In a later patch we want to enhance the logic for the branch selection.
Rewrite the current logic to be in C, so we can directly use C when
we enhance the logic.
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There is an optimization used in "git diff $treeA $treeB" to borrow
an already checked-out copy in the working tree when it is known to
be the same as the blob being compared, expecting that open/mmap of
such a file is faster than reading it from the object store, which
involves inflating and applying delta. This however kicked in even
when the checked-out copy needs to go through the convert-to-git
conversion (including the clean filter), which defeats the whole
point of the optimization. The optimization has been disabled when
the conversion is necessary.
* jk/diff-do-not-reuse-wtf-needs-cleaning:
diff: do not reuse worktree files that need "clean" conversion
Code cleanup.
* rs/submodule-config-code-cleanup:
submodule-config: fix test binary crashing when no arguments given
submodule-config: combine early return code into one goto
submodule-config: passing name reference for .gitmodule blobs
submodule-config: use explicit empty string instead of strbuf in config_from()
"git push" and "git clone" learned to give better progress meters
to the end user who is waiting on the terminal.
* jk/push-progress:
receive-pack: send keepalives during quiet periods
receive-pack: turn on connectivity progress
receive-pack: relay connectivity errors to sideband
receive-pack: turn on index-pack resolving progress
index-pack: add flag for showing delta-resolution progress
clone: use a real progress meter for connectivity check
check_connected: add progress flag
check_connected: relay errors to alternate descriptor
check_everything_connected: use a struct with named options
check_everything_connected: convert to argv_array
rev-list: add optional progress reporting
check_everything_connected: always pass --quiet to rev-list
"git fetch" exchanges batched have/ack messages between the sender
and the receiver, initially doubling every time and then falling
back to enlarge the window size linearly. The "smart http"
transport, being an half-duplex protocol, outgrows the preset limit
too quickly and becomes inefficient when interacting with a large
repository. The internal mechanism learned to grow the window size
more aggressively when working with the "smart http" transport.
* jt/fetch-large-handshake-window-on-http:
fetch-pack: grow stateless RPC windows exponentially
"git jump" script (in contrib/) has been updated a bit.
* jk/git-jump:
contrib/git-jump: fix typo in README
contrib/git-jump: add whitespace-checking mode
contrib/git-jump: fix greedy regex when matching hunks
"git status" learned to suggest "merge --abort" during a conflicted
merge, just like it already suggests "rebase --abort" during a
conflicted rebase.
* mm/status-suggest-merge-abort:
status: suggest 'git merge --abort' when appropriate
Users of the parse_options_concat() API function need to allocate
extra slots in advance and fill them with OPT_END() when they want
to decide the set of supported options dynamically, which makes the
code error-prone and hard to read. This has been corrected by tweaking
the API to allocate and return a new copy of "struct option" array.
* jk/parse-options-concat:
parse_options: allocate a new array when concatenating
"git push" learned to accept and pass extra options to the
receiving end so that hooks can read and react to them.
* sb/push-options:
add a test for push options
push: accept push options
receive-pack: implement advertising and receiving push options
push options: {pre,post}-receive hook learns about push options
Dumb http transport on the client side has been optimized.
* ew/http-walker:
list: avoid incompatibility with *BSD sys/queue.h
http-walker: reduce O(n) ops with doubly-linked list
http: avoid disconnecting on 404s for loose objects
http-walker: remove unused parameter from fetch_object
The build procedure for "git persistent-https" helper (in contrib/)
has been updated so that it can be built with more recent versions
of Go.
* pm/build-persistent-https-with-recent-go:
contrib/persistent-https: use Git version for build label
contrib/persistent-https: update ldflags syntax for Go 1.7+
"git merge" in Git v2.9 was taught to forbid merging an unrelated
lines of history by default, but that is exactly the kind of thing
the "--rejoin" mode of "git subtree" (in contrib/) wants to do.
"git subtree" has been taught to use the "--allow-unrelated-histories"
option to override the default.
* da/subtree-2.9-regression:
subtree: fix "git subtree split --rejoin"
t7900-subtree.sh: fix quoting and broken && chains
"git commit --help" said "--no-verify" is only about skipping the
pre-commit hook, and failed to say that it also skipped the
commit-msg hook.
