This is slightly more verbose, but will let us annotate the
extensions with further options in future commits.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We have a static array of extensions, but hardcode the size
of the array in our loops. Let's pull out this magic number,
which will make it easier to change.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This commit extends more the functionality of `pack-objects` by allowing
it to write out a `.bitmap` index next to any written packs, together
with the `.idx` index that currently gets written.
If bitmap writing is enabled for a given repository (either by calling
`pack-objects` with the `--write-bitmap-index` flag or by having
`pack.writebitmaps` set to `true` in the config) and pack-objects is
writing a packfile that would normally be indexed (i.e. not piping to
stdout), we will attempt to write the corresponding bitmap index for the
packfile.
Bitmap index writing happens after the packfile and its index has been
successfully written to disk (`finish_tmp_packfile`). The process is
performed in several steps:
1. `bitmap_writer_set_checksum`: this call stores the partial
checksum for the packfile being written; the checksum will be
written in the resulting bitmap index to verify its integrity
2. `bitmap_writer_build_type_index`: this call uses the array of
`struct object_entry` that has just been sorted when writing out
the actual packfile index to disk to generate 4 type-index bitmaps
(one for each object type).
These bitmaps have their nth bit set if the given object is of
the bitmap's type. E.g. the nth bit of the Commits bitmap will be
1 if the nth object in the packfile index is a commit.
This is a very cheap operation because the bitmap writing code has
access to the metadata stored in the `struct object_entry` array,
and hence the real type for each object in the packfile.
3. `bitmap_writer_reuse_bitmaps`: if there exists an existing bitmap
index for one of the packfiles we're trying to repack, this call
will efficiently rebuild the existing bitmaps so they can be
reused on the new index. All the existing bitmaps will be stored
in a `reuse` hash table, and the commit selection phase will
prioritize these when selecting, as they can be written directly
to the new index without having to perform a revision walk to
fill the bitmap. This can greatly speed up the repack of a
repository that already has bitmaps.
4. `bitmap_writer_select_commits`: if bitmap writing is enabled for
a given `pack-objects` run, the sequence of commits generated
during the Counting Objects phase will be stored in an array.
We then use that array to build up the list of selected commits.
Writing a bitmap in the index for each object in the repository
would be cost-prohibitive, so we use a simple heuristic to pick
the commits that will be indexed with bitmaps.
The current heuristics are a simplified version of JGit's
original implementation. We select a higher density of commits
depending on their age: the 100 most recent commits are always
selected, after that we pick 1 commit of each 100, and the gap
increases as the commits grow older. On top of that, we make sure
that every single branch that has not been merged (all the tips
that would be required from a clone) gets their own bitmap, and
when selecting commits between a gap, we tend to prioritize the
commit with the most parents.
Do note that there is no right/wrong way to perform commit
selection; different selection algorithms will result in
different commits being selected, but there's no such thing as
"missing a commit". The bitmap walker algorithm implemented in
`prepare_bitmap_walk` is able to adapt to missing bitmaps by
performing manual walks that complete the bitmap: the ideal
selection algorithm, however, would select the commits that are
more likely to be used as roots for a walk in the future (e.g.
the tips of each branch, and so on) to ensure a bitmap for them
is always available.
5. `bitmap_writer_build`: this is the computationally expensive part
of bitmap generation. Based on the list of commits that were
selected in the previous step, we perform several incremental
walks to generate the bitmap for each commit.
The walks begin from the oldest commit, and are built up
incrementally for each branch. E.g. consider this dag where A, B,
C, D, E, F are the selected commits, and a, b, c, e are a chunk
of simplified history that will not receive bitmaps.
A---a---B--b--C--c--D
\
E--e--F
We start by building the bitmap for A, using A as the root for a
revision walk and marking all the objects that are reachable
until the walk is over. Once this bitmap is stored, we reuse the
bitmap walker to perform the walk for B, assuming that once we
reach A again, the walk will be terminated because A has already
been SEEN on the previous walk.
This process is repeated for C, and D, but when we try to
generate the bitmaps for E, we can reuse neither the current walk
nor the bitmap we have generated so far.
What we do now is resetting both the walk and clearing the
bitmap, and performing the walk from scratch using E as the
origin. This new walk, however, does not need to be completed.
Once we hit B, we can lookup the bitmap we have already stored
for that commit and OR it with the existing bitmap we've composed
so far, allowing us to limit the walk early.
