The checkout machinery has been taught to perform the actual
write-out of the files in parallel when able.
* mt/parallel-checkout-part-2:
parallel-checkout: add design documentation
parallel-checkout: support progress displaying
parallel-checkout: add configuration options
parallel-checkout: make it truly parallel
unpack-trees: add basic support for parallel checkout
Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Make parallel checkout configurable by introducing two new settings:
checkout.workers and checkout.thresholdForParallelism. The first defines
the number of workers (where one means sequential checkout), and the
second defines the minimum number of entries to attempt parallel
checkout.
To decide the default value for checkout.workers, the parallel version
was benchmarked during three operations in the linux repo, with cold
cache: cloning v5.8, checking out v5.8 from v2.6.15 (checkout I) and
checking out v5.8 from v5.7 (checkout II). The four tables below show
the mean run times and standard deviations for 5 runs in: a local file
system on SSD, a local file system on HDD, a Linux NFS server, and
Amazon EFS (all on Linux). Each parallel checkout test was executed with
the number of workers that brings the best overall results in that
environment.
Local SSD:
Sequential 10 workers Speedup
Clone 8.805 s ± 0.043 s 3.564 s ± 0.041 s 2.47 ± 0.03
Checkout I 9.678 s ± 0.057 s 4.486 s ± 0.050 s 2.16 ± 0.03
Checkout II 5.034 s ± 0.072 s 3.021 s ± 0.038 s 1.67 ± 0.03
Local HDD:
Sequential 10 workers Speedup
Clone 32.288 s ± 0.580 s 30.724 s ± 0.522 s 1.05 ± 0.03
Checkout I 54.172 s ± 7.119 s 54.429 s ± 6.738 s 1.00 ± 0.18
Checkout II 40.465 s ± 2.402 s 38.682 s ± 1.365 s 1.05 ± 0.07
Linux NFS server (v4.1, on EBS, single availability zone):
Sequential 32 workers Speedup
Clone 240.368 s ± 6.347 s 57.349 s ± 0.870 s 4.19 ± 0.13
Checkout I 242.862 s ± 2.215 s 58.700 s ± 0.904 s 4.14 ± 0.07
Checkout II 65.751 s ± 1.577 s 23.820 s ± 0.407 s 2.76 ± 0.08
EFS (v4.1, replicated over multiple availability zones):
Sequential 32 workers Speedup
Clone 922.321 s ± 2.274 s 210.453 s ± 3.412 s 4.38 ± 0.07
Checkout I 1011.300 s ± 7.346 s 297.828 s ± 0.964 s 3.40 ± 0.03
Checkout II 294.104 s ± 1.836 s 126.017 s ± 1.190 s 2.33 ± 0.03
The above benchmarks show that parallel checkout is most effective on
repositories located on an SSD or over a distributed file system. For
local file systems on spinning disks, and/or older machines, the
parallelism does not always bring a good performance. For this reason,
the default value for checkout.workers is one, a.k.a. sequential
checkout.
To decide the default value for checkout.thresholdForParallelism,
another benchmark was executed in the "Local SSD" setup, where parallel
checkout showed to be beneficial. This time, we compared the runtime of
a `git checkout -f`, with and without parallelism, after randomly
removing an increasing number of files from the Linux working tree. The
"sequential fallback" column below corresponds to the executions where
checkout.workers was 10 but checkout.thresholdForParallelism was equal
to the number of to-be-updated files plus one (so that we end up writing
sequentially). Each test case was sampled 15 times, and each sample had
a randomly different set of files removed. Here are the results:
sequential fallback 10 workers speedup
10 files 772.3 ms ± 12.6 ms 769.0 ms ± 13.6 ms 1.00 ± 0.02
20 files 780.5 ms ± 15.8 ms 775.2 ms ± 9.2 ms 1.01 ± 0.02
50 files 806.2 ms ± 13.8 ms 767.4 ms ± 8.5 ms 1.05 ± 0.02
100 files 833.7 ms ± 21.4 ms 750.5 ms ± 16.8 ms 1.11 ± 0.04
200 files 897.6 ms ± 30.9 ms 730.5 ms ± 14.7 ms 1.23 ± 0.05
500 files 1035.4 ms ± 48.0 ms 677.1 ms ± 22.3 ms 1.53 ± 0.09
1000 files 1244.6 ms ± 35.6 ms 654.0 ms ± 38.3 ms 1.90 ± 0.12
2000 files 1488.8 ms ± 53.4 ms 658.8 ms ± 23.8 ms 2.26 ± 0.12
From the above numbers, 100 files seems to be a reasonable default value
for the threshold setting.
