Summary:
It's possible that the non-mergedriver code path decides to create ("get") a
file, because it only exists in one side of the merge parents, while
mergedriver decides to remove that file. That makes the file exist in both
"g" (get, aka. create) and "r" (remove) lists and result in incorrect
dirstate state.
Fix it by detecting the case and remove the conflicted files from the "g" list.
Reviewed By: DurhamG
Differential Revision: D14145990
fbshipit-source-id: f4e5d9787dca5cda6cae270f8a84296ecddf553c
Summary: the push-path will be used for Mononoke write path roll out.
Reviewed By: markbt
Differential Revision: D14153207
fbshipit-source-id: 158e6e08dfa017c98e668fdfc37138dd7b7d76c7
Summary:
We can't run in parallel at the moment as the log file and the lock file are
shared.
Every path maintains independent backup state (the previous diff).
The secondary backup state doesn't affect smartlog (only the main one)
The issue with this approach is that we maintain backup lock a bit longer.
Unfortunately, the progress in smartlog doesn't show anything about the second backup.
I added 'finished', it makes it easier to compare in the logs.
Reviewed By: markbt
Differential Revision: D14149399
fbshipit-source-id: f90e8aac6cb8dee53d5c7468bd6adba067e13362
Summary:
We are going to support 2 different backends of Commit Cloud: Mercurial and
Mononoke.
Each of them should maintain local backup state separately.
Output of some tests have been slightly changes, this is because a separate backup
state, the same error appears earlier when we are trying the backup stacks.
The idea is to have separate backup states for different remote paths, but
there will be only one cloud sync state for the current source of
truth. We could include there the remote path and then validate that cloud sync
state is correct if the remote path has been changed.
However, for backup states it is much easier to have them separately (and we
will backup in 2 places)
Reviewed By: markbt
Differential Revision: D14138496
fbshipit-source-id: 0a7a763a395be5456cbd724bff7ebc069f03fb0e
Summary: The `hg update` command has short flags for `--merge` and `--clean`, namely `-m` and `-C`. Let's add the short versions to `hg next` and `hg prev` as well.
Reviewed By: quark-zju
Differential Revision: D14155218
fbshipit-source-id: d51f5f658b525809e4c512ffa300dea4b4cabcdf
Summary:
We're encountering an issue that I think is caused by invalid data
coming back from memcache, possibly due to our recently introduced connection
reusing. Let's add some checksums to verify that the data we put in is identical
to the data we get out for a given key.
Reviewed By: kulshrax
Differential Revision: D14141683
fbshipit-source-id: 206b51b862db7d54def02f5310b90f473d5a0d03
Summary: We weren't passing the `--merge` flag from `hg prev` and `hg next` to the underlying invocation of `hg update`. This diff fixes the problem so that `--merge` now works as expected.
Reviewed By: DurhamG
Differential Revision: D14154244
fbshipit-source-id: 4a2412cca714ec8f269eb5f2989e39821642fbb3
Summary:
Backs out D13944829 that add preloading the manifest to pushrebase. In
a situation where you receive tree commits, but the repo is hybrid, this causes
it to try to lookup the bundled tree manifests in the flat manifest revlog.
Let's just back this out for now until we can come up with an appropriate fix,
or move all our repos to treeonly.
Reviewed By: quark-zju
Differential Revision: D14151876
fbshipit-source-id: 0f7419d5b9bcd9d7ce363f4349e9e2e4a86cf713
Summary: Add the current clock value to the output of `hg debugstatus`
Reviewed By: quark-zju
Differential Revision: D14150641
fbshipit-source-id: 917ac3095bc933c042c0f057d0dbda38ef710844
Summary:
If treestate is in use, fsmonitor state is stored in the treestate, rather than
in a separate fsmonitor.state file. Update rage to understand this.
Reviewed By: quark-zju
Differential Revision: D14131131
fbshipit-source-id: d80d766625d7915b6a76f66f33e763eed23677e9
Summary:
This is a perf optimization. `unbundlereplay.respondlightly` instructs the
server to not produce the pushback parts regardless of what `replycaps` part
of the incoming bundle says. This is important, since the mononoke-hg sync will
send all the bundles in a searialized way, so we want to optimize time where
possible.
Reviewed By: StanislavGlebik
Differential Revision: D14131575
fbshipit-source-id: afec15347d43fa52b1ec64b4ac8ece5b227ccf7d
Summary:
Add a `pager.stderr` config option that when set to false, disables the
redirection of stderr to the pager.
This means progress bars will be rendered, and other information will appear
synchronously, rather than buffered by the pager, but also that Mercurial will
start outputting stderr over the top of any output that the pager produces,
potentially corrupting the pager display.