After all the bitmaps have been generated, another iteration
through the list of commits is performed to find the best XOR
offsets for compression before writing them to disk. Because of
the incremental nature of these bitmaps, XORing one of them with
its predecesor results in a minimal "bitmap delta" most of the
time. We can write this delta to the on-disk bitmap index, and
then re-compose the original bitmaps by XORing them again when
loaded.
This is a phase very similar to pack-object's `find_delta` (using
bitmaps instead of objects, of course), except the heuristics
have been greatly simplified: we only check the 10 bitmaps before
any given one to find best compressing one. This gives good
results in practice, because there is locality in the ordering of
the objects (and therefore bitmaps) in the packfile.
6. `bitmap_writer_finish`: the last step in the process is
serializing to disk all the bitmap data that has been generated
in the two previous steps.
The bitmap is written to a tmp file and then moved atomically to
its final destination, using the same process as
`pack-write.c:write_idx_file`.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The bitmap reachability index used to speed up the counting objects
phase during `pack-objects` can also be used to optimize a normal
rev-list if the only thing required are the SHA1s of the objects during
the list (i.e., not the path names at which trees and blobs were found).
Calling `git rev-list --objects --use-bitmap-index [committish]` will
perform an object iteration based on a bitmap result instead of actually
walking the object graph.
These are some example timings for `torvalds/linux` (warm cache,
best-of-five):
$ time git rev-list --objects master > /dev/null
real 0m34.191s
user 0m33.904s
sys 0m0.268s
$ time git rev-list --objects --use-bitmap-index master > /dev/null
real 0m1.041s
user 0m0.976s
sys 0m0.064s
Likewise, using `git rev-list --count --use-bitmap-index` will speed up
the counting operation by building the resulting bitmap and performing a
fast popcount (number of bits set on the bitmap) on the result.
Here are some sample timings of different ways to count commits in
`torvalds/linux`:
$ time git rev-list master | wc -l
399882
real 0m6.524s
user 0m6.060s
sys 0m3.284s
$ time git rev-list --count master
399882
real 0m4.318s
user 0m4.236s
sys 0m0.076s
$ time git rev-list --use-bitmap-index --count master
399882
real 0m0.217s
user 0m0.176s
sys 0m0.040s
This also respects negative refs, so you can use it to count
a slice of history:
$ time git rev-list --count v3.0..master
144843
real 0m1.971s
user 0m1.932s
sys 0m0.036s
$ time git rev-list --use-bitmap-index --count v3.0..master
real 0m0.280s
user 0m0.220s
sys 0m0.056s
Though note that the closer the endpoints, the less it helps. In the
traversal case, we have fewer commits to cross, so we take less time.
But the bitmap time is dominated by generating the pack revindex, which
is constant with respect to the refs given.
Note that you cannot yet get a fast --left-right count of a symmetric
difference (e.g., "--count --left-right master...topic"). The slow part
of that walk actually happens during the merge-base determination when
we parse "master...topic". Even though a count does not actually need to
know the real merge base (it only needs to take the symmetric difference
of the bitmaps), the revision code would require some refactoring to
handle this case.
Additionally, a `--test-bitmap` flag has been added that will perform
the same rev-list manually (i.e. using a normal revwalk) and using
bitmaps, and verify that the results are the same. This can be used to
exercise the bitmap code, and also to verify that the contents of the
.bitmap file are sane.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In this patch, we use the bitmap API to perform the `Counting Objects`
phase in pack-objects, rather than a traditional walk through the object
graph. For a reasonably-packed large repo, the time to fetch and clone
is often dominated by the full-object revision walk during the Counting
Objects phase. Using bitmaps can reduce the CPU time required on the
server (and therefore start sending the actual pack data with less
delay).
For bitmaps to be used, the following must be true:
1. We must be packing to stdout (as a normal `pack-objects` from
`upload-pack` would do).
2. There must be a .bitmap index containing at least one of the
"have" objects that the client is asking for.
3. Bitmaps must be enabled (they are enabled by default, but can be
disabled by setting `pack.usebitmaps` to false, or by using
`--no-use-bitmap-index` on the command-line).
If any of these is not true, we fall back to doing a normal walk of the
object graph.
Here are some sample timings from a full pack of `torvalds/linux` (i.e.
something very similar to what would be generated for a clone of the
repository) that show the speedup produced by various
methods:
[existing graph traversal]
$ time git pack-objects --all --stdout --no-use-bitmap-index \
</dev/null >/dev/null
Counting objects: 3237103, done.
Compressing objects: 100% (508752/508752), done.
