* Re: [PATCH] push options: fail properly in the stateless case
From: Junio C Hamano @ 2017-02-08 18:11 UTC (permalink / raw)
To: Stefan Beller; +Cc: git, jrnieder
In-Reply-To: <20170208010954.19478-1-sbeller@google.com>
Stefan Beller <sbeller@google.com> writes:
> When using non-builtin protocols relying on a transport helper
> (such as http), push options are not propagated to the helper.
>
> Fix this by propagating the push options to the transport helper and
The description up to this point is VERY readable and sensible. But
that makes the title sound a bit strange. I read it as if it were
saying "stateless case can never support push-options so fail if the
caller attempts to use one", but that does not seem to be what is
going on.
> adding a test that push options using http fail properly.
Sounds sensible. What end-user visible effect does this fix have?
IOW, what feature do we use "push-option" for?
Ahh, OK, so you need to describe that there are two issues in order
to be understood by the readers:
(1) the helper protocol does not propagate push-option
(2) the http helper is not prepared to handle push-option
You fix (1), and you take advantage of the fact (2) to ensure that
(1) is fixed in the new test.
With such an understanding, the title makes (sort of) sense and you
wouldn't have to be asked "what end-user visible effect/benefit does
this have?"
> +'option push-option <c-string>::
> + Transmit this push option.
> +
There is no "c-string" in the current documentation used or
defined. The closest thing I found is
... that field will be quoted in the manner of a C string ...
in git-status page, but I do not think you send the value for an
push-option after running quote_c_style(), so I am puzzled.
I'd rather see 'option push-option <string>' as the bullet item, and
in its description say how arbitrary values (if you allow them, that
is) can be used, e.g. "Transmit <string> encoded in such and such
way a the value of the push-option".
^ permalink raw reply
* Re: [PATCH RFC 0/2] Kill manual ref parsing code in worktree.c
From: Junio C Hamano @ 2017-02-08 18:18 UTC (permalink / raw)
To: Nguyễn Thái Ngọc Duy; +Cc: git, Michael Haggerty
In-Reply-To: <20170208113144.8201-1-pclouds@gmail.com>
Nguyễn Thái Ngọc Duy <pclouds@gmail.com> writes:
> A hundred years ago I added this code because a "standalone" ref
> parsing code was not available from refs.c and the file was going through
> some heavy changes that refactoring its ref parsing code out was not
> an option. I promised to kill this parse_ref() eventually. I'm
> fulfilling it today (or soon).
;-)
Thanks. And thanks for looping Michael in. I'd appreciate his
input in this area.
> I would really like to you double check the approach I'm using here
> (using submodule interface for accessing refs from another worktree)
> since that may be the way forward to fix the "gc losing objects" in
> multi worktrees. I've given it lots of thoughts in the last 24 hours.
> Still can't find any fundamental flaw...
I see that you posted a separate message outlining the idea
yesterday and I didn't see any response (and I was sick and lacked
energy to think it through); I think the basic approach to use "an
API to bring set of refs hidden from your normal view" is sensible.
Except for the unfortunate naming of the interface that makes it
sound as if it is only to access submodules, but that is where the
feature original came from, so let's not complain too loudly ;-)
> Nguyễn Thái Ngọc Duy (2):
> refs.c: add resolve_ref_submodule()
> worktree.c: use submodule interface to access refs from another worktree
>
> branch.c | 3 +-
> refs.c | 20 +++++++++----
> refs.h | 3 ++
> worktree.c | 99 +++++++++++++++-----------------------------------------------
> worktree.h | 2 +-
> 5 files changed, 44 insertions(+), 83 deletions(-)
^ permalink raw reply
* Software Freedom Conservancy donations
From: Jeff King @ 2017-02-08 18:34 UTC (permalink / raw)
To: git
The recent thread about the git-scm.com website raised some discussion
about money, and a lot of people offered financial assistance. As I said
in that thread, we _don't_ have a dire need for money to keep hosting
the site.
But we do sometimes need money for other things (like conference
travel), and we rely on services (like accounting and legal advice)
provided by the Software Freedom Conservancy which are free to us, but
funded ultimately by donations.