* os/no-verify-skips-commit-msg-too:
commit: describe that --no-verify skips the commit-msg hook in the help text
It's natural to expect %f to be an actual file on disk; help avoid that
mistake.
Signed-off-by: Joey Hess <joeyh@joeyh.name>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Previously, we simply treated hard links as if they were plain files
with size 0, ignoring the link type "1" and hence the link target.
What we should do instead, of course, is to use the link target to get
at the import mark for the contents, even if we cannot recreate the hard
link per se, as Git has no concept of hard links.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This strdup was added as part of 58dbfa2 (blame: accept
multiple -L ranges, 2013-08-06) to be consistent with
parse_opt_string_list(), which appends to the same list.
But as of 7a7a517 (parse_opt_string_list: stop allocating
new strings, 2016-06-13), we should stop using strdup (to
match parse_opt_string_list, and for all the reasons
described in that commit; namely that it does nothing useful
and causes us to leak the memory).
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
On Windows, it is already pretty expensive to try to recreate the stat()
data that Git assumes is cheap to obtain. To make things halfway decent
in performance, we even have to skip emulating the inode and to
determine the number of hard links.
This is not a huge problem, usually, as either the size or the mtime or
the ctime are tell-tale enough to say when a file has changed, and even
if not, those changes are typically made after the index file was
written, triggering a rehashing of the files' contents.
The t4130-apply-criss-cross-rename test case, however, requires the
inode to determine that files of equal size were swapped, as renaming
files does not update their mtime. Every once in a while, t4130 fails
on Windows because of this missing piece.
Equal file sizes are not crucial for the test cases, however. Hence,
generate files with different sizes so that there is some property that
the swapped files can be discovered reliably even on Windows.
Helped-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The API documentation said that the hashmap_entry structure to be
embedded in the caller's structure is to be treated as opaque, which
left the reader wondering if it can safely be discarded when it no
longer is necessary. If the hashmap_entry structure had references
to external resources such as allocated memory or an open file
descriptor, merely free(3)ing the containing structure (when the
caller's structure is on the heap) or letting it go out of scope
(when it is on the stack) would end up leaking the external
resource.
Document that there is no need for hashmap_entry_clear() that
corresponds to hashmap_entry_init() to give the API users a little
bit of peace of mind.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we compute the date to go in author/committer lines of
commits, or tagger lines of tags, we get the current date
once and then cache it for the rest of the program. This is
a good thing in some cases, like "git commit", because it
means we do not racily assign different times to the
author/committer fields of a single commit object.
But as more programs start to make many commits in a single
process (e.g., the recently builtin "git am"), it means that
you'll get long strings of commits with identical committer
timestamps (whereas before, we invoked "git commit" many
times and got true timestamps).
This patch addresses it by letting callers reset the cached
time, which means they'll get a fresh time on their next
call to git_committer_info() or git_author_info(). The first
caller to do so is "git am", which resets the time for each
patch it applies.
It would be nice if we could just do this automatically
before filling in the ident fields of commit and tag
objects. Unfortunately, it's hard to know where a particular
logical operation begins and ends.
For instance, if commit_tree_extended() were to call
reset_ident_date() before getting the committer/author
ident, that doesn't quite work; sometimes the author info is
passed in to us as a parameter, and it may or may not have
come from a previous call to ident_default_date(). So in
those cases, we lose the property that the committer and the
author timestamp always match.
You could similarly put a date-reset at the end of
commit_tree_extended(). That actually works in the current
code base, but it's fragile. It makes the assumption that
after commit_tree_extended() finishes, the caller has no
other operations that would logically want to fall into the
same timestamp.
So instead we provide the tool to easily do the reset, and
let the high-level callers use it to annotate their own
logical operations.
There's no automated test, because it would be inherently
racy (it depends on whether the program takes multiple
seconds to run). But you can see the effect with something
like:
# make a fake 100-patch series
top=$(git rev-parse HEAD)
bottom=$(git rev-list --first-parent -100 HEAD | tail -n 1)
git log --format=email --reverse --first-parent \
--binary -m -p $bottom..$top >patch
# now apply it; this presumably takes multiple seconds
git checkout --detach $bottom
git am <patch
# now count the number of distinct committer times;
# prior to this patch, there would only be one, but
# now we'd typically see several.
git log --format=%ct $bottom.. | sort -u
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Helped-by: Paul Tan <pyokagan@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The branch field will be used in a later patch by `submodule update`.
