This commit introduces the infrastructure for the core.fsync
configuration knob. The repository components we want to sync
are identified by flags so that we can turn on or off syncing
for specific components.
If core.fsyncObjectFiles is set and the core.fsync configuration
also includes FSYNC_COMPONENT_LOOSE_OBJECT, we will fsync any
loose objects. This picks the strictest data integrity behavior
if core.fsync and core.fsyncObjectFiles are set to conflicting values.
This change introduces the currently unused fsync_component
helper, which will be used by a later patch that adds fsyncing to
the refs backend.
Actual configuration and documentation of the fsync components
list are in other patches in the series to separate review of
the underlying mechanism from the policy of how it's configured.
Helped-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Neeraj Singh <neerajsi@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The order in which various files that make up a single (conceptual)
packfile has been reevaluated and straightened up. This matters in
correctness, as an incomplete set of files must not be shown to a
running Git.
* tb/pack-finalize-ordering:
pack-objects: rename .idx files into place after .bitmap files
pack-write: split up finish_tmp_packfile() function
builtin/index-pack.c: move `.idx` files into place last
index-pack: refactor renaming in final()
builtin/repack.c: move `.idx` files into place last
pack-write.c: rename `.idx` files after `*.rev`
pack-write: refactor renaming in finish_tmp_packfile()
bulk-checkin.c: store checksum directly
pack.h: line-wrap the definition of finish_tmp_packfile()
The code that optionally creates the *.rev reverse index file has
been optimized to avoid needless computation when it is not writing
the file out.
* ab/reverse-midx-optim:
pack-write: skip *.rev work when not writing *.rev
Split up the finish_tmp_packfile() function and use the split-up version
in pack-objects.c in preparation for moving the step of renaming the
*.idx file later as part of a function change.
Since the only other caller of finish_tmp_packfile() was in
bulk-checkin.c, and it won't be needing a change to its *.idx renaming,
provide a thin wrapper for the old function as a static function in that
file. If other callers end up needing the simpler version it could be
moved back to "pack-write.c" and "pack.h".
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We treat the presence of an `.idx` file as the indicator of whether or
not it's safe to use a packfile. But `finish_tmp_packfile()` (which is
responsible for writing the pack and moving the temporary versions of
all of its auxiliary files into place) is inconsistent about the write
order.
Specifically, it moves the `.rev` file into place after the `.idx`,
leaving open the possibility to open a pack which looks "ready" (because
the `.idx` file exists and is readable) but appears momentarily to not
have a `.rev` file. This causes Git to fall back to generating the
pack's reverse index in memory.
Though racy, this amounts to an unnecessary slow-down at worst, and
doesn't affect the correctness of the resulting reverse index.
Close this race by moving the .rev file into place before moving the
.idx file into place.
This still leaves the issue of `.idx` files being renamed into place
before the auxiliary `.bitmap` file is renamed when in pack-object.c's
write_pack_file() "write_bitmap_index" is true. That race will be
addressed in subsequent commits.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Refactor the renaming in finish_tmp_packfile() into a helper function.
The callers are now expected to pass a "name_buffer" ending in
"pack-OID." instead of the previous "pack-", we then append "pack",
"idx" or "rev" to it.
By doing the strbuf_setlen() in rename_tmp_packfile() we reuse the
buffer and avoid the repeated allocations we'd get if that function had
its own temporary "struct strbuf".
This approach of reusing the buffer does make the last user in
pack-object.c's write_pack_file() slightly awkward, since we needlessly
do a strbuf_setlen() before calling strbuf_release() for consistency. In
subsequent changes we'll move that bitmap writing code around, so let's
not skip the strbuf_setlen() now.
The previous strbuf_reset() idiom originated with 5889271114
(finish_tmp_packfile():use strbuf for pathname construction,
2014-03-03), which in turn was a minimal adjustment of pre-strbuf code
added in 0e990530ae (finish_tmp_packfile(): a helper function,
2011-10-28).
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix a performance regression introduced in a587b5a786 (pack-write.c:
extract 'write_rev_file_order', 2021-03-30) and stop needlessly
allocating the "pack_order" array and sorting it with
"pack_order_cmp()", only to throw that work away when we discover that
we're not writing *.rev files after all.
This redundant work was not present in the original version of this
code added in 8ef50d9958 (pack-write.c: prepare to write 'pack-*.rev'
files, 2021-01-25). There we'd call write_rev_file() from
e.g. finish_tmp_packfile(), but we'd "return NULL" early in
write_rev_file() if not doing a "WRITE_REV" or "WRITE_REV_VERIFY".