I went into detail on our project activities and finances recently at:
http://public-inbox.org/git/20170202024501.57hrw4657tsqerqq@sigill.intra.peff.net/
If you are interested in contributing financially, you can either donate
to the Git project (10% of which goes to Conservancy's general fund, and
the rest ends up in Git's account):
https://git-scm.com/sfc
Or you can donate directly to Conservancy's general fund:
https://sfconservancy.org/donate/
They also have an annual Supporter program, which includes the
all-important t-shirt:
https://sfconservancy.org/supporter/
Through February 13th, there's an anonymous donor matching all Supporter
signups and renewals, so any donation you make before then will go twice
as far.
-Peff
^ permalink raw reply
* Re: Trying to use xfuncname without success.
From: René Scharfe @ 2017-02-08 18:37 UTC (permalink / raw)
To: Jack Adrian Zappa; +Cc: git-mailing-list
In-Reply-To: <CAKepmagp2mXNviA2VdT=3EQtZi2LkA_5oG6=AbfkBGKP9Hqiiw@mail.gmail.com>
Am 08.02.2017 um 18:11 schrieb Jack Adrian Zappa:
> Thanks Rene, but you seem to have missed the point. NOTHING is
> working. No matter what I put there, it doesn't seem to get matched.
I'm not so sure about that. With your example I get this diff without
setting diff.natvis.xfuncname:
diff --git a/a.natvis b/a.natvis
index 7f9bdf5..bc3c090 100644
--- a/a.natvis
+++ b/a.natvis
@@ -19,7 +19,7 @@ xmlns="http://schemas.microsoft.com/vstudio/debugger/natvis/2010">
<!-- Non-blank line -->
- <Item Name="added var">added_var</Item>
+ <Item Name="added var">added_vars</Item>
<Item Name="var2">var2</Item>
Note the XML namespace in the hunk header. It's put there by the
default rule because "xmlns" starts at the beginning of the line. Your
diff has nothing there, which means the default rule is not used, i.e.
your user-defined rule is in effect.
Come to think of it, this line break in the middle of the AutoVisualizer
tab might have been added by your email client unintentionally, so that
we use different test files, which then of course results in different
diffs. Is that the case?
Anyway, if I run the following two commands:
$ git config diff.natvis.xfuncname "^[\t ]*<Type[\t ]+Name=\"([^\"]+)\".*$"
$ echo '*.natvis diff=natvis' >.gitattributes
... then I get this, both on Linux (git version 2.11.1) and on Windows
(git version 2.11.1.windows.1):
diff --git a/a.natvis b/a.natvis
index 7f9bdf5..bc3c090 100644
--- a/a.natvis
+++ b/a.natvis
@@ -19,7 +19,7 @@ test
<!-- Non-blank line -->
- <Item Name="added var">added_var</Item>
+ <Item Name="added var">added_vars</Item>
<Item Name="var2">var2</Item>
> Just to be sure, I tested your regex and again it didn't work.
At this point I'm out of ideas, sorry. :( The only way I was able to
break it was due to mistyping the extension as "netvis" several times
for some reason.
René
^ permalink raw reply related
* Re: [PATCH] rev-parse --git-path: fix output when running in a subdirectory
From: Junio C Hamano @ 2017-02-08 18:47 UTC (permalink / raw)
To: Johannes Schindelin; +Cc: git, Nguyễn Thái Ngọc Duy
In-Reply-To: <50fe3ea3302c40f4c96eaa5a568837e3334f9dc4.1486555851.git.johannes.schindelin@gmx.de>
Johannes Schindelin <johannes.schindelin@gmx.de> writes:
> The problem lies deeper, much deeper...
> ... but it bought us many, many problems.
I think you are making a mountain out of molehill here.
This looks like as an opposite problem of a bug that forgets to
prepend the prefix to relative pathnames given by the end user.
I do agree that some calling scripts may find it more convenient if
the output were relative to their current directory, and in that
sense, this is worth addressing.
However.
How long has "rev-parse --git-path" been around? Had scripts in the
wild chance to learn to live with the "output is relative to the top
of the working tree" reality? I think the answers are "since 2.5" and
"yes".
I do not think we can make this unconditionally without breaking
users. We instead need something like a new "--git-path-relative"
option, similar in the spirit that output from "git diff" can be
made relative to the current directory with the "--relative" option.
Assuming that we are discussing the new behaviour that is
conditionally triggered, let's see the implementation.