Reviewed By: simpkins
Differential Revision: D14130965
fbshipit-source-id: eba8e31fdad05c935eae13e1900c5f7e5af7ecbc
Summary:
A lot of code is duplicated between data stores, and history stores, and one
reason for it is the absence of common trait between these 2. By adding a new
Store trait it will make it easier to write generic code that works accross
data and history store.
Reviewed By: quark-zju
Differential Revision: D14091899
fbshipit-source-id: deef1d43a7d300cb3607c67554ad54f20c870e23
Summary:
Prepare the code to use 'known' wireprotocol command to check what is backed up
on the server instead of slower 'lookup'. We don't need to reestablish ssh
connection, and we can test all the nodes in one go.
Also test the 'known' request with Mononoke is working (all commands have been
tested with isbackedupnodes => isbackedupnodes2)
Reviewed By: markbt
Differential Revision: D14131383
fbshipit-source-id: 5a150b64d0e84a02357c2367879b2da8493d9167
Summary:
Use the same way to check if the commits have been backed up in Mononoke and Mercurial.
The only common way to check is 'lookup' command because Mercurial doesn't support discovery for commit cloud commits, the command 'known' is also not supported.
Also we have to go one by one because lookup doesn't got any better API.
It is still much faster than backup commits that are already there.
Introduce such check for pushbackup as well.
Remove hacky way to check it from cloud sync.
For commit cloud in Mononoke we will have backfill, so the server side check will be heavily used when you go to Mononoke at the first time.
Unfortunately connection pool module in mercurial is not good enough in detecting closed connections and can easily return a broken connection on the next call.
Reviewed By: markbt
Differential Revision: D14085849
fbshipit-source-id: d76d9a71f9efdbdfec4de3198cd428b6b693418d
Summary:
This is to be used from Mononoke->hg sync.
Currently expects only `pushrebase` bundles, e.g. one head and one book to
move.
Reviewed By: StanislavGlebik
Differential Revision: D14116130
fbshipit-source-id: 959a6e51f51e21da5592c84188e294a33057ffaa
Summary: When comparing the number of files changed, fsmonitor was incorrectly using the length of the result dict (always 4), rather than the number of files watchman returned. Use the right list of files instead.
Reviewed By: markbt
Differential Revision: D14123604
fbshipit-source-id: 94684f1f189d045b2f6a880180b15e52ba9bba8c
Summary:
this will be used to understand if server supports 'listkeyspatterns'
(Mononoke doesn't support this)
Reviewed By: DurhamG
Differential Revision: D14107742
fbshipit-source-id: c9a42e8516eb5660ab2f498996b211db7086bcb1
Summary:
In order to move the types in `edenapi-types` (containing types shared between Mercurial and Mononoke) to the `types` crate, we need to move a few types from the `revisionstore` crate into this crate first, because `revisionstore` depends on `types`, which would create a circular dependency since `edenapi-types` uses types from `revisionstore`.
In particular, this diff moves the `Key` and `NodeInfo` types into their own modules in the `types` crate.
Reviewed By: quark-zju
Differential Revision: D14114166
fbshipit-source-id: 8f9e78d610425faec9dc89ecc9e450651d24177a
Summary:
the early exit logic was incorrect, basically if there is at least 1 bookmark
in the repo and the backup state was not empty, it didn't catch that nothing has been changed.
bookmarks are dicts, so it is fine to compare them
if any bookmarks in the repo, everytime `hg pushbackup` established a connection to mercurial
Reviewed By: quark-zju
Differential Revision: D14103938
fbshipit-source-id: 0edc4d9e186199670765fd2362236330e831c7d9
Summary:
There are still users of "hg id -b", namely shell prompt scripts.
However, it seems fine to remove "branch" from the formatted (JSON or other
formats) output.
Reviewed By: ikostia
Differential Revision: D14076296
fbshipit-source-id: a77f4b9e73161d322da3f1bdec993082ac463939
Summary: Without named branch, it is not meaningful to close a branch.
Differential Revision: D14076295
fbshipit-source-id: 527ae457cb14f9415fa7269b7dc123252e7d6867
Summary:
Unfortunately, there are still users of the "hg branch" command so we cannot
remove it right now. But I'm updating internal tools to not call it.
Reviewed By: ikostia
Differential Revision: D14076300
fbshipit-source-id: a5322d9c9ce8148c6ad0800653bd829736556386
Summary:
It seems nobody uses it from dev_commands logging. Search for "allowbranches"
in all internal repos suggests nobody turned it on. The tests were also
migrated to not use the command so let's just remove it.
Reviewed By: ikostia
Differential Revision: D14076298
fbshipit-source-id: a1cb03b2da289c01b15f0580ffe77531941be213
Summary: This would make the test runner run tests with named branch disabled.
