command.Find is included and lots of stuff is badly
hacked to make it compile.
Benchmarking vs master, this git-annex find is significantly faster!
Specifically:
num files old new speedup
48500 4.77 3.73 28%
12500 1.36 1.02 66%
20 0.075 0.074 0% (so startup time is unchanged)
That's without really finishing the optimization. Things still to do:
* Eliminate all the fromRawFilePath, toRawFilePath, encodeBS,
decodeBS conversions.
* Use versions of IO actions like getFileStatus that take a RawFilePath.
* Eliminate some Data.ByteString.Lazy.toStrict, which is a slow copy.
* Use ByteString for parsing git config to speed up startup.
It's likely several of those will speed up git-annex find further.
And other commands will certianly benefit even more.
This will speed up the common case where a Key is deserialized from
disk, but is then serialized to build eg, the path to the annex object.
Previously attempted in 4536c93bb2
and reverted in 96aba8eff7.
The problems mentioned in the latter commit are addressed now:
Read/Show of KeyData is backwards-compatible with Read/Show of Key from before
this change, so Types.Distribution will keep working.
The Eq instance is fixed.
Also, Key has smart constructors, avoiding needing to remember to update
the cached serialization.
Used git-annex benchmark:
find is 7% faster
whereis is 3% faster
get when all files are already present is 5% faster
Generally, the benchmarks are running 0.1 seconds faster per 2000 files,
on a ram disk in my laptop.
Rescued from commit 11d6e2e260 which removed
db benchmarks in favor of benchmarking arbitrary git-annex commands. Which
is nice and general, but microbenchmarks are useful too.
using a blake2 variant optimised for 4-way CPUs
This had been deferred because the Debian package of cryptonite, and
possibly other builds, was broken for blake2bp, but I've confirmed #892855
is fixed.
This commit was sponsored by Brett Eisenberg on Patreon.
Drop support for building with ghc older than 8.4.4, and with older
versions of serveral haskell libraries than will be included in Debian 10.
The only remaining version ifdefs in the entire code base are now a couple
for aws!
This commit should only be merged after the Debian 10 release.
And perhaps it will need to wait longer than that; it would make
backporting new versions of git-annex to Debian 9 (stretch) which
has been actively happening as recently as this year.
This commit was sponsored by Ilya Shlyakhter.
This does not change the overall license of the git-annex program, which
was already AGPL due to a number of sources files being AGPL already.
Legally speaking, I'm adding a new license under which these files are
now available; I already released their current contents under the GPL
license. Now they're dual licensed GPL and AGPL. However, I intend
for all my future changes to these files to only be released under the
AGPL license, and I won't be tracking the dual licensing status, so I'm
simply changing the license statement to say it's AGPL.
(In some cases, others wrote parts of the code of a file and released it
under the GPL; but in all cases I have contributed a significant portion
of the code in each file and it's that code that is getting the AGPL
license; the GPL license of other contributors allows combining with
AGPL code.)
Adding that field broke the Read/Show serialization back-compat,
and also the Eq and Ord instances were not blinded to it, which broke
git annex fsck and probably more.
I think that the new approach used in formatKeyVariety will be nearly
as fast, but have not benchmarked it.
This reverts commit 4536c93bb2.
That broke Read/Show of a Key, and unfortunately Key is read in at least
one place; the GitAnnexDistribution data type.
It would be worth bringing this optimisation back, but it would need
either a custom Read/Show instance that preserves back-compat, or
wrapping Key in a data type that contains the serialization, or changing
how GitAnnexDistribution is serialized.
Also, the Eq instance would need to compare keys with and without a
cached seralization the same.
This will speed up the common case where a Key is deserialized from
disk, but is then serialized to build eg, the path to the annex object.
It means that every place a Key has any of its fields changed, the cache
has to be dropped. I've grepped and found them all. But, it would be
better to avoid that gotcha somehow..
The new parser is significantly stricter than the old one:
The old file2key allowed the fields to come in any order,
but the new one requires the fixed order that git-annex has always used.
Hopefully this will not cause any breakage.
And the old file2key allowed eg SHA1-m1-m2-m3-m4-m5-m6--xxxx
while the new does not allow duplication of fields. This could potentially
improve security, because allowing lots of extra junk like that in a key
could potentially be used in a SHA1 collision attack, although the current
attacks need binary data and not this kind of structured numeric data.