Total 3237103 (delta 2699584), reused 3237103 (delta 2699584)
real 0m44.111s
user 0m42.396s
sys 0m3.544s
[bitmaps only, without partial pack reuse; note that
pack reuse is automatic, so timing this required a
patch to disable it]
$ time git pack-objects --all --stdout </dev/null >/dev/null
Counting objects: 3237103, done.
Compressing objects: 100% (508752/508752), done.
Total 3237103 (delta 2699584), reused 3237103 (delta 2699584)
real 0m5.413s
user 0m5.604s
sys 0m1.804s
[bitmaps with pack reuse (what you get with this patch)]
$ time git pack-objects --all --stdout </dev/null >/dev/null
Reusing existing pack: 3237103, done.
Total 3237103 (delta 0), reused 0 (delta 0)
real 0m1.636s
user 0m1.460s
sys 0m0.172s
Signed-off-by: Vicent Marti <tanoku@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This function actually does three things:
1. Check whether we've already added the object to our
packing list.
2. Check whether the object meets our criteria for adding.
3. Actually add the object to our packing list.
It's a little hard to see these three phases, because they
happen linearly in the rather long function. Instead, this
patch breaks them up into three separate helper functions.
The result is a little easier to follow, though it
unfortunately suffers from some optimization
interdependencies between the stages (e.g., during step 3 we
use the packing list index from step 1 and the packfile
information from step 2).
More importantly, though, the various parts can be
composed differently, as they will be in the next patch.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A bitmap index is a `.bitmap` file that can be found inside
`$GIT_DIR/objects/pack/`, next to its corresponding packfile, and
contains precalculated reachability information for selected commits.
The full specification of the format for these bitmap indexes can be found
in `Documentation/technical/bitmap-format.txt`.
For a given commit SHA1, if it happens to be available in the bitmap
index, its bitmap will represent every single object that is reachable
from the commit itself. The nth bit in the bitmap is the nth object in
the packfile; if it's set to 1, the object is reachable.
By using the bitmaps available in the index, this commit implements
several new functions:
- `prepare_bitmap_git`
- `prepare_bitmap_walk`
- `traverse_bitmap_commit_list`
- `reuse_partial_packfile_from_bitmap`
The `prepare_bitmap_walk` function tries to build a bitmap of all the
objects that can be reached from the commit roots of a given `rev_info`
struct by using the following algorithm:
- If all the interesting commits for a revision walk are available in
the index, the resulting reachability bitmap is the bitwise OR of all
the individual bitmaps.
- When the full set of WANTs is not available in the index, we perform a
partial revision walk using the commits that don't have bitmaps as
roots, and limiting the revision walk as soon as we reach a commit that
has a corresponding bitmap. The earlier OR'ed bitmap with all the
indexed commits can now be completed as this walk progresses, so the end
result is the full reachability list.
- For revision walks with a HAVEs set (a set of commits that are deemed
uninteresting), first we perform the same method as for the WANTs, but
using our HAVEs as roots, in order to obtain a full reachability bitmap
of all the uninteresting commits. This bitmap then can be used to:
a) limit the subsequent walk when building the WANTs bitmap
b) finding the final set of interesting commits by performing an
AND-NOT of the WANTs and the HAVEs.
If `prepare_bitmap_walk` runs successfully, the resulting bitmap is
stored and the equivalent of a `traverse_commit_list` call can be
performed by using `traverse_bitmap_commit_list`; the bitmap version
of this call yields the objects straight from the packfile index
(without having to look them up or parse them) and hence is several
orders of magnitude faster.
As an extra optimization, when `prepare_bitmap_walk` succeeds, the
`reuse_partial_packfile_from_bitmap` call can be attempted: it will find
the amount of objects at the beginning of the on-disk packfile that can
be reused as-is, and return an offset into the packfile. The source
packfile can then be loaded and the bytes up to `offset` can be written
directly to the result without having to consider the entires inside the
packfile individually.
If the `prepare_bitmap_walk` call fails (e.g. because no bitmap files
are available), the `rev_info` struct is left untouched, and can be used
to perform a manual rev-walk using `traverse_commit_list`.
Hence, this new set of functions are a generic API that allows to
perform the equivalent of
git rev-list --objects [roots...] [^uninteresting...]
for any set of commits, even if they don't have specific bitmaps
generated for them.
In further patches, we'll use this bitmap traversal optimization to
speed up the `pack-objects` and `rev-list` commands.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is the technical documentation for the JGit-compatible Bitmap v1
on-disk format.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
EWAH is a word-aligned compressed variant of a bitset (i.e. a data
structure that acts as a 0-indexed boolean array for many entries).