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Internally we call the underscore version of relative_path, but externally
we present an API with no underscores.
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When depth is given the user may have a reasonable expectation that
any remote operation is using the given depth. Add a test to demonstrate
we still get the desired sha1 even if the depth is too short to
include the actual commit.
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The prior hard coded depth was chosen to be exactly the length from the
recorded gitlink to the tip of the remote, so if you add more commits
to the remote before, this test will not test its intention any more.
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The case statement to check the file mode of a staged file appears
a number of times.
Simplify the test by utilizing a test_mode_in_index helper function.
Signed-off-by: Ingo Brückl <ib@wupperonline.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Depending on the underlying platform a chmod may be a noop. Although it
wouldn't harm the result of the '--chmod=-x' test, there is a more
robust way to make sure the --chmod option works both ways.
Merge the two separate tests for the --chmod option into one, checking
both permissions on the same file.
Signed-off-by: Ingo Brückl <ib@wupperonline.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When an earlier test that has prerequisite is skipped, files
used by later tests may be left in the working tree in an
unexpected state. For example, a test runs this sequence:
echo foo >xfoo1 && chmod 755 xfoo1
to create an executable file xfoo1, expecting that xfoo1
does not exist before it runs in the test sequence.
However, the absence of this file depends on "git reset
--hard" done in an earlier test, that is skipped when SANITY
prerequisite is not met, and worse yet, xfoo1 originally is
created as a symbolic link, which means the chmod does not
affect the modes of xfoo1 as this test expects.
Fix this by starting the test with "rm -f xfoo1" to make
sure the file is created from scratch, and do the same to
other similar tests.
Signed-off-by: Ingo Brückl <ib@wupperonline.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The result of st_mult() is the same no matter the order of its
arguments. It invokes the macro unsigned_mult_overflows(), which
divides the second parameter by the first one. Pass constants
first to allow that division to be done already at compile time.
Signed-off-by: Rene Scharfe <l.s.r@web.de>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Replace uses of strbuf_addf() for adding strings with more lightweight
strbuf_addstr() calls.
In http-push.c it becomes easier to see what's going on without having
to verfiy that the definition of PROPFIND_ALL_REQUEST doesn't contain
any format specifiers.
Signed-off-by: Rene Scharfe <l.s.r@web.de>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This helps users who would prefer format-patch to default to --from,
and makes it easier to change the default in the future.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix a case where an html link can be generated from unescaped input
resulting in invalid strict xhtml or potentially injected code.
An overview of a repo with a tag "1.0.0&0.0.1" would previously result
in an unescaped ampersand in the link body.
Signed-off-by: Andreas Brauchli <a.brauchli@elementarea.net>
Acked-by: Jakub Narębski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Ever since 66a155b (Enable output buffering in merge-recursive.,
2007-01-14), we had a problem: When the merge failed in a fatal way, all
regular output was swallowed because we called die() and did not get a
chance to drain the output buffers.
To fix this, several modifications were necessary:
- we needed to stop die()ing, to give callers a chance to do something
when an error occurred (in this case, flush the output buffers),
- we needed to delay printing the error message so that the caller can
print the buffered output before that, and
- we needed to make sure that the output buffers are flushed even when
the return value indicates an error.
The first two changes were introduced through earlier commits in this
patch series, and this commit addresses the third one.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The recursive merge machinery accumulates its output in an output
buffer, to be flushed at the end of merge_recursive(). At this point,
we forgot to release the output buffer.
When calling merge_trees() (i.e. the non-recursive part of the recursive
merge) directly, the output buffer is never flushed because the caller
may be merge_recursive() which wants to flush the output itself.
For the same reason, merge_trees() cannot release the output buffer: it
may still be needed.