Note: Up to 1000 files, we observe a drop in the execution time of the
parallel code with an increase in the number of files. This is a rather
odd behavior, but it was observed in multiple repetitions. Above 1000
files, the execution time increases according to the number of files, as
one would expect.
About the test environments: Local SSD tests were executed on an
i7-7700HQ (4 cores with hyper-threading) running Manjaro Linux. Local
HDD tests were executed on an Intel(R) Xeon(R) E3-1230 (also 4 cores
with hyper-threading), HDD Seagate Barracuda 7200.14 SATA 3.1, running
Debian. NFS and EFS tests were executed on an Amazon EC2 c5n.xlarge
instance, with 4 vCPUs. The Linux NFS server was running on a m6g.large
instance with 2 vCPUSs and a 1 TB EBS GP2 volume. Before each timing,
the linux repository was removed (or checked out back to its previous
state), and `sync && sysctl vm.drop_caches=3` was executed.
Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use multiple worker processes to distribute the queued entries and call
write_pc_item() in parallel for them. The items are distributed
uniformly in contiguous chunks. This minimizes the chances of two
workers writing to the same directory simultaneously, which could affect
performance due to lock contention in the kernel. Work stealing (or any
other format of re-distribution) is not implemented yet.
The protocol between the main process and the workers is quite simple.
They exchange binary messages packed in pkt-line format, and use
PKT-FLUSH to mark the end of input (from both sides). The main process
starts the communication by sending N pkt-lines, each corresponding to
an item that needs to be written. These packets contain all the
necessary information to load, smudge, and write the blob associated
with each item. Then it waits for the worker to send back N pkt-lines
containing the results for each item. The resulting packet must contain:
the identification number of the item that it refers to, the status of
the operation, and the lstat() data gathered after writing the file (iff
the operation was successful).
For now, checkout always uses a hardcoded value of 2 workers, only to
demonstrate that the parallel checkout framework correctly divides and
writes the queued entries. The next patch will add user configurations
and define a more reasonable default, based on tests with the said
settings.
Co-authored-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This new interface allows us to enqueue some of the entries being
checked out to later uncompress them, apply in-process filters, and
write out the files in parallel. For now, the parallel checkout
machinery is enabled by default and there is no user configuration, but
run_parallel_checkout() just writes the queued entries in sequence
(without spawning additional workers). The next patch will actually
implement the parallelism and, later, we will make it configurable.
Note that, to avoid potential data races, not all entries are eligible
for parallel checkout. Also, paths that collide on disk (e.g.
case-sensitive paths in case-insensitive file systems), are detected by
the parallel checkout code and skipped, so that they can be safely
sequentially handled later. The collision detection works like the
following:
- If the collision was at basename (e.g. 'a/b' and 'a/B'), the framework
detects it by looking for EEXIST and EISDIR errors after an
open(O_CREAT | O_EXCL) failure.
- If the collision was at dirname (e.g. 'a/b' and 'A'), it is detected
at the has_dirs_only_path() check, which is done for the leading path
of each item in the parallel checkout queue.
Both verifications rely on the fact that, before enqueueing an entry for
parallel checkout, checkout_entry() makes sure that there is no file at
the entry's path and that its leading components are all real
directories. So, any later change in these conditions indicates that
there was a collision (either between two parallel-eligible entries or
between an eligible and an ineligible one).
After all parallel-eligible entries have been processed, the collided
(and thus, skipped) entries are sequentially fed to checkout_entry()
again. This is similar to the way the current code deals with
collisions, overwriting the previously checked out entries with the
subsequent ones. The only difference is that, since we no longer create
the files in the same order that they appear on index, we are not able
to determine which of the colliding entries will survive on disk (for
the classic code, it is always the last entry).
Co-authored-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Builds on top of the sparse-index infrastructure to mark operations
that are not ready to mark with the sparse index, causing them to
fall back on fully-populated index that they always have worked with.