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add and apply a semantic patch for using xopen() instead of calling
open(2) and die() or die_errno() explicitly. This makes the error
messages more consistent and shortens the code.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
An on-disk reverse-index to map the in-pack location of an object
back to its object name across multiple packfiles is introduced.
* tb/reverse-midx:
midx.c: improve cache locality in midx_pack_order_cmp()
pack-revindex: write multi-pack reverse indexes
pack-write.c: extract 'write_rev_file_order'
pack-revindex: read multi-pack reverse indexes
Documentation/technical: describe multi-pack reverse indexes
midx: make some functions non-static
midx: keep track of the checksum
midx: don't free midx_name early
midx: allow marking a pack as preferred
t/helper/test-read-midx.c: add '--show-objects'
builtin/multi-pack-index.c: display usage on unrecognized command
builtin/multi-pack-index.c: don't enter bogus cmd_mode
builtin/multi-pack-index.c: split sub-commands
builtin/multi-pack-index.c: define common usage with a macro
builtin/multi-pack-index.c: don't handle 'progress' separately
builtin/multi-pack-index.c: inline 'flags' with options
Existing callers provide the reverse index code with an array of 'struct
pack_idx_entry *'s, which is then sorted by pack order (comparing the
offsets of each object within the pack).
Prepare for the multi-pack index to write a .rev file by providing a way
to write the reverse index without an array of pack_idx_entry (which the
MIDX code does not have).
Instead, callers can invoke 'write_rev_index_positions()', which takes
an array of uint32_t's. The ith entry in this array specifies the ith
object's (in index order) position within the pack (in pack order).
Expose this new function for use in a later patch, and rewrite the
existing write_rev_file() in terms of this new function.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The approach to "fsck" the incoming objects in "index-pack" is
attractive for performance reasons (we have them already in core,
inflated and ready to be inspected), but fundamentally cannot be
applied fully when we receive more than one pack stream, as a tree
object in one pack may refer to a blob object in another pack as
".gitmodules", when we want to inspect blobs that are used as
".gitmodules" file, for example. Teach "index-pack" to emit
objects that must be inspected later and check them in the calling
"fetch-pack" process.
* jt/transfer-fsck-across-packs:
fetch-pack: print and use dangling .gitmodules
fetch-pack: with packfile URIs, use index-pack arg
http-fetch: allow custom index-pack args
http: allow custom index-pack args
Teach index-pack to print dangling .gitmodules links after its "keep" or
"pack" line instead of declaring an error, and teach fetch-pack to check
such lines printed.
This allows the tree side of the .gitmodules link to be in one packfile
and the blob side to be in another without failing the fsck check,
because it is now fetch-pack which checks such objects after all
packfiles have been downloaded and indexed (and not index-pack on an
individual packfile, as it is before this commit).
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This patch prepares for callers to be able to write reverse index files
to disk.
It adds the necessary machinery to write a format-compliant .rev file
from within 'write_rev_file()', which is called from
'finish_tmp_packfile()'.
Similar to the process by which the reverse index is computed in memory,
these new paths also have to sort a list of objects by their offsets
within a packfile. These new paths use a qsort() (as opposed to a radix
sort), since our specialized radix sort requires a full revindex_entry
struct per object, which is more memory than we need to allocate.
The qsort is obviously slower, but the theoretical slowdown would
require a repository with a large amount of objects, likely implying
that the time spent in, say, pack-objects during a repack would dominate
the overall runtime.
Signed-off-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
write_promisor_file() already uses xfopen(), so it would die
if the file cannot be opened for writing. To be consistent
with this behavior and not overlook issues, let's also die if
there are errors when we are actually writing to the file.
Suggested-by: Jeff King <peff@peff.net>
Suggested-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Let's replace the 2 different pieces of code that write a
promisor file in 'builtin/repack.c' and 'fetch-pack.c'
with a new function called 'write_promisor_file()' in
'pack-write.c' and 'pack.h'.
This might also help us in the future, if we want to put
back the ref names and associated hashes that were in
the promisor files we are repacking in 'builtin/repack.c'
as suggested by a NEEDSWORK comment just above the code
we are refactoring.