> - puts(git_path("%s", argv[i + 1]));
> + path = git_path("%s", argv[i + 1]);
> + if (prefix && !is_absolute_path(path))
> + path = real_path(path);
> + puts(path);
Duy, want to help me out here? I am wondering if using a logic
similar to what is used by "cd t && git grep -e foo :/" to emit
paths as "../Documentation/CodingGuidelines" as relative to the
current working directory is more appropriate than forcing the
absolute path output here (and if so, it may be preferrable to use
the relative_path() helper to do so), or the paths to files in
$GIT_DIR are conceptually different enough from paths to files in
the working tree and it will be more robust to have the output as an
absolute path.
I am leaning toward the latter (i.e. the above use of real_path() is
simple and good), but I haven't thought things through and since we
have an area expert here in the thread...
^ permalink raw reply
* Re: "disabling bitmap writing, as some objects are not being packed"?
From: Jeff King @ 2017-02-08 19:08 UTC (permalink / raw)
To: David Turner; +Cc: Junio C Hamano, Duy Nguyen, Git Mailing List
In-Reply-To: <1486580742.1938.52.camel@novalis.org>
On Wed, Feb 08, 2017 at 02:05:42PM -0500, David Turner wrote:
> On Wed, 2017-02-08 at 09:44 -0800, Junio C Hamano wrote:
> > Duy Nguyen <pclouds@gmail.com> writes:
> >
> > > On second thought, perhaps gc.autoDetach should default to false if
> > > there's no tty, since its main point it to stop breaking interactive
> > > usage. That would make the server side happy (no tty there).
> >
> > Sounds like an idea, but wouldn't that keep the end-user coming over
> > the network waiting after accepting a push until the GC completes, I
> > wonder. If an impatient user disconnects, would that end up killing
> > an ongoing GC? etc.
>
> Regardless, it's impolite to keep the user waiting. So, I think we
> should just not write the "too many unreachable loose objects" message
> if auto-gc is on. Does that sound OK?
I thought the point of that message was to prevent auto-gc from kicking
in over and over again due to objects that won't actually get pruned.
I wonder if you'd want to either bump the auto-gc object limit, or
possibly reduce the gc.pruneExpire limit to keep this situation from
coming up in the first place (or at least mitigating the amount of time
it's the case).
-Peff
^ permalink raw reply
* Re: gitconfig get out of sync with submodule entries on branch switch
From: Stefan Beller @ 2017-02-08 19:07 UTC (permalink / raw)
To: Benjamin Schindler; +Cc: git@vger.kernel.org
In-Reply-To: <7e54658a-dcb2-64a7-3c67-0c4fa221b2fb@gmail.com>
On Mon, Feb 6, 2017 at 4:17 AM, Benjamin Schindler
<beschindler@gmail.com> wrote:
>
>
> On 06.02.2017 11:35, Stefan Beller wrote:
>>
>> Answering the original email, as I feel we're going down the wrong rabbit
>> hole in the existing thread.
>>
>> On Mon, Jan 30, 2017 at 8:21 AM, Benjamin Schindler
>> <beschindler@gmail.com> wrote:
>>>
>>> Hi
>>>
>>> Consider the following usecase: I have the master branch where I have a
>>> submodule A. I create a branch where I rename the submodule to be in the
>>> directory B. After doing all of this, everything looks good.
>>> Now, I switch back to master. The first oddity is, that it fails to
>>> remove
>>> the folder B because there are still files in there:
>>>
>>> bschindler@metis ~/Projects/submodule_test (testbranch) $ git checkout
>>> master
>>> warning: unable to rmdir other_submodule: Directory not empty
>>> Switched to branch 'master'
>>
>>
>> checkout currently doesn't support submodules, so it should neither
>> try to delete B nor try to repopulate A when switching back to master.
>> checkout ought to not even touch the existing submodule B.
>
>
> Well, it tried to remove the folder (the rmdir warning) but it failed so in
> some sense you are right. Is there a technical reason for this default
> though? Here, I frequently have to point out to people that they need to
> initialize/update the submodule on e.g. clone.
The reasoning is more based on historical momentum.
Back then when gitlinks/submodules were invented (git repositories
inside other git repositories! How frickin' cool is that?) nobody quite knew
how they were going to be used eventually. So the safe play at the time was
to not touch them at all. (Also easy to implement, but that was not the point
as I learned).