Forgetting to release the output buffer did not matter much when running
git-checkout, or git-merge-recursive, because we exited after the
operation anyway. Ever since cherry-pick learned to pick a commit range,
however, this memory leak had the potential of becoming a problem.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since 66a155b (Enable output buffering in merge-recursive., 2007-01-14),
we already accumulate the output in a buffer. The idea was to avoid
interfering with the progress output that goes to stderr, which is
unbuffered, when we write to stdout, which is buffered.
We extend that buffering to allow the caller to handle the output
(possibly suppressing it). This will help us when extending the
sequencer to do rebase -i's brunt work: it does not want the picks to
print anything by default but instead determine itself whether to print
the output or not.
Note that we also redirect the error messages into the output buffer
when the caller asked not to flush the output buffer, for two reasons:
1) to retain the correct output order, and 2) to allow the caller to
suppress *all* output.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 66a155b (Enable output buffering in merge-recursive., 2007-01-14), we
changed the code such that it prints the output in one go, to avoid
interfering with the progress output.
Let's make sure that the same holds true when outputting the commit
title: previously, we used several printf() statements to stdout and
assumed that stdout's buffer is large enough to hold the entire
commit title.
Apart from making that speculation unnecessary, we change the code to
add the message to the output buffer before flushing for another reason:
the next commit will introduce a new level of output buffering, where
the caller can request the output not to be flushed, but to be retained
for further processing.
This latter feature will be needed when teaching the sequencer to do
rebase -i's brunt work: it wants to control the output of the
cherry-picks (i.e. recursive merges).
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The data structure passed to the recursive merge machinery has a feature
where the caller can ask for the output to be buffered into a strbuf, by
setting the field 'buffer_output'.
Previously, we died without flushing, losing accumulated output. With
this patch, we show the output first, and only then print the error
message.
Currently, the only user of that buffering is merge_recursive() itself,
to avoid the progress output to interfere.
In the next patches, we will introduce a new buffer_output mode that
forces merge_recursive() to retain the output buffer for further
processing by the caller. If the caller asked for that, we will then
also write the error messages into the output buffer. This is necessary
to give the caller more control not only how to react in case of errors
but also control how/if to display the error messages.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This will allow a diff patch id to be created using only the header data
so that the contents of the file will not have to be loaded.
Signed-off-by: Kevin Willford <kcwillford@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The cherry_pick_list was looping through the original side checking the
seen indicator and setting the cherry_flag on the commit. If we save
off the commit in the patch_id we can set the cherry_flag on the correct
commit when running through the other side when a patch_id match is found.
Signed-off-by: Kevin Willford <kcwillford@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This change will use the hashmap from the hashmap.h to keep track of the
patch_ids that have been encountered instead of using an internal
implementation. This simplifies the implementation of the patch ids.
Signed-off-by: Kevin Willford <kcwillford@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In want_object_in_pack(), we can exit early from our loop if
neither "local" nor "ignore_pack_keep" are set. If they are,
however, we must examine each pack to see if it has the
object and is non-local or has a ".keep".
It's quite common for there to be no non-local or .keep
packs at all, in which case we know ahead of time that
looking further will be pointless. We can pre-compute this
by simply iterating over the list of packs ahead of time,
and dropping the flags if there are no packs that could
match.
Another similar strategy would be to modify the loop in
want_object_in_pack() to notice that we have already found
the object once, and that we are looping only to check for
"local" and "keep" attributes. If a pack has neither of
those, we can skip the call to find_pack_entry_one(), which
is the expensive part of the loop.
This has two advantages:
- it isn't all-or-nothing; we still get some improvement
when there's a small number of kept or non-local packs,
and a large number of non-kept local packs
- it eliminates any possible race where we add new
non-local or kept packs after our initial scan. In
practice, I don't think this race matters; we already
cache the packed_git information, so somebody who adds a
new pack or .keep file after we've started will not be
noticed at all, unless we happen to need to call
reprepare_packed_git() because a lookup fails.
In other words, we're already racy, and the race is not
a big deal (losing the race means we might include an
object in the pack that would not otherwise be, which is
an acceptable outcome).