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Call hashwrite_be64() to write a 64-bit value instead of open-coding it
using htonl() and hashwrite(). This shortens the code, gets rid of a
buffer and several magic numbers, and makes the intent clearer.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
hashwrite() already buffers writes, so pass the fanout table entries
individually via hashwrite_be32(), which also does the endianess
conversion for us. This avoids a memory copy, shortens the code and
reduces the number of magic numbers.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Call hashwrite_be32() instead of open-coding it. This shortens the code
a bit and makes it easier to read.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The index-pack documentation explicitly states that the pack
name is derived from the sorted list of object names, but
since commit 1190a1acf8 ("pack-objects: name pack files
after trailer hash") that isn't true anymore.
Be less explicit in the docs as to what the exact output is,
and just say that it's whatever goes into the pack name.
Also update a comment on write_idx_file() since it no longer
modifies the sha1 variable (it's const now anyway), as noted
by Junio.
Fixes: 1190a1acf8 ("pack-objects: name pack files after trailer hash")
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Pack checksums always use the current hash algorithm in use, so switch
from sha1_to_hex to hash_to_hex.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This rounds out the previous three patches, covering the
inequality logic for the "hash" variant of the functions.
As with the previous three, the accompanying code changes
are the mechanical result of applying the coccinelle patch;
see those patches for more discussion.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Using the more restrictive oideq() should, in the long run,
give the compiler more opportunities to optimize these
callsites. For now, this conversion should be a complete
noop with respect to the generated code.
The result is also perhaps a little more readable, as it
avoids the "zero is equal" idiom. Since it's so prevalent in
C, I think seasoned programmers tend not to even notice it
anymore, but it can sometimes make for awkward double
negations (e.g., we can drop a few !!oidcmp() instances
here).
This patch was generated almost entirely by the included
coccinelle patch. This mechanical conversion should be
completely safe, because we check explicitly for cases where
oidcmp() is compared to 0, which is what oideq() is doing
under the hood. Note that we don't have to catch "!oidcmp()"
separately; coccinelle's standard isomorphisms make sure the
two are treated equivalently.
I say "almost" because I did hand-edit the coccinelle output
to fix up a few style violations (it mostly keeps the
original formatting, but sometimes unwraps long lines).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If we want to use a hashfile on the temporary file for a lockfile, then
we need finalize_hashfile() to fully write the trailing hash but also keep
the file descriptor open.
Do this by adding a new CSUM_HASH_IN_STREAM flag along with a functional
change that checks this flag before writing the checksum to the stream.
This differs from previous behavior since it would be written if either
CSUM_CLOSE or CSUM_FSYNC is provided.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The hashclose() method behaves very differently depending on the flags
parameter. In particular, the file descriptor is not always closed.
Perform a simple rename of "hashclose()" to "finalize_hashfile()" in
preparation for functional changes.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Rename struct sha1file to struct hashfile, along with all of its related
functions.
The transformation in this commit was made by global search-and-replace.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Convert various uses of hardcoded 20- and 40-based numbers to use
the_hash_algo, along with direct calls to SHA-1. Adjust the names of
variables to refer to "hash" instead of "sha1".
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When a caller tries to read a particular set of bytes via
read_in_full(), there are three possible outcomes:
1. An error, in which case -1 is returned and errno is
set.
2. A short read, in which fewer bytes are returned and
errno is unspecified (we never saw a read error, so we
may have some random value from whatever syscall failed
last).
3. The full read completed successfully.
Many callers handle cases 1 and 2 together by just checking
the result against the requested size. If their combined
error path looks at errno (e.g., by calling die_errno), they
may report a nonsense value.
Let's fix these sites by having them distinguish between the
two error cases. That avoids the random errno confusion, and
lets us give more detailed error messages.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Convert struct pack_idx_entry to use struct object_id by changing the
definition and applying the following semantic patch, plus the standard
object_id transforms:
@@
struct pack_idx_entry E1;
@@
- E1.sha1
+ E1.oid.hash
@@
struct pack_idx_entry *E1;
@@
- E1->sha1
+ E1->oid.hash
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The odb_mkstemp() function expects the caller to provide a
fixed buffer to write the resulting tempfile name into. But
it creates the template using snprintf without checking the
return value. This means we could silently truncate the
filename.
In practice, it's unlikely that the truncation would end in
the template-pattern that mkstemp needs to open the file. So
we'd probably end up failing either way, unless the path was
specially crafted.
The simplest fix would be to notice the truncation and die.