* ds/sparse-index-protections: (26 commits)
name-hash: use expand_to_path()
sparse-index: expand_to_path()
name-hash: don't add directories to name_hash
revision: ensure full index
resolve-undo: ensure full index
read-cache: ensure full index
pathspec: ensure full index
merge-recursive: ensure full index
entry: ensure full index
dir: ensure full index
update-index: ensure full index
stash: ensure full index
rm: ensure full index
merge-index: ensure full index
ls-files: ensure full index
grep: ensure full index
fsck: ensure full index
difftool: ensure full index
commit: ensure full index
checkout: ensure full index
...
Document changes in -m and --diff-merges=m semantics, as well as new
--diff-merges=on option.
Signed-off-by: Sergey Organov <sorganov@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
New log.diffMerges configuration variable sets the format that
--diff-merges=on will be using. The default is "separate".
t4013: add the following tests for log.diffMerges config:
* Test that wrong values are denied.
* Test that the value of log.diffMerges properly affects both
--diff-merges=on and -m.
t9902: fix completion tests for log.d* to match log.diffMerges.
Added documentation for log.diffMerges.
Signed-off-by: Sergey Organov <sorganov@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Let -m option (and --diff-merges=m) enable the default format instead
of "separate", to be able to tune it with log.diffMerges option.
Signed-off-by: Sergey Organov <sorganov@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Split set_diff_merges() into separate parsing and execution functions,
the former to be reused for parsing of configuration values later in
the patch series.
Signed-off-by: Sergey Organov <sorganov@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Introduce the notion of default diff format for merges, and the option
"on" to select it. The default format is "separate" and can't yet
be changed, so effectively "on" is just a synonym for "separate"
for now. Add corresponding test to t4013.
This is in preparation for introducing log.diffMerges configuration
option that will let --diff-merges=on to be configured to any
supported format.
Signed-off-by: Sergey Organov <sorganov@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The prefetch task in "git maintenance" assumed that "git fetch"
from any remote would fetch all its local branches, which would
fetch too much if the user is interested in only a subset of
branches there.
* ds/maintenance-prefetch-fix:
maintenance: respect remote.*.skipFetchAll
maintenance: use 'git fetch --prefetch'
fetch: add --prefetch option
maintenance: simplify prefetch logic
The dependencies for config-list.h and command-list.h were broken
when the former was split out of the latter, which has been
corrected.
* sg/bugreport-fixes:
Makefile: add missing dependencies of 'config-list.h'
"git push --quiet --set-upstream" was not quiet when setting the
upstream branch configuration, which has been corrected.
* ow/push-quiet-set-upstream:
transport: respect verbosity when setting upstream
When packet_write() fails, we gave an extra error message
unnecessarily, which has been corrected.
* mt/pkt-write-errors:
pkt-line: do not report packet write errors twice
Plug the ort merge backend throughout the rest of the system, and
start testing it as a replacement for the recursive backend.
* en/ort-readiness:
Add testing with merge-ort merge strategy
t6423: mark remaining expected failure under merge-ort as such
Revert "merge-ort: ignore the directory rename split conflict for now"
merge-recursive: add a bunch of FIXME comments documenting known bugs
merge-ort: write $GIT_DIR/AUTO_MERGE whenever we hit a conflict
t: mark several submodule merging tests as fixed under merge-ort
merge-ort: implement CE_SKIP_WORKTREE handling with conflicted entries
t6428: new test for SKIP_WORKTREE handling and conflicts
merge-ort: support subtree shifting
merge-ort: let renormalization change modify/delete into clean delete
merge-ort: have ll_merge() use a special attr_index for renormalization
merge-ort: add a special minimal index just for renormalization
merge-ort: use STABLE_QSORT instead of QSORT where required
Various rename detection optimization to help "ort" merge strategy
backend.
* en/ort-perf-batch-10:
diffcore-rename: determine which relevant_sources are no longer relevant
merge-ort: record the reason that we want a rename for a file
diffcore-rename: add computation of number of unknown renames
diffcore-rename: check if we have enough renames for directories early on
diffcore-rename: only compute dir_rename_count for relevant directories
merge-ort: record the reason that we want a rename for a directory
merge-ort, diffcore-rename: tweak dirs_removed and relevant_source type
diffcore-rename: take advantage of "majority rules" to skip more renames
If a remote has the skipFetchAll setting enabled, then that remote is
not intended for frequent fetching. It makes sense to not fetch that
data during the 'prefetch' maintenance task. Skip that remote in the
iteration without error. The skip_default_update member is initialized
in remote.c:handle_config() as part of initializing the 'struct remote'.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'prefetch' maintenance task previously forced the following refspec
for each remote:
+refs/heads/*:refs/prefetch/<remote>/*
If a user has specified a more strict refspec for the remote, then this
prefetch task downloads more objects than necessary.