Reviewed By: phillco
Differential Revision: D14059835
fbshipit-source-id: 9fbad5688bd8c69d32016b2ec1f931102638da68
Summary:
Migrate away from using the "hg branch" command to "hg commit --extra branch="
instead. In the future the branch namespace would be removed, so we create
local tags instead.
`test-revset2` was split from `test-revset` and has the same header. Do the
same change to it. Besides, de-duplicate tests about the `tag()` revset.
Reviewed By: ikostia
Differential Revision: D14058435
fbshipit-source-id: b59fadc43939d85d14bbf9f81227c523b65557a0
Summary:
"st_size" returned by watchman is sometimes 0. We suspect it's incorrect.
It could be the file on disk is actually empty, or the kernel made it
look empty to watchman in the past, or there is something wrong with
serialization, etc. Let's first add a real os.stat to test if the file
is really empty.
Reviewed By: phillco
Differential Revision: D14090851
fbshipit-source-id: bd75cf86bfd117285658ad0dd7fb28d0e8b5585c
Summary: Otherwise mergedrivers become impossible to debug if they don't import correctly.
Reviewed By: ikostia
Differential Revision: D14071200
fbshipit-source-id: 3e677089e6b008d892158290c392e1f9d68a67ae
Summary:
During repack, the repacked files are deleted without any verification. Since
Adam saw some data loss, it's possible that somehow repack didn't fully repack
a packfile but it was deleted. Let's verify that the entire packfile was
repacked before deleting it.
Since repack is mostly a background operation, we don't have a way to notify
the user, but we can log the error to a scuba table to analyse further.
Reviewed By: DurhamG
Differential Revision: D14069766
fbshipit-source-id: 4358a87deeb9732eec1afdfb742e8d81db41cd87
Summary:
Removing files on Windows is hard. It can fail for many reasons, many of which
involves another process having the file opened in some way. One way to solve
this problem is that renaming the file isn't as restrictive as removing it.
Since hg repack will attempt removing any temporary files it will also try to
remove the packfiles that we failed to remove earlier.
Reviewed By: DurhamG
Differential Revision: D14030445
fbshipit-source-id: 1f3799e021c2e0451943a1d5bd4cd25ed608ffb6
Summary:
Preloading all the pack files on initialization ties the lifetime of the
packfiles to the repo. For normal operations, this is fine, as packfiles are
mostly read. During a repack however, we need to be able to remove them, and
while having an open file handle allows deletion on unix OSes, it prevents it
on Windows.
The Rust repack now succeeds on Windows.
Reviewed By: DurhamG
Differential Revision: D14013786
fbshipit-source-id: 99279d4af67a0dfe8679159e9409186f56a09296
Summary:
Since september (D9840431) rcutil.rccomponents is no longer source of truth
about which config files we load. In fact we use it only in the
`hg config --debug`. This leads to situations where the debug command mentions
loading a specific config file and then ignores it completely which is very
confusing.
Let's remove it.
Reviewed By: suitingtseng
Differential Revision: D14083220
fbshipit-source-id: 362fd9bf574e24639f99a1203206184da42d1e24
Summary:
When using query strings or fragments in an URL we should treat repository paths
as the same repo. This allows servers to use query strings for metadata, without
treating the urls as different servers. By introducing normalization in our grouping,
we will consider the normalized result to be the qualifier for what is the same repo,
rather then the full absolute path - which includes query strings and fragments.
Reviewed By: DurhamG
Differential Revision: D14051479
fbshipit-source-id: c82fe041467f6bd6af210688c0be873e2da93781
Summary: We were reading the `edenapi.url` config item without explicitly setting it up with the `configitem()` function. Not sure what negative impact this would have, but it's probably a good idea to have the explicit call in place.
Reviewed By: quark-zju
Differential Revision: D14075080
fbshipit-source-id: bb4e25de273341768f850f1d5aab6ac21e7f2fc5
Summary: Now that the `edenapi` module in bindings is always available for all platforms, we no longer need a try block around the import.
Reviewed By: quark-zju
Differential Revision: D14075082
fbshipit-source-id: e3f45e67ef4572e58f85875af12390ea5d697d43
Summary: Move the edenapi Python bindings into the common `bindings` crate.
Reviewed By: quark-zju
Differential Revision: D13963179
fbshipit-source-id: 76dead82af992615a9e452ee6fbb9f66639c822c
Summary:
Switch from using OpenSSL (via `native-tls`) to [Rustls](https://github.com/ctz/rustls), a pure-Rust TLS implementation based on the `ring` crypto crate.
Unlike `native-tls`, Rustls supports ALPN, which means it can be used along with Hyper to perform HTTP/2 requests over TLS. (OpenSSL also supports ALPN, but older versions of Windows' `schannel` library do not, and as such `native-tls` doesn't support ALPN either regardless of platform.)