It uses a 64-bit run-length encoding (RLE) compression scheme,
trading some compression for better processing speed.
The goal of this word-aligned implementation is not to achieve
the best compression, but rather to improve query processing time.
As it stands right now, this EWAH implementation will always be more
efficient storage-wise than its uncompressed alternative.
EWAH arrays will be used as the on-disk format to store reachability
bitmaps for all objects in a repository while keeping reasonable sizes,
in the same way that JGit does.
This EWAH implementation is a mostly straightforward port of the
original `javaewah` library that JGit currently uses. The library is
self-contained and has been embedded whole (4 files) inside the `ewah`
folder to ease redistribution.
The library is re-licensed under the GPLv2 with the permission of Daniel
Lemire, the original author. The source code for the C version can
be found on GitHub:
https://github.com/vmg/libewok
The original Java implementation can also be found on GitHub:
https://github.com/lemire/javaewah
[jc: stripped debug-only code per Peff's $gmane/239768]
Signed-off-by: Vicent Marti <tanoku@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Helped-by: Ramsay Jones <ramsay@ramsay1.demon.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The POSIX standard doesn't currently define a `ntohll`/`htonll`
function pair to perform network-to-host and host-to-network
swaps of 64-bit data. These 64-bit swaps are necessary for the on-disk
storage of EWAH bitmaps if they are not in native byte order.
Many thanks to Ramsay Jones <ramsay@ramsay1.demon.co.uk> and
Torsten Bögershausen <tboegi@web.de> for cygwin/mingw/msvc
portability fixes.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The `git_open_noatime` helper can be of general interest for other
consumers of git's different on-disk formats.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This commit enables users of `struct rev_info` to peform custom limiting
during a revision walk (i.e. `get_revision`).
If the field `include_check` has been set to a callback, this callback
will be issued once for each commit before it is added to the "pending"
list of the revwalk. If the include check returns 0, the commit will be
marked as added but won't be pushed to the pending list, effectively
limiting the walk.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As the pack-objects system grows beyond the single
pack-objects.c file, more parts (like the soon-to-exist
bitmap code) will need to compute hashes for matching
deltas. Factor out name_hash to make it available to other
files.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The hash table that stores the packing list for a given `pack-objects`
run was tightly coupled to the pack-objects code.
In this commit, we refactor the hash table and the underlying storage
array into a `packing_data` struct. The functionality for accessing and
adding entries to the packing list is hence accessible from other parts
of Git besides the `pack-objects` builtin.
This refactoring is a requirement for further patches in this series
that will require accessing the commit packing list from outside of
`pack-objects`.
The hash table implementation has been minimally altered: we now
use table sizes which are always a power of two, to ensure a uniform
index distribution in the array.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Allow users to efficiently lookup consecutive entries that are expected
to be found on the same revindex by exporting `find_revindex_position`:
this function takes a pointer to revindex itself, instead of looking up
the proper revindex for a given packfile on each call.
Signed-off-by: Vicent Marti <tanoku@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We are passed a "void *" and write it out without ever
touching it; let's indicate that by using "const".
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git ls-files -k" needs to crawl only the part of the working tree
that may overlap the paths in the index to find killed files, but
shared code with the logic to find all the untracked files, which
made it unnecessarily inefficient.
* jc/ls-files-killed-optim:
dir.c::test_one_path(): work around directory_exists_in_index_icase() breakage
t3010: update to demonstrate "ls-files -k" optimization pitfalls
ls-files -k: a directory only can be killed if the index has a non-directory
dir.c: use the cache_* macro to access the current index
"git branch --track" had a minor regression in v1.8.3.2 and later
that made it impossible to base your local work on anything but a
local branch of the upstream repository you are tracking from.
* jh/checkout-auto-tracking:
t3200: fix failure on case-insensitive filesystems
branch.c: Relax unnecessary requirement on upstream's remote ref name
t3200: Add test demonstrating minor regression in 41c21f2
Refer to branch.<name>.remote/merge when documenting --track
t3200: Minor fix when preparing for tracking failure
t2024: Fix &&-chaining and a couple of typos
When there is no sufficient overlap between old and new history
during a "git fetch" into a shallow repository, objects that the
sending side knows the receiving end has were unnecessarily sent.