However, we can observe that most callers immediately
xstrdup() the result anyway. So instead, let's switch to
using a strbuf, which is easier for them (and isn't a big
deal for the other 2 callers, who can just strbuf_release
when they're done with it).
Note that many of the callers used static buffers, but this
was purely to avoid putting a large buffer on the stack. We
never passed the static buffers out of the function, so
there's no complicated memory handling we need to change.
Signed-off-by: Jeff King <peff@peff.net>
The odb_mkstemp function does not return an error; it dies
on failure instead. But many of its callers compare the
resulting descriptor against -1 and die themselves.
Mostly this is just pointless, but it does raise a question
when looking at the callers: if they show the results of the
"template" buffer after a failure, what's in it? The answer
is: it doesn't matter, because it cannot happen.
So let's make that clear by removing the bogus error checks.
In bitmap_writer_finish(), we can drop the error-handling
code entirely. In the other two cases, it's shared with the
open() in another code path; we can just move the
error-check next to that open() call.
And while we're at it, let's flesh out the function's
docstring a bit to make the error behavior clear.
Signed-off-by: Jeff King <peff@peff.net>
The encode_in_pack_object_header() writes a variable-length
header to an output buffer, but it doesn't actually know
long the buffer is. At first glance, this looks like it
might be possible to overflow.
In practice, this is probably impossible. The smallest
buffer we use is 10 bytes, which would hold the header for
an object up to 2^67 bytes. Obviously we're not likely to
see such an object, but we might worry that an object could
lie about its size (causing us to overflow before we realize
it does not actually have that many bytes). But the argument
is passed as a uintmax_t. Even on systems that have __int128
available, uintmax_t is typically restricted to 64-bit by
the ABI.
So it's unlikely that a system exists where this could be
exploited. Still, it's easy enough to use a normal out/len
pair and make sure we don't write too far. That protects the
hypothetical 128-bit system, makes it harder for callers to
accidentally specify a too-small buffer, and makes the
resulting code easier to audit.
Note that the one caller in fast-import tried to catch such
a case, but did so _after_ the call (at which point we'd
have already overflowed!). This check can now go away.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Apply the semantic patch contrib/coccinelle/qsort.cocci to the code
base, replacing calls of qsort(3) with QSORT. The resulting code is
shorter and supports empty arrays with NULL pointers.
Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The point of this function is to drop an entry from the
"packed_git" cache that points to a file we might be
overwriting, because our contents may not be the same (and
hence the only caller was pack-objects as it moved a
temporary packfile into place).
In older versions of git, this could happen because the
names of packfiles were derived from the set of objects they
contained, not the actual bits on disk. But since 1190a1a
(pack-objects: name pack files after trailer hash,
2013-12-05), the name reflects the actual bits on disk, and
any two packfiles with the same name can be used
interchangeably.
Dropping this function not only saves a few lines of code,
it makes the lifetime of "struct packed_git" much easier to
reason about: namely, we now do not ever free these structs.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Get rid of magic string length constants by using skip_prefix() instead
of memcmp() and use xstrfmt() for building a string instead of a
PATH_MAX-sized buffer, snprintf() and xstrdup().
Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The old version fixes a maximum length on the buffer, which could be a problem
if one is not certain of the length of get_object_directory().
Using strbuf can avoid the protential bug.
Helped-by: Michael Haggerty <mhagger@alum.mit.edu>
Helped-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Sun He <sunheehnus@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Borrow the bitmap index into packfiles from JGit to speed up
enumeration of objects involved in a commit range without having to
fully traverse the history.
* jk/pack-bitmap: (26 commits)
ewah: unconditionally ntohll ewah data
ewah: support platforms that require aligned reads
read-cache: use get_be32 instead of hand-rolled ntoh_l
block-sha1: factor out get_be and put_be wrappers
do not discard revindex when re-preparing packfiles
pack-bitmap: implement optional name_hash cache
t/perf: add tests for pack bitmaps
t: add basic bitmap functionality tests
count-objects: recognize .bitmap in garbage-checking
repack: consider bitmaps when performing repacks
repack: handle optional files created by pack-objects
repack: turn exts array into array-of-struct
repack: stop using magic number for ARRAY_SIZE(exts)
pack-objects: implement bitmap writing
rev-list: add bitmap mode to speed up object lists
pack-objects: use bitmaps when packing objects
pack-objects: split add_object_entry
pack-bitmap: add support for bitmap indexes
documentation: add documentation for the bitmap format
ewah: compressed bitmap implementation
...