And now we have a lot of people expecting just that. So we cannot change
the behavior overnight. So we'd first implement the alternative (e.g. the
--recurse-submodules flag for checkout) and then once we do a major
release we may want to flip the default.
>
>>
>>>
>>> Git submodule deinit on B fails because the submodule is not known to git
>>> anymore (after all, the folder B exists only in the other branch). I can
>>> easily just remove the folder B from disk and initialize the submodule A
>>> again, so all seems good.
>>
>>
>> by initializing you mean populating(?), i.e.
>>
>> git submodule update
>>
>> would work without the --init flag or preceding "git submodule init A".
>> That ought to not redownload A, but just put files back in the working
>> tree
>> from the submodule git directory inside the superprojects git dir.
>>
>>>
>>> However, what is not good is that the submodule b is still known in
>>> .git/config.
>>
>>
>> Oh, I see. You did not just rename the path, but also the name
>> in the .gitmodules?
>
>
> I wasn't even aware that the submodule name was something different from the
> path because the name is by default set to be the path to it. So yes, I
> didn't just relocate it, it had a different name.
Yeah the path/name is tricky and usually only differs when you move the
submodule inside the working tree. (As the name stays constant we know
where the git directory is expected: .git/modules<name> so we can check there
instead of re-cloning to the "new" submodule-path.)
>
>>
>>> This is in particular a problem for us, because I know a number
>>> of tools which use git config to retrieve the submodule list. Is it
>>> therefore a bug that upon branch switch, the submodule gets deregistered,
>>> but its entry in .git/config remains?
>>
>>
>> The config remains as it indicates that you express(ed) interest in
>> submodule A, such that when switching branches
>>
>> master->renamedToB->master
>>
>> then we'd still care about A. As for the tools, I'd rather see them use
>>
>> git submodule status/summary
>>
>> instead of directly looking at the config, because the config may
>> change in the future.
>
>
> That was my feeling but its good to know to have more solid reasons why that
> would be.
The reasons here are backward/forward compatibility.
When trying to change the behavior of submodule related things, there is no
clear distinction of plumbing/porcelain as we have in the rest of Git.
So even in gits own test suite we sometimes rely on things that may change
later on, and that makes it very hard to move forward, which is why I try to
have an opinion on how to do things properly.
Thanks,
Stefan
^ permalink raw reply
* Re: [PATCH 2/2] worktree.c: use submodule interface to access refs from another worktree
From: Stefan Beller @ 2017-02-08 19:50 UTC (permalink / raw)
To: Nguyễn Thái Ngọc Duy
Cc: git@vger.kernel.org, Junio C Hamano, Michael Haggerty
In-Reply-To: <20170208113144.8201-3-pclouds@gmail.com>
On Wed, Feb 8, 2017 at 3:31 AM, Nguyễn Thái Ngọc Duy <pclouds@gmail.com> wrote:
> (**) At this point, we may want to rename refs *_submodule API to
> something more neutral, maybe s/_submodule/_remote/
I agree on (**), except that I am not sure if /_remote/ is a better name,
because there is already a concept of a "remote" as well as
"remote-tracking" in Git. (Usually it is not reachable on the same
FS, but resides on another machine).
So if we were to do s/_submodule/_remote/, we'd have e.g.
for_each_ref_remote
which could mean that we do networking things to obtain the
actual remote refs or just talk about remote tracking refs.
My gut reaction would be to s/submodule/alternative/ here,
but we also have a thing called alternates already.
So we're looking for a name that describes refs for both
worktrees as well as submodules. (Not sure if we can generalize
to also include alternates, too)
And the one thing they share is that they have their refs
"not at the usual place", e.g. not at .git/refs but rather at
.git/{modules,worktrees}.
Recently we had a tangential discussion in submodule land
about the different places, when adding the
"git submodule absorbgitdirs" command, that moves
the submodule/.git directory into .git/modules/<name>.
We chose "absorb" here as the name, because it
was moved into the .git/ area.
So maybe one of:
s/submodule_ref/unusual_ref/
(to emphasize it is not a regular ref inside the repo, so:)
s/submodule_ref/irregular_ref/
s/resolve_ref_submodule/resolve_ref_out_of_place/
s/resolve_ref_submodule/resolve_ref_gitlink_pointed/
s/resolve_ref_submodule/resolve_ref_linked_pointer/
would be my current thinking.