* nd/fetch-into-shallow:
Add testcase for needless objects during a shallow fetch
list-objects: mark more commits as edges in mark_edges_uninteresting
list-objects: reduce one argument in mark_edges_uninteresting
upload-pack: delegate rev walking in shallow fetch to pack-objects
shallow: add setup_temporary_shallow()
shallow: only add shallow graft points to new shallow file
move setup_alternate_shallow and write_shallow_commits to shallow.c
Cleanups and tweaks for credential handling to work with ancient versions
of the gnome-keyring library that are still in use.
* bc/gnome-keyring:
contrib/git-credential-gnome-keyring.c: support really ancient gnome-keyring
contrib/git-credential-gnome-keyring.c: support ancient gnome-keyring
contrib/git-credential-gnome-keyring.c: report failure to store password
contrib/git-credential-gnome-keyring.c: use glib messaging functions
contrib/git-credential-gnome-keyring.c: use glib memory allocation functions
contrib/git-credential-gnome-keyring.c: use secure memory for reading passwords
contrib/git-credential-gnome-keyring.c: use secure memory functions for passwds
contrib/git-credential-gnome-keyring.c: use gnome helpers in keyring_object()
contrib/git-credential-gnome-keyring.c: set Gnome application name
contrib/git-credential-gnome-keyring.c: ensure buffer is non-empty before accessing
contrib/git-credential-gnome-keyring.c: strlen() returns size_t, not ssize_t
contrib/git-credential-gnome-keyring.c: exit non-zero when called incorrectly
contrib/git-credential-gnome-keyring.c: add static where applicable
contrib/git-credential-gnome-keyring.c: *style* use "if ()" not "if()" etc.
contrib/git-credential-gnome-keyring.c: remove unused die() function
contrib/git-credential-gnome-keyring.c: remove unnecessary pre-declarations
Explain how '.' can be used to refer to the "current repository"
in the documentation.
* po/dot-url:
doc/cli: make "dot repository" an independent bullet point
config doc: update dot-repository notes
doc: command line interface (cli) dot-repository dwimmery
"git cherry-pick" without further options would segfault.
Could use a follow-up to handle '-' after argv[1] better.
* hu/cherry-pick-previous-branch:
cherry-pick: handle "-" after parsing options
Make "git grep" and "git show" pay attention to --textconv when
dealing with blob objects.
* mg/more-textconv:
grep: honor --textconv for the case rev:path
grep: allow to use textconv filters
t7008: demonstrate behavior of grep with textconv
cat-file: do not die on --textconv without textconv filters
show: honor --textconv for blobs
diff_opt: track whether flags have been set explicitly
t4030: demonstrate behavior of show with textconv
Document rules to use GIT_REFLOG_ACTION variable in the scripted
Porcelain. git-rebase--interactive locally violates them, but it
is a leaf user that does not call out to or dot-source other
scripts, so it does not urgently need to be fixed.
* jc/reflog-doc:
setup_reflog_action: document the rules for using GIT_REFLOG_ACTION
Rewrite "git repack" in C.
* sb/repack-in-c:
repack: improve warnings about failure of renaming and removing files
repack: retain the return value of pack-objects
repack: rewrite the shell script in C
Some progress and diagnostic messages from "git clone" were
incorrectly sent to the standard output stream, not to the standard
error stream.
* jk/clone-progress-to-stderr:
clone: always set transport options
clone: treat "checking connectivity" like other progress
clone: send diagnostic messages to stderr
The option to gpg sign a merge commit is available but was not
documented. Use wording from the git-commit(1) manpage.
Signed-off-by: Nicolas Vigier <boklm@mars-attacks.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"format-patch --from=<whom>" forgot to omit unnecessary in-body
from line, i.e. when <whom> is the same as the real author.
* jk/format-patch-from:
format-patch: print in-body "From" only when needed
Clean up the internal of the name-hash mechanism used to work
around case insensitivity on some filesystems to cleanly fix a
long-standing API glitch where the caller of cache_name_exists()
that ask about a directory with a counted string was required to
have '/' at one location past the end of the string.
* es/name-hash-no-trailing-slash-in-dirs:
dir: revert work-around for retired dangerous behavior
name-hash: stop storing trailing '/' on paths in index_state.dir_hash
employ new explicit "exists in index?" API
name-hash: refactor polymorphic index_name_exists()
"git filter-branch" in a repository with many refs blew limit of
command line length.
* lc/filter-branch-too-many-refs:
Allow git-filter-branch to process large repositories with lots of branches.
"git checkout [--detach] <commit>" was listed poorly in the
synopsis section of its documentation.
* jc/checkout-detach-doc:
checkout: update synopsys and documentation on detaching HEAD
Jeff King