Rustls also builds on Windows without any special configuration, sidestepping the issues we've been having with OpenSSL in the Windows build.
Reviewed By: quark-zju
Differential Revision: D14070084
fbshipit-source-id: 25268c58a88177f4708370696d326b4c0bdc89a0
Summary:
Add a new `init` module to the `bindings` crate. This is intended as a place to put global Rust initialization code for Mercurial's Rust extensions. Ordinarily, such code would go at the start of `main()`, but since `hg` doesn't have a Rust main, putting initialization here at least guarantees that it will happen before any of the Rust extension code runs.
Right now, the only thing initialized in the new module is `env_logger`.
Reviewed By: quark-zju
Differential Revision: D14072560
fbshipit-source-id: 0f2770d0a3a6e9c6b7fe68eb62007cc091adad59
Summary:
if server already have everything, the function should return True
otherwise we treat it as fail to backup in pushbackupbundlestacks function
failed to push stack bundle rooted at ...
Reviewed By: quark-zju
Differential Revision: D14048141
fbshipit-source-id: 288291db2f31b0d284c97fecbd61049a58c045de
Summary:
Monotone repos are rare these days. Drop support for it.
This also solves an issue that test-convert-mtn.t uses named branches.
Differential Revision: D14059836
fbshipit-source-id: 1e9d4fe6fdc295393ff67c5e068b230b9ed0c0af
Summary:
The (disabled by default) `transplant` command is similar to `graft` or
`rebase`. It makes sense to removeit.
This diff removes the extension.
Reviewed By: singhsrb
Differential Revision: D14000099
fbshipit-source-id: a03e4925cefa4236bd9d62cfe9d33d140707bd7c
Summary: They will be replaced by successors and predecessors.
Reviewed By: singhsrb
Differential Revision: D14000097
fbshipit-source-id: 07f0c97edd7eb16ad665c00886c00212d516ae72
Summary:
`test-patchbomb.t` uses named branches and does not look simple to fix.
We don't use email patches internally, and it's not hard to write scripts
around `hg export`. Therefore drop the extension and its tests.
Reviewed By: singhsrb
Differential Revision: D13978577
fbshipit-source-id: 19867ae68c19c996bfce064eb2234705939db9ea
Summary: Named branches are going away. Remove the logic around it.
Reviewed By: phillco
Differential Revision: D13978575
fbshipit-source-id: d6e28d7cadffa612f74a2afc12800829d6113dfa
Summary:
The refresh function is intended to load on-disk packfiles that aren't yet
present in the cache, but it had several issues that resulted in unexpected
behavior. Mainly, the new files would be added in the wrong order in the cache,
and runonpacks would not update the last used packs.
Differential Revision: D14013785
fbshipit-source-id: da4151b859441f0ba51863f3907852922c255f82
Summary:
Previously, the logic for chosing between the rust repack, and the python one
was deeply integrated with the python repack. This resulted in work being
performed unnecessarily.
The drawbacks of this method is that in the case of fallback to Python, the
repacklock will be dropped and re-obtained, as well as the prerepack hook will
be ran twice. Since the python code will soon be entirely replaced this is
probably not important.
Differential Revision: D14013790
fbshipit-source-id: 7c754db3fb984cc9e6b7df1a4e32b72fa4a1531b
Summary: Most of the code is identical between these 2 functions, so let's merge them.
Differential Revision: D14013787
fbshipit-source-id: f0469e05fcf723db02d9f4c28097e61f958284a5
Summary:
This should auto recover from a state where watchman returns too many files,
which makes "hg status" slow.
The default config value 200 is not chosen very scientifically, but it should
cover the 6s status case where there are 90k changed files reported.
Differential Revision: D14036494
fbshipit-source-id: 368f53e99e2e54343a6ac9145cbb86e4ac65e4ac
Summary:
To connect to the memcache client it used a path from a loose files data store (`remotefilelogcontentstore._path`), which is a path to the hg cache, but pack files store doesn't have the same field.
I replaced the path with cache path from shallowutil.