The previous change introduced the '--prefetch' option to 'git fetch'
which manipulates the remote's refspec to place all resulting refs into
refs/prefetch/, with further partitioning based on the destinations of
those refspecs.
Update the documentation to be more generic about the destination refs.
Do not mention custom refspecs explicitly, as that does not need to be
highlighted in this documentation. The important part of placing refs in
refs/prefetch/ remains.
Reported-by: Tom Saeger <tom.saeger@oracle.com>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The --prefetch option will be used by the 'prefetch' maintenance task
instead of sending refspecs explicitly across the command-line. The
intention is to modify the refspec to place all results in
refs/prefetch/ instead of anywhere else.
Create helper method filter_prefetch_refspec() to modify a given refspec
to fit the rules expected of the prefetch task:
* Negative refspecs are preserved.
* Refspecs without a destination are removed.
* Refspecs whose source starts with "refs/tags/" are removed.
* Other refspecs are placed within "refs/prefetch/".
Finally, we add the 'force' option to ensure that prefetch refs are
replaced as necessary.
There are some interesting cases that are worth testing.
An earlier version of this change dropped the "i--" from the loop that
deletes a refspec item and shifts the remaining entries down. This
allowed some refspecs to not be modified. The subtle part about the
first --prefetch test is that the "refs/tags/*" refspec appears directly
before the "refs/heads/bogus/*" refspec. Without that "i--", this
ordering would remove the "refs/tags/*" refspec and leave the last one
unmodified, placing the result in "refs/heads/*".
It is possible to have an empty refspec. This is typically the case for
remotes other than the origin, where users want to fetch a specific tag
or branch. To correctly test this case, we need to further remove the
upstream remote for the local branch. Thus, we are testing a refspec
that will be deleted, leaving nothing to fetch.
Helped-by: Tom Saeger <tom.saeger@oracle.com>
Helped-by: Ramsay Jones <ramsay@ramsayjones.plus.com>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Apparently this is not supported with Microsoft's Universal C Runtime.
So let's not actually do that.
Instead, just return success because we _know_ that we expect the `NUL`
device to be present.
Side note: it is possible to turn off the "Null device driver" and
thereby disable `NUL`. Too many things are broken if this driver is
disabled, therefore it is not worth bothering to try to detect its
presence when `access()` is called.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
On write() errors, packet_write() dies with the same error message that
is already printed by its callee, packet_write_gently(). This produces
an unnecessarily verbose and repetitive output:
error: packet write failed
fatal: packet write failed: <strerror() message>
In addition to that, packet_write_gently() does not always fulfill its
caller expectation that errno will be properly set before a non-zero
return. In particular, that is not the case for a "data exceeds max
packet size" error. So, in this case, packet_write() will call
die_errno() and print an strerror(errno) message that might be totally
unrelated to the actual error.
Fix both those issues by turning packet_write() and
packet_write_gently() into wrappers to a common lower level function
that doesn't print the error message, but instead returns it on a buffer
for the caller to die() or error() as appropriate.
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Replace GIT_CONFIG_NOSYSTEM mechanism to decline from reading the
system-wide configuration file with GIT_CONFIG_SYSTEM that lets
users specify from which file to read the system-wide configuration
(setting it to an empty file would essentially be the same as
setting NOSYSTEM), and introduce GIT_CONFIG_GLOBAL to override the
per-user configuration in $HOME/.gitconfig.
* ps/config-global-override:
config: allow overriding of global and system configuration
config: unify code paths to get global config paths
config: rename `git_etc_config()`
Handling of "promisor packs" that allows certain objects to be
missing and lazily retrievable has been optimized (a bit).
* jk/promisor-optim:
revision: avoid parsing with --exclude-promisor-objects
lookup_unknown_object(): take a repository argument
is_promisor_object(): free tree buffer after parsing
Documentation updates, with unrelated comment updates, too.
* ab/usage-error-docs:
api docs: document that BUG() emits a trace2 error event
api docs: document BUG() in api-error-handling.txt
usage.c: don't copy/paste the same comment three times
"git apply" now takes "--3way" and "--cached" at the same time, and
work and record results only in the index.