However, it also has a disadvantage: we still loop over the
rest of the packs for each object to check their flags. This
is much less expensive than doing the object lookup, but
still not free. So if we wanted to implement that strategy
to cover the non-all-or-nothing cases, we could do so in
addition to this one (so you get the most speedup in the
all-or-nothing case, and the best we can do in the other
cases). But given that the all-or-nothing case is likely the
most common, it is probably not worth the trouble, and we
can revisit this later if evidence points otherwise.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When pack-objects collects the list of objects to pack
(either from stdin, or via its internal rev-list), it
filters each one through want_object_in_pack().
This function loops through each existing packfile, looking
for the object. When we find it, we mark the pack/offset
combo for later use. However, we can't just return "yes, we
want it" at that point. If --honor-pack-keep is in effect,
we must keep looking to find it in _all_ packs, to make sure
none of them has a .keep. Likewise, if --local is in effect,
we must make sure it is not present in any non-local pack.
As a result, the sum effort of these calls is effectively
O(nr_objects * nr_packs). In an ordinary repository, we have
only a handful of packs, and this doesn't make a big
difference. But in pathological cases, it can slow the
counting phase to a crawl.
This patch notices the case that we have neither "--local"
nor "--honor-pack-keep" in effect and breaks out of the loop
early, after finding the first instance. Note that our worst
case is still "objects * packs" (i.e., we might find each
object in the last pack we look in), but in practice we will
often break out early. On an "average" repo, my git.git with
8 packs, this shows a modest 2% (a few dozen milliseconds)
improvement in the counting-objects phase of "git
pack-objects --all <foo" (hackily instrumented by sticking
exit(0) right after list_objects).
But in a much more pathological case, it makes a bigger
difference. I ran the same command on a real-world example
with ~9 million objects across 1300 packs. The counting time
dropped from 413s to 45s, an improvement of about 89%.
Note that this patch won't do anything by itself for a
normal "git gc", as it uses both --honor-pack-keep and
--local.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Each pack has an index for looking up entries in O(log n)
time, but if we have multiple packs, we have to scan through
them linearly. This can produce a measurable overhead for
some operations.
We dealt with this long ago in f7c22cc (always start looking
up objects in the last used pack first, 2007-05-30), which
keeps what is essentially a 1-element most-recently-used
cache. In theory, we should be able to do better by keeping
a similar but longer cache, that is the same length as the
pack-list itself.
Since we now have a convenient generic MRU structure, we can
plug it in and measure. Here are the numbers for running
p5303 against linux.git:
Test HEAD^ HEAD
------------------------------------------------------------------------
5303.3: rev-list (1) 31.56(31.28+0.27) 31.30(31.08+0.20) -0.8%
5303.4: repack (1) 40.62(39.35+2.36) 40.60(39.27+2.44) -0.0%
5303.6: rev-list (50) 31.31(31.06+0.23) 31.23(31.00+0.22) -0.3%
5303.7: repack (50) 58.65(69.12+1.94) 58.27(68.64+2.05) -0.6%
5303.9: rev-list (1000) 38.74(38.40+0.33) 31.87(31.62+0.24) -17.7%
5303.10: repack (1000) 367.20(441.80+4.62) 342.00(414.04+3.72) -6.9%
The main numbers of interest here are the rev-list ones
(since that is exercising the normal object lookup code
path). The single-pack case shouldn't improve at all; the
260ms speedup there is just part of the run-to-run noise
(but it's important to note that we didn't make anything
worse with the overhead of maintaining our cache). In the
50-pack case, we see similar results. There may be a slight
improvement, but it's mostly within the noise.
The 1000-pack case does show a big improvement, though. That
carries over to the repack case, as well. Even though we
haven't touched its pack-search loop yet, it does still do a
lot of normal object lookups (e.g., for the internal
revision walk), and so improves.
As a point of reference, I also ran the 1000-pack test
against a version of HEAD^ with the last_found_pack
optimization disabled. It takes ~60s, so that gives an
indication of how much even the single-element cache is
helping.