Reviewed By: singhsrb, liubov-dmitrieva
Differential Revision: D14020122
fbshipit-source-id: 29bc8a01fc42d43eddc6ae67cf4e41fc552e117c
Summary:
This back ports Yuya's patch from upstream:
changeset: 443029011990c75c533ec0454fd0f5c1060d4690
branch: stable
user: Yuya Nishihara <yuya@tcha.org>
date: Sun, 01 Jul 2018 06:36:53 -0800
encoding: alias cp65001 to utf-8 on Windows
As far as I can tell, cp65001 is the Windows name for UTF-8. I don't know
how different it is from the UTF-8, but Python 3 appears to have introduced
new codec for cp65001, so the alias is enabled only for Python 2.
https://bugs.python.org/issue13216
This patch is untested, but hopefully fixes the following issue.
https://bitbucket.org/tortoisehg/thg/issues/5127/
More context:
Windows 10 Build 17035 (November 2017) introduced a
"Use Unicode UTF-8 for worldwide language support" checkbox, which makes the
"A" flavored APIs use UTF-8 instead of some non-UTF-8 local encoding. With
that checkbox ticked, hg would crash in cmd.exe complaining
"LookupError: unknown encoding: cp65001"
{F150484107}
Reviewed By: phillco
Differential Revision: D14017361
fbshipit-source-id: a648b2a9bc341d796532c6fd2730c083402b3736
Summary:
The super python builtin expects a type as an argument, while a string was
passed, this was causing trouble for some users.
Reviewed By: DurhamG
Differential Revision: D14012716
fbshipit-source-id: 6714eb20745428818721ec3df85588a7b4c7ebb7
Summary:
It would be massy if there are 2 chg servers running: one for fb hg, one for
upstream hg, and they share a same socket path.
Change socket path and the commandserver name so fb-hg chg can only talk to
fb-hg chg servers.
Reviewed By: markbt
Differential Revision: D13869319
fbshipit-source-id: f9d42af9bdfc542207f23c536b478fd5ef8d02a0
Summary: My earlier diff broke master, this is the fix.
Reviewed By: markbt
Differential Revision: D14005894
fbshipit-source-id: 6ce43913fb501791592512f2f12cd67ef27f0457
Summary:
Currently if default path and infinitepush both point to the remote path
from the repo, the activepath function returns 'infinitepush' rather than
'default' because of the logic I have fixed.
It makes your smartlog looks like.
```
o 037dddb5 62 minutes ago infinitepush/fbobjc/stable remote/fbobjc/stable
.
o 66e0e00c Today at 05:58 infinitepush/fbandroid/stable
remote/fbandroid/stable
.
@ 28c3b5b0 Today at 05:40 mlesyk D14003881 T27597699
. [fbar][prod_role] fix region detection part
.
o 407504d6 Today at 04:58 infinitepush/fbsource/stable
remote/fbsource/stable
```
When you pull it updates infinitepush/master, etc instead of remote/master,
etc.
This only happens if they both point to the same url.
This is not desirable for infinitepush.
I also introduced infinitepush-other that we will use for the secondary commit cloud backend.
It would have the same issue if it matches the url of the default path.
Reviewed By: markbt
Differential Revision: D14005639
fbshipit-source-id: f3ea8098e1c418e2efd43b5430bb4252f6ad6b2c
Summary:
this fixes `hg-new-workdir` by matching the function signature
and propagating the arguments.
Reviewed By: singhsrb
Differential Revision: D13992348
fbshipit-source-id: 582025affe2e0511f239e82bc880b2fecea84ac7
Summary:
logexchange was a subset of remotenames, added by:
changeset: 5a62910948d2d4bac5defe305d0ddb22f0fda549 D1547
user: Pulkit Goyal <7895pulkit@gmail.com>
date: Wed, 04 Oct 2017 10:32:02 -0800
summary: remotenames: move function to pull remotenames from the remoterepo to core
Since we use the full remotenames instead, remove logexchange and its test,
which depends on named branch.
Differential Revision: D13954458
fbshipit-source-id: d565c131100ef90f3cf69e9051643ac8e5846f0d
Summary:
If the history portion of a loose file was corrupt, users would often
get errors like "ValueError: substring not found" when trying to split on the
null separators. Let's catch that and handle it where appropriate. For cache
stores, we move it out of the way and return a KeyError, so the union store will
move on to try to fetch it from the server again. For local stores, we return a
ValueError and surface it to the user since their local data may be corrupt.
Reviewed By: quark-zju
Differential Revision: D13976499
fbshipit-source-id: 39e63f19fc752ca0179fbc5b0908a28d46de3a3a
Summary:
simplecache can be configured with memcache backend
this will potentially save us a lot of disk space and allow us to have a shared cache between hg hosts.
getting rid of the caches on disks will save 966GB of disk space per machine.
P60876105
Reviewed By: quark-zju
Differential Revision: D13950137
fbshipit-source-id: 015e971f1bfed334edb4fe2381c2c6336b84b161
Summary:
reuse the connection for other requests
I made it global
Reviewed By: quark-zju
Differential Revision: D13964705
fbshipit-source-id: ed759e36f15b91a838dc2c20b58384ae54ad5d48
Summary:
simplecache extension support both memcache and local disk as a cache store.
we are going to use memcache on hg servers
getting rid of the caches on disks will save 966GB of disk space per machine.