* jz/apply-3way-cached:
git-apply: allow simultaneous --cached and --3way options
The command line completion (in contrib/) has learned that
CHERRY_PICK_HEAD is a possible pseudo-ref.
* ab/complete-cherry-pick-head:
bash completion: complete CHERRY_PICK_HEAD
"git apply --3way" has always been "to fall back to 3-way merge
only when straight application fails". Swap the order of falling
back so that 3-way is always attempted first (only when the option
is given, of course) and then straight patch application is used as
a fallback when it fails.
* jz/apply-run-3way-first:
git-apply: try threeway first when "--3way" is used
A command such as `git push -qu origin feature` will print "Branch
'feature' set up to track remote branch 'feature' from 'origin'." even
when --quiet is passed. In this case it's because install_branch_config() is
always called with BRANCH_CONFIG_VERBOSE.
struct transport keeps track of the desired verbosity. Fix the above
issue by passing BRANCH_CONFIG_VERBOSE conditionally based on that.
Signed-off-by: Øystein Walle <oystwa@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The same "do not capitalize the first word" rule is applied to both
our patch titles and error messages, but the existing description
was fuzzy in two aspects.
* For error messages, it was not said that this was only about the
first word that begins the sentence.
* For both, it was not clear when a capital letter there was not an
error. We avoid capitalizing the first word when the only reason
you would capitalize it is because it happens to be the first
word in the sentence. If a proper noun, which is usually spelled
in capital letters, happens to come at the beginning of the
sentence, it should be kept in capital letters.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A sparse-index loads the name-hash data for its entries, including the
sparse-directory entries. If a caller asks for a path that is contained
within a sparse-directory entry, we need to expand to a full index and
recalculate the name hash table before returning the result. Insert
calls to expand_to_path() to protect against this case.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some users of the index API have a specific path they are looking for,
but choose to use index_file_exists() to rely on the name-hash hashtable
instead of doing binary search with index_name_pos(). These users only
need to know a yes/no answer, not a position within the cache array.
When the index is sparse, the name-hash hash table does not contain the
full list of paths within sparse directories. It _does_ contain the
directory names for the sparse-directory entries.
Create a helper function, expand_to_path(), for intended use with the
name-hash hashtable functions. The integration with name-hash.c will
follow in a later change.
The solution here is to use ensure_full_index() when we determine that
the requested path is within a sparse directory entry. This will
populate the name-hash hashtable as the index is recomputed from
scratch.
There may be cases where the caller is trying to find an untracked path
that is not in the index but also is not within a sparse directory
entry. We want to minimize the overhead for these requests. If we used
index_name_pos() to find the insertion order of the path, then we could
determine from that position if a sparse-directory exists. (In fact,
just calling index_name_pos() in that case would lead to expanding the
index to a full index.) However, this takes O(log N) time where N is the
number of cache entries.
To keep the performance of this call based mostly on the input string,
use index_file_exists() to look for the ancestors of the path. Using the
heuristic that a sparse directory is likely to have a small number of
parent directories, we start from the bottom and build up. Use a string
buffer to allow mutating the path name to terminate after each slash for
each hashset test.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Sparse directory entries represent a directory that is outside the
sparse-checkout definition. These are not paths to blobs, so should not
be added to the name_hash table. Instead, they should be added to the
directory hashtable when 'ignore_case' is true.
Add a condition to avoid placing sparse directories into the name_hash
hashtable. This avoids filling the table with extra entries that will
never be queried.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before iterating over all index entries, ensure that a sparse index is
expanded to a full index to avoid unexpected behavior. This case could
be integrated later by ensuring that we walk the tree in the
sparse-directory entry, but the current behavior is only expecting
blobs. Save this integration for later when it can be properly tested.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before iterating over all cache entries, ensure that a sparse index is
expanded to a full index to avoid unexpected behavior.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before iterating over all cache entries, ensure that a sparse index is
expanded to a full index to avoid unexpected behavior.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before iterating over all cache entries, ensure that a sparse index is
expanded to a full index to avoid unexpected behavior.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before iterating over all cache entries, ensure that a sparse index is
expanded to a full index to avoid unexpected behavior.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano
Matheus Tavares
Sergey Organov
Derrick Stolee
Johannes Schindelin
Øystein Walle