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 sha1write function returns an int, but it will always be
"0". The failure-prone parts of the function happen in the
"flush" callback, which cannot pass an error back to us. So
we just end up calling die() during the flush.
Let's just drop the return value altogether, as it only
confuses callers into thinking that it might be useful.
Only one call site actually checked the return value. We can
drop that check, since it just led to a die() anyway.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Our current scheme for naming packfiles is to calculate the
sha1 hash of the sorted list of objects contained in the
packfile. This gives us a unique name, so we are reasonably
sure that two packs with the same name will contain the same
objects.
It does not, however, tell us that two such packs have the
exact same bytes. This makes things awkward if we repack the
same set of objects. Due to run-to-run variations, the bytes
may not be identical (e.g., changed zlib or git versions,
different source object reuse due to new packs in the
repository, or even different deltas due to races during a
multi-threaded delta search).
In theory, this could be helpful to a program that cares
that the packfile contains a certain set of objects, but
does not care about the particular representation. In
practice, no part of git makes use of that, and in many
cases it is potentially harmful. For example, if a dumb http
client fetches the .idx file, it must be sure to get the
exact .pack that matches it. Similarly, a partial transfer
of a .pack file cannot be safely resumed, as the actual
bytes may have changed. This could also affect a local
client which opened the .idx and .pack files, closes the
.pack file (due to memory or file descriptor limits), and
then re-opens a changed packfile.
In all of these cases, git can detect the problem, as we
have the sha1 of the bytes themselves in the pack trailer
(which we verify on transfer), and the .idx file references
the trailer from the matching packfile. But it would be
simpler and more efficient to actually get the correct
bytes, rather than noticing the problem and having to
restart the operation.
This patch simply uses the pack trailer sha1 as the pack
name. It should be similarly unique, but covers the exact
representation of the objects. Other parts of git should not
care, as the pack name is returned by pack-objects and is
essentially opaque.
One test needs to be updated, because it actually corrupts a
pack and expects that re-packing the corrupted bytes will
use the same name. It won't anymore, but we can easily just
use the name that pack-objects hands back.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
On Solaris the system headers define the "tmpfile" name, which'll
cause Git compiled with Sun Studio 12 Update 1 to whine about us
redefining the name:
"pack-write.c", line 76: warning: name redefined by pragma redefine_extname declared static: tmpfile (E_PRAGMA_REDEFINE_STATIC)
"sha1_file.c", line 2455: warning: name redefined by pragma redefine_extname declared static: tmpfile (E_PRAGMA_REDEFINE_STATIC)
"fast-import.c", line 858: warning: name redefined by pragma redefine_extname declared static: tmpfile (E_PRAGMA_REDEFINE_STATIC)
"builtin/index-pack.c", line 175: warning: name redefined by pragma redefine_extname declared static: tmpfile (E_PRAGMA_REDEFINE_STATIC)
Just renaming the "tmpfile" variable to "tmp_file" in the relevant
places is the easiest way to fix this.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jc/stream-to-pack:
bulk-checkin: replace fast-import based implementation
csum-file: introduce sha1file_checkpoint
finish_tmp_packfile(): a helper function
create_tmp_packfile(): a helper function
write_pack_header(): a helper function
Conflicts:
pack.h
When receive-pack & fetch-pack are run and store the pack obtained over
the wire to a local repository, they internally run the index-pack command
with the --strict option. Make sure that we reject incoming packfile that
records objects twice to avoid spreading such a damage.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Factor out a small logic out of the private write_pack_file() function
in builtin/pack-objects.c.
This changes the order of finishing multi-pack generation slightly. The
code used to
- adjust shared perm of temporary packfile
- rename temporary packfile to the final name
- update mtime of the packfile under the final name
- adjust shared perm of temporary idxfile
- rename temporary idxfile to the final name
but because the helper does not want to do the mtime thing, the updated
code does that step first and then all the rest.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A pack v2 .idx file usually records offset using 64-bit representation
only when the offset does not fit within 31-bit, but you can handcraft
your .idx file to record smaller offset using 64-bit, storing all zero
in the upper 4-byte. By inspecting the original idx file when running
index-pack --verify, encode such low offsets that do not need to be in
64-bit but are encoded using 64-bit just like the original idx file so
that we can still validate the pack/idx pair by comparing the idx file
recomputed with the original.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