P60876105
this will also allow hg machines to share the same cached data
Reviewed By: DurhamG
Differential Revision: D13927163
fbshipit-source-id: 3155c61da4a82dd7a790a8ccf8dbc701fe3957e1
Summary:
`hg status` should show slashes in filepaths standard for the diff format. Now it shows windows' backslash.
More details:
https://fb.workplace.com/groups/scm/permalink/1933867456662865/
Reviewed By: quark-zju
Differential Revision: D13944478
fbshipit-source-id: d34d2c6bc0b9966e1806e36668c21e8abfc83a02
Summary:
It looks like D13924759 accidentally replace `prog.value` with `prog` in one
location, which is causing crashes of the form:
TypeError: unsupported operand type(s) for +=: 'nullbar' and 'int'
Reviewed By: quark-zju
Differential Revision: D13984835
fbshipit-source-id: cbf1a52d8048b623beac6048224767f0abb7f5c0
Summary:
On linux, ask /proc/self/mountinfo for the block device information, and then
ask udev for the filesystem type. It works for block devices, and is what
util-linux tools like findmnt, lsblk do.
This detects ntfs-3g mounts of block devices as `fuse.ntfs`.
Reviewed By: DurhamG
Differential Revision: D13982877
fbshipit-source-id: eacea6a949b54d9a6483f2c55c437718eb89c243
Summary:
The constant is not exposed in linux/magic.h, perhaps because fuse is an
optional feature. It's unlikely to be changed. Define it so fuse fstype
can be detected.
Reviewed By: DurhamG
Differential Revision: D13969640
fbshipit-source-id: 5084bdacf193261a187a74546a82995f2dd2d7c6
Summary:
This is breaking the tests. It has existed for years, so not sure why
the recent reformatting broke it now.
Reviewed By: quark-zju
Differential Revision: D13976462
fbshipit-source-id: 08f3a9f08ec54449b4554017ea9806f3ee51afcb
Summary: We don't really need to check misses here as neither of request fucnctions return misses after fetching from the server. They raise exception if there are any misses.
Differential Revision: D13908252
fbshipit-source-id: 80b6e4fb79b419b5e6d3b343b3bd46d3f8dc7d3c
Summary: Implemented logic for fetching data from the scmmemcache and the server in pack files format. By default `prefetch` will still use loose files, unless the `remotefilelog.fetchpacks` config flag will be set to true.
Differential Revision: D13905733
fbshipit-source-id: cc559f5bc600067199a91efe6cf44b1bde6bf69b
Summary: New get/set protocols for pack files. Also removed hitcount from `receive` method, because it is not used.
Differential Revision: D13924759
fbshipit-source-id: 1382b6f2bc865a19125f3ccd17a779cd16f08b50
Summary: To add the data fetched from the server to the scmmemcache, we need to know pasths to the packfiles created.
Differential Revision: D13905734
fbshipit-source-id: 7c470b560e32e63e7df4f960be92bc9b6da9de3d
Summary: Moved fetching loose files from the server to the separate function
Reviewed By: DurhamG
Differential Revision: D13855807
fbshipit-source-id: 39821644abcdc7375ba1b9ef9519fd2083eb0a71
Summary: Move scmmemcache set/get commands protocol to the cache wrapper to isolate it from the main logic
Reviewed By: DurhamG
Differential Revision: D13855809
fbshipit-source-id: 95b70f7f9d10d97beff107f9f6e5e4f5a8a307c1
Summary: List of missed files is converted to a set many times. So changing the type to a set.
Differential Revision: D13924760
fbshipit-source-id: 65fe52ff6113724d8f7f7a8d5b3c1501f245eedc
Summary: Move scmmemcache receive logic to the cache wrapper to isolate it from the main prefetch logic
Reviewed By: DurhamG
Differential Revision: D13855808
fbshipit-source-id: d0da84857797e3e155507fcf284b39a146677e54
Summary:
In some cases we are mounting ntfs in a linux host using ntfs.3g and we
want to prevent it from writing symlinks that can't be read from Windows.
Some background, ntfs.3g supports symlinks, but it produces IntxLNK which cause
problems when mounted into a real Windows machine. We don't support symlinks on
Windows anyway, so we want to just disable this. Unfortunately it's difficult to
detect if we're on a ntfs.3g fuse mount, so instead we're just going to have the
machine set up the repo with this extension enabled.
Differential Revision: D13964546
fbshipit-source-id: 7545538e51be4c986a50161165d7cc2e64280555
Summary: In Rust it is typically more idiomatic to have a static method on a struct to produce a builder, since this means the builder doesn't need to be explicitly imported to construct a new instance of the struct.