Thanks,
Stefan
^ permalink raw reply
* Re: [RFC] mailmap.blob overrides default .mailmap
From: Jeff King @ 2017-02-08 19:50 UTC (permalink / raw)
To: Cornelius Weig; +Cc: Stefan Beller, git@vger.kernel.org
In-Reply-To: <53836bcd-1d4d-13fb-a523-1258017d19c9@tngtech.com>
On Tue, Feb 07, 2017 at 11:45:31PM +0100, Cornelius Weig wrote:
> On the other hand, a checked-in .mailmap file and a mailmap-blob are
> both as in-history as the other to me. Now consider the following
> settings:
I think it depends how you use them. You could point mailmap.blob to
some other ref entirely (even one that you fetched from another
repository).
I'd expect normal use to point it to HEAD:.mailmap, though (and that was
certainly the use case I wrote it for). On the other hand, the point of
pointing it to that particular blob is that it works even when you
_don't_ have a checkout (and this kicks in automatically in a bare
repo).
> $ git config --unset mailmap.file
> $ git config mailmap.blob HEAD:.mailmap
> $ sed -i 's:peff@peff.com:no-valid-address:' .mailmap
> $ git log -1 --author 'Jeff King'
In case anybody wants to experiment, there are a bunch of things that
make this a non-working example (at least on git.git):
- my address is actually peff.net :)
- There mailmap which mentions peff.net maps peff@github.com to
peff.net, so this change would require --author=peff@github.com.
- We don't apply mailmaps for the default output of "git log". You can
format with "%aN %aE", or just use "git shortlog -ns --author=peff"
which does map.
But that aside, yeah, you can make an argument to expect one way or the
other, depending on the situation you set up. I don't have a strong
feeling about it, but my gut feeling is that no ordering is
significantly better than the other, and that puts me in favor of
leaving it as-is purely out of inertia and backwards-compatibility.
-Peff
^ permalink raw reply
* Re: Trying to use xfuncname without success.
From: Samuel Lijin @ 2017-02-08 20:24 UTC (permalink / raw)
To: René Scharfe; +Cc: Jack Adrian Zappa, git-mailing-list
In-Reply-To: <1aa20b4e-782f-a650-eab8-51218b838337@web.de>
On Windows 7, it works for me in both CMD and Git Bash:
$ git --version
git version 2.11.0.windows.3
$ git diff HEAD^ --word-diff
diff --git a/test.natvis b/test.natvis
index 93396ad..1233b8c 100644
--- a/test.natvis
+++ b/test.natvis
@@ -18,6 +18,7 @@ test
<!-- Non-blank line -->
{+<Item Name="added var">added_var</Item>+}
<Item Name="var2">var2</Item>
On Wed, Feb 8, 2017 at 12:37 PM, René Scharfe <l.s.r@web.de> wrote:
> Am 08.02.2017 um 18:11 schrieb Jack Adrian Zappa:
>> Thanks Rene, but you seem to have missed the point. NOTHING is
>> working. No matter what I put there, it doesn't seem to get matched.
>
> I'm not so sure about that. With your example I get this diff without
> setting diff.natvis.xfuncname:
>
> diff --git a/a.natvis b/a.natvis
> index 7f9bdf5..bc3c090 100644
> --- a/a.natvis
> +++ b/a.natvis
> @@ -19,7 +19,7 @@ xmlns="http://schemas.microsoft.com/vstudio/debugger/natvis/2010">
>
>
> <!-- Non-blank line -->
> - <Item Name="added var">added_var</Item>
> + <Item Name="added var">added_vars</Item>
>
>
> <Item Name="var2">var2</Item>
>
> Note the XML namespace in the hunk header. It's put there by the
> default rule because "xmlns" starts at the beginning of the line. Your
> diff has nothing there, which means the default rule is not used, i.e.
> your user-defined rule is in effect.
>
> Come to think of it, this line break in the middle of the AutoVisualizer
> tab might have been added by your email client unintentionally, so that
> we use different test files, which then of course results in different
> diffs. Is that the case?