For comparison, here's a smaller repository, git.git:
Test HEAD^ HEAD
---------------------------------------------------------------------
5303.3: rev-list (1) 1.56(1.54+0.01) 1.54(1.51+0.02) -1.3%
5303.4: repack (1) 1.84(1.80+0.10) 1.82(1.80+0.09) -1.1%
5303.6: rev-list (50) 1.58(1.55+0.02) 1.59(1.57+0.01) +0.6%
5303.7: repack (50) 2.50(3.18+0.04) 2.50(3.14+0.04) +0.0%
5303.9: rev-list (1000) 2.76(2.71+0.04) 2.24(2.21+0.02) -18.8%
5303.10: repack (1000) 13.21(19.56+0.25) 11.66(18.01+0.21) -11.7%
You can see that the percentage improvement is similar.
That's because the lookup we are optimizing is roughly
O(nr_objects * nr_packs). Since the number of packs is
constant in both tests, we'd expect the improvement to be
linear in the number of objects. But the whole process is
also linear in the number of objects, so the improvement
is a constant factor.
The exact improvement does also depend on the contents of
the packs. In p5303, the extra packs all have 5 first-parent
commits in them, which is a reasonable simulation of a
pushed-to repository. But it also means that only 250
first-parent commits are in those packs (compared to almost
50,000 total in linux.git), and the rest are in the huge
"base" pack. So once we start looking at history in taht big
pack, that's where we'll find most everything, and even the
1-element cache gets close to 100% cache hits. You could
almost certainly show better numbers with a more
pathological case (e.g., distributing the objects more
evenly across the packs). But that's simply not that
realistic a scenario, so it makes more sense to focus on
these numbers.
The implementation itself is a straightforward application
of the MRU code. We provide an MRU-ordered list of packs
that shadows the packed_git list. This is easy to do because
we only create and revise the pack list in one place. The
"reprepare" code path actually drops the whole MRU and
replaces it for simplicity. It would be more efficient to
just add new entries, but there's not much point in
optimizing here; repreparing happens rarely, and only after
doing a lot of other expensive work. The key things to keep
optimized are traversal (which is just a normal linked list,
albeit with one extra level of indirection over the regular
packed_git list), and marking (which is a constant number of
pointer assignments, though slightly more than the old
last_found_pack was; it doesn't seem to create a measurable
slowdown, though).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There are a few places in Git that would benefit from a fast
most-recently-used cache (e.g., the list of packs, which we
search linearly but would like to order based on locality).
This patch introduces a generic list that can be used to
store arbitrary pointers in most-recently-used order.
The implementation is just a doubly-linked list, where
"marking" an item as used moves it to the front of the list.
Insertion and marking are O(1), and iteration is O(n).
There's no lookup support provided; if you need fast
lookups, you are better off with a different data structure
in the first place.
There is also no deletion support. This would not be hard to
do, but it's not necessary for handling pack structs, which
are created and never removed.
Signed-off-by: Jeff King <peff@peff.net>
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>
Git's pack storage does efficient (log n) lookups in a
single packfile's index, but if we have multiple packfiles,
we have to linearly search each for a given object. This
patch introduces some timing tests for cases where we have a
large number of packs, so that we can measure any
improvements we make in the following patches.
The main thing we want to time is object lookup. To do this,
we measure "git rev-list --objects --all", which does a
fairly large number of object lookups (essentially one per
object in the repository).
However, we also measure the time to do a full repack, which
is interesting for two reasons. One is that in addition to
the usual pack lookup, it has its own linear iteration over
the list of packs. And two is that because it it is the tool
one uses to go from an inefficient many-pack situation back
to a single pack, we care about its performance not only at
marginal numbers of packs, but at the extreme cases (e.g.,
if you somehow end up with 5,000 packs, it is the only way
to get back to 1 pack, so we need to make sure it performs
well).
We measure the performance of each command in three
scenarios: 1 pack, 50 packs, and 1,000 packs.
The 1-pack case is a baseline; any optimizations we do to
handle multiple packs cannot possibly perform better than
this.
The 50-pack case is as far as Git should generally allow
your repository to go, if you have auto-gc enabled with the
default settings. So this represents the maximum performance
improvement we would expect under normal circumstances.
The 1,000-pack case is hopefully rare, though I have seen it
in the wild where automatic maintenance was broken for some
time (and the repository continued to receive pushes). This
represents cases where we care less about general
performance, but want to make sure that a full repack
command does not take excessively long.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