Reviewed By: DurhamG
Differential Revision: D13925323
fbshipit-source-id: c06d5d42ba941dbbb2c619f9470e79fa23f35f68
Summary: Rename Mononoke API to Eden API, per war room discussion.
Reviewed By: quark-zju
Differential Revision: D13908195
fbshipit-source-id: 94a2fe93f8a89d0c5e9b6a24939cc4760cfaade0
Summary:
The default value is specified in configitems.py. Remove the `default`
argument.
Reviewed By: phillco
Differential Revision: D13909730
fbshipit-source-id: ca45c63be7d48e04e71a6f5b3b682f625a3030f6
Summary:
This reverts the upstream changeset:
changeset: 66ecde8a704d9fa92bd99720923524c4bd3f315b
user: Boris Feld <boris.feld@octobus.net>
date: Wed, 15 Nov 2017 18:52:42 -0800
summary: server: introduce a 'experimental.single-head-per-branch' option
Named branches are going to be removed. Remove experimental dependent feature.
Reviewed By: phillco
Differential Revision: D13909727
fbshipit-source-id: 3baf101e5a038b58bc01506826b631a2a58a7782
Summary:
Lots of places use it to print some thing like "fetched x files in y seconds (1KB/s)".
In tests, these have to be manually replaced to "* (glob)" patterns.
This diff makes test output static without using globs. So `run-tests.py -i` would just fix things.
Reviewed By: phillco
Differential Revision: D13954456
fbshipit-source-id: d139c7c5e268d3bab562c266335d740986780924
Summary:
We're seeing issues where a file show up as "M" while it isn't expected so. When
running "hg status" again, the file disappeared. Existing logs suggest that
watchman does not think anything has changed between the two "status" commands.
Therefore add more logs of the internals to try to get more insights.
Differential Revision: D13954209
fbshipit-source-id: 0b89a6da5e8702e707440ef2e1823eb4968c10be
Summary:
pushrebase has logic that loops through all the files and builds a list
of what has changed. Unfortunately, Mercurial has some optimizations where if
the manifest isn't loaded already, it tries to only load the manifest delta, and
checks if this might be a good idea by first checking if the file is in the
commit metadata file list. The commit metadata is a list, which makes it a O(n)
scan to check containment. Since we do this for every file, it becomes O(n^2).
To avoid this, let's just make sure the actual manifest is loaded.
Once every repo is a tree repo, we can get rid of the manifest delta
optimization and get rid of the need to prefetch here.
Reviewed By: singhsrb
Differential Revision: D13944829
fbshipit-source-id: c0f33ca650b7956a1f39e961c94678a6f7f380b6
Summary:
During pushrebase there is code that says 'for each file, do some work
that involves listing the changed files'. Because we reparse the list each time,
this becomes n^2 behavior. Let's cache the list to avoid this.
Reviewed By: singhsrb
Differential Revision: D13944830
fbshipit-source-id: 4fd0dbf26b6cee9d0c01879cafb9837262f7afa0
Summary: Extend the option to disable `incoming` to also disable `outgoing`.
Reviewed By: markbt
Differential Revision: D13877971
fbshipit-source-id: c4f994d536dda5db7c0e9f14656be48fc757c237
Summary:
When doing lfs prefetch, pass the filenames that correspond to each oid to lfs
so that it can include them in the progress bar.
Reviewed By: DurhamG
Differential Revision: D13716078
fbshipit-source-id: c91eca6b342996c0c3d1ee43f3b1e98db4ed64d4
Summary:
If the repository was mounted into a different directory than the
source of truth, our watchman integration would not work. To fix this, Mercurial
now applies the same canonicalization logic to the path before passing it to
watchman. This logic is stolen verbatim from watchmans integration test path
utility.
This also requires updating our copy of pywatchman to match the upstream
pywatchman.
Reviewed By: quark-zju
Differential Revision: D13886207
fbshipit-source-id: ea72166edf2eb30f2bee992d9aa936e2e9729e63
Summary: In repositories which change quickly, the incoming and outgoing commands aren't useful. Add an option to disable these commands.
Reviewed By: markbt
Differential Revision: D13859604
fbshipit-source-id: c4784b7eafbe3c4a6cf03686f1431204089b0155
Summary:
The simplecache test attempts to connect to a memcache instance on localhost.
This means the test fails if a memcache instance is not available.
Remove the use of memcache in the test. We still test local files, and assume
that memcache itself works.
Also corrupt the correct key when testing corruption handling, and document the
config options.
Reviewed By: quark-zju
Differential Revision: D13818471
fbshipit-source-id: 6ca7f575aa813f95773144be5337796f029ffd90
Summary:
Move all Rust bindings to a single python extension, `bindings`. This should
improve compilation time and make things simpler.