Speed improved of course, and fairly substantially, in microbenchmarks:
benchmarking old/key2file
time 2.264 μs (2.257 μs .. 2.273 μs)
1.000 R² (1.000 R² .. 1.000 R²)
mean 2.265 μs (2.260 μs .. 2.275 μs)
std dev 21.17 ns (13.06 ns .. 39.26 ns)
benchmarking new/key2file'
time 1.744 μs (1.741 μs .. 1.747 μs)
1.000 R² (1.000 R² .. 1.000 R²)
mean 1.745 μs (1.742 μs .. 1.751 μs)
std dev 13.55 ns (9.099 ns .. 21.89 ns)
benchmarking old/file2key
time 6.114 μs (6.102 μs .. 6.129 μs)
1.000 R² (1.000 R² .. 1.000 R²)
mean 6.118 μs (6.106 μs .. 6.143 μs)
std dev 55.00 ns (30.08 ns .. 100.2 ns)
benchmarking new/file2key'
time 1.791 μs (1.782 μs .. 1.801 μs)
1.000 R² (0.999 R² .. 1.000 R²)
mean 1.792 μs (1.785 μs .. 1.804 μs)
std dev 32.46 ns (20.59 ns .. 50.82 ns)
variance introduced by outliers: 19% (moderately inflated)
This will be used to protect against CVE-2018-10859, where an encrypted
special remote is fed the wrong encrypted data, and so tricked into
decrypting something that the user encrypted with their gpg key and did
not store in git-annex.
It also protects against CVE-2018-10857, where a remote follows a http
redirect to a file:// url or to a local private web server. While that's
already been prevented in git-annex's own use of http, external special
remotes, hooks, etc use other http implementations and could still be
vulnerable.
The policy is not yet enforced, this commit only adds the appropriate
metadata to remotes.
This commit was sponsored by Boyd Stephen Smith Jr. on Patreon.
Some blake hash varieties were not yet available in that version.
Rather than tracking exact details of what cryptonite supported when,
disable blake unless using a current cryptonite.
There are a lot of different variants and sizes, I suppose we might as well
export all the common ones.
Bump dep to cryptonite to 0.16, earlier versions lacked BLAKE2 support.
Even android has 0.16 or newer.
On Debian, Blake2bp_512 is buggy, so I have omitted it for now.
http://bugs.debian.org/892855
This commit was sponsored by andrea rota.
Note that GPGHMAC keys are not cryptographically secure, because their
content has no relation to the name of the key. So, things that use this
function to avoid sending keys to a remote will need to special case in
support for those keys. If GPGHMAC keys were accepted as
cryptographically secure, symlinks using them could be committed to a
git repo, and their content would be accepted into the repo, with no
guarantee that two repos got the same content, which is what we're aiming
to prevent.
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.
I am not happy that I had to put backend-specific code in file2key. But
it would be very difficult to avoid this layering violation.
Most of the time, when parsing a Key from a symlink target, git-annex
never looks up its Backend at all, so adding this check to a method of
the Backend object would not work.
The Key could be made to contain the appropriate
Backend, but since Backend is parameterized on an "a" that is fixed to
the Annex monad later, that would need Key to change to "Key a".
The only way to clean this up that I can see would be to have the Key
contain a LowlevelBackend, and put the validation in LowlevelBackend.
Perhaps later, but that would be an extensive change, so let's not do
it in this commit which may want to cherry-pick to backports.
This commit was sponsored by Ethan Aubin.
For use with tor hidden services, and perhaps other transports later.
Based on Utility.SimpleProtocol, it's a line-based protocol,
interspersed with transfers of bytestrings of a specified size.
Implementation of the local and remote sides of the protocol is done
using a free monad. This lets monadic code be included here, without
tying it to any particular way to get bytes peer-to-peer.
This adds a dependency on the haskell package "free", although that
was probably pulled in transitively from other dependencies already.
This commit was sponsored by Jeff Goeke-Smith on Patreon.
This is needed because when preferred content matches on files,
the second pass would otherwise want to drop all keys. Using a bloom filter
avoids this, and in the case of a false positive, a key will be left
undropped that preferred content would allow dropping. Chances of that
happening are a mere 1 in 1 million.
This fixes all instances of " \t" in the code base. Most common case
seems to be after a "where" line; probably vim copied the two space layout
of that line.
Done as a background task while listening to episode 2 of the Type Theory
podcast.
Leverage the new chunked remotes to automatically resume downloads.
Sort of like rsync, although of course not as efficient since this
needs to start at a chunk boundry.
But, unlike rsync, this method will work for S3, WebDAV, external
special remotes, etc, etc. Only directory special remotes so far,
but many more soon!
This implementation will also properly handle starting a download
from one remote, interrupting, and resuming from another one, and so on.
(Resuming interrupted chunked uploads is similarly doable, although
slightly more expensive.)
This commit was sponsored by Thomas Djärv.
This avoids a proliferation of hash directories when using new-style
chunking, and should improve performance since chunks are accessed
in sequence and so should have a common locality.
Of course, when a chunked key is encrypted, its hash directories have no
relation to the parent key.
This commit was sponsored by Christian Kellermann.
Added new fields for chunk number, and chunk size. These will not appear
in normal keys ever, but will be used for chunked data stored on special
remotes.
This commit was sponsored by Jouni K Seppanen.
Found this in failed armhf build log, where quickcheck found a way to break
prop_idempotent_key_decode. The "s" indicates size, but since nothing comes
after it, that's not valid. When encoding the resulting key, no size was
present, so it encoded to "a--a".
Also, "a-sX--a" is not legal, since X is not a number. Not found by
quickcheck.
A file named "foo-" or "foo-bar" was taken as a key's file, with a backend
of "foo", and an empty keyName. This led to various problems, especially
because converting that key back to a file did not yeild the same filename.