>
> Anyway, if I run the following two commands:
>
> $ git config diff.natvis.xfuncname "^[\t ]*<Type[\t ]+Name=\"([^\"]+)\".*$"
> $ echo '*.natvis diff=natvis' >.gitattributes
>
> ... then I get this, both on Linux (git version 2.11.1) and on Windows
> (git version 2.11.1.windows.1):
>
> diff --git a/a.natvis b/a.natvis
> index 7f9bdf5..bc3c090 100644
> --- a/a.natvis
> +++ b/a.natvis
> @@ -19,7 +19,7 @@ test
>
>
> <!-- Non-blank line -->
> - <Item Name="added var">added_var</Item>
> + <Item Name="added var">added_vars</Item>
>
>
> <Item Name="var2">var2</Item>
>
>> Just to be sure, I tested your regex and again it didn't work.
>
> At this point I'm out of ideas, sorry. :( The only way I was able to
> break it was due to mistyping the extension as "netvis" several times
> for some reason.
>
> René
^ permalink raw reply related
* [PATCH] diff: print line prefix for --name-only output
From: Jeff King @ 2017-02-08 20:31 UTC (permalink / raw)
To: Duy Nguyen; +Cc: Davide Del Vento, Git Mailing List
In-Reply-To: <CACsJy8BmxpTW5pCLNEQa9sm8teU0O+_Xu+td+uh1K0Bn=yu=yw@mail.gmail.com>
On Wed, Feb 08, 2017 at 01:24:34PM +0700, Duy Nguyen wrote:
> On Wed, Feb 8, 2017 at 6:11 AM, Davide Del Vento <ddvento@ucar.edu> wrote:
> > `git log --graph --name-only` works fine, but the name is not
> > properly indented as it is for `git log --graph --name-status`
> >
> > I tested this in git v1.9.1 the only one I have access at the moment
>
> Confirmed still happens on master. --stat plays nicely with --graph
> though, so it's probably just some small fixes somewhere in diff*.c.
Yep. Looks like every format except --name-only handles this correctly.
-- >8 --
Subject: diff: print line prefix for --name-only output
If you run "git log --graph --name-only", the pathnames are
not indented to go along with their matching commits (unlike
all of the other diff formats). We need to output the line
prefix for each item before writing it.
The tests cover both --name-status and --name-only. The
former actually gets this right already, because it builds
on the --raw format functions. It's only --name-only which
uses its own code (and this fix mirrors the code in
diff_flush_raw()).
Note that the tests don't follow our usual style of setting
up the "expect" output inside the test block. This matches
the surrounding style, but more importantly it is easier to
read: we don't have to worry about embedded single-quotes,
and the leading indentation is more obvious.
Signed-off-by: Jeff King <peff@peff.net>
---
diff.c | 1 +
t/t4202-log.sh | 48 ++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 49 insertions(+)
diff --git a/diff.c b/diff.c
index d91a34490..a79f3408d 100644
--- a/diff.c
+++ b/diff.c
@@ -4450,6 +4450,7 @@ static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
name_a = p->two->path;
name_b = NULL;
strip_prefix(opt->prefix_length, &name_a, &name_b);
+ fprintf(opt->file, "%s", diff_line_prefix(opt));
write_name_quoted(name_a, opt->file, opt->line_termination);
}
}
diff --git a/t/t4202-log.sh b/t/t4202-log.sh
index 08ea725de..48b55bfd2 100755
--- a/t/t4202-log.sh
+++ b/t/t4202-log.sh
@@ -1212,6 +1212,54 @@ test_expect_success 'log --line-prefix="*** " --graph with diff and stats' '
test_i18ncmp expect actual.sanitized
'
+cat >expect <<-\EOF
+* reach
+|
+| A reach.t
+* Merge branch 'tangle'
+* Merge branch 'side'
+|\
+| * side-2
+|
+| A 2
+* Second
+|
+| A one
+* sixth
+
+ D a/two
+EOF
+
+test_expect_success 'log --graph with --name-status' '
+ git log --graph --format=%s --name-status tangle..reach >actual &&
+ sanitize_output <actual >actual.sanitized &&
+ test_cmp expect actual.sanitized
+'
+
+cat >expect <<-\EOF
+* reach
+|
+| reach.t
+* Merge branch 'tangle'
+* Merge branch 'side'