Reviewed By: quark-zju
Differential Revision: D13923866
fbshipit-source-id: 560592b5a6c0c4f1b836c755ef123666a1059164
Summary:
We will probably not be supporting secret phase in the future anyway
so lets get rid of it.
Reviewed By: quark-zju
Differential Revision: D13928907
fbshipit-source-id: 7edcd84d2ee3c37124dcf9d3ba92c0610d074157
Summary:
Add an option that let's us specify the date of each commit in a pushrebase.
It should be specified via a config option, which accepts a name of a file in
which json dictionary should be stored. Key of this dictionary is a commit hash, value
is a date.
Reviewed By: ikostia
Differential Revision: D13803174
fbshipit-source-id: 6271c18c61399b89c92dce7a4fe63c8fae8dae7c
Summary: Previously, `hg debuggetfile` would allow fetching a single file from the API server. This diff updates the command to `hg debuggetfiles`, which accepts a list of filenode/path pairs on stdin, and fetches all the files concurrently, writing the contents to a single datapack.
Reviewed By: DurhamG
Differential Revision: D13893894
fbshipit-source-id: 36fc54f1870273ab4b447de5d615b3fb6aaa3378
Summary: Expose `MononokeClient`'s multi-get functionality through the Python bindings using a builder-style API.
Differential Revision: D13893748
fbshipit-source-id: 312e7449eb8300432b1f1de28e0c6fc3e47d1d53
Summary: Give MononokeClient the ability to fetch multiple files concurrently. Right now this functionality is not exposed via the Python bindings, so as far as the Mercurial Python code is concerned, nothing has changed. The multi-get functionality will be used later in the stack.
Differential Revision: D13893575
fbshipit-source-id: c9e514fbeb41bbb37f52f6df3920eb01a66df293
Summary: Add a new `get_file()` method to `MononokeClient` that fetches Mercurial file content from the API server and writes it to a datapack in the cache. This functionality is exposed via the new `hg debuggetfile` debug command, which takes a filenode and file path and fetches the corresponding file.
Differential Revision: D13889829
fbshipit-source-id: 2b68bf114ee72d641de7a1043cca1975e34cf4e6
Summary: Rename this debug command to make it obvious what it does (i.e., perform a health check) and make it print more useful output (i.e., the name of the server for which it performs the health check).
Reviewed By: DurhamG
Differential Revision: D13890867
fbshipit-source-id: 8fc96bcc06d04611a308ecc0b870049936f1cb29
Summary:
D13853115 adds `edenscm/` to `sys.path` and code still uses `import mercurial`.
That has nasty problems if both `import mercurial` and
`import edenscm.mercurial` are used, because Python would think `mercurial.foo`
and `edenscm.mercurial.foo` are different modules so code like
`try: ... except mercurial.error.Foo: ...`, or `isinstance(x, mercurial.foo.Bar)`
would fail to handle the `edenscm.mercurial` version. There are also some
module-level states (ex. `extensions._extensions`) that would cause trouble if
they have multiple versions in a single process.
Change imports to use the `edenscm` so ideally the `mercurial` is no longer
imported at all. Add checks in extensions.py to catch unexpected extensions
importing modules from the old (wrong) locations when running tests.
Reviewed By: phillco
Differential Revision: D13868981
fbshipit-source-id: f4e2513766957fd81d85407994f7521a08e4de48
Summary:
A while back, as a startup optimization, we started to hard code
the path to the watchman socket in the mercurial config. This meant
that we could avoid running a subprocess for every command.
Today, a chef configuration change was rolled out that caused watchman
to be stopped on all of the corpnet macs with the result being that
all mercurial commands are slow because it never uses and never restarts
watchman.
This diff adds some fallback logic to allow spawning the server in
the case that we have a connection error.
Reviewed By: DurhamG
Differential Revision: D13855696
fbshipit-source-id: 44c0ad36b6e80520912e42bcea9e04f4f165b588
Summary:
Follow up to D13853115 to fix test failures:
- Update an `__import__` statement in dispatch.py to use the new
`edenscm.mercurial` package path.
- Make sure the `__version__.py` file is generated at the correct location by
the buck build files.
Reviewed By: quark-zju
Differential Revision: D13858016
fbshipit-source-id: 54ac4a02274cf921397932e9625f422c5f427623
Summary:
Move top-level Python packages `mercurial`, `hgext` and `hgdemandimport` to
a new top-level package `edenscm`. This allows the Python packages provided by
the upstream Mercurial to be installed side-by-side.
To maintain compatibility, `edenscm/` gets added to `sys.path` in
`mercurial/__init__.py`.
Reviewed By: phillco, ikostia
Differential Revision: D13853115
fbshipit-source-id: b296b0673dc54c61ef6a591ebc687057ff53b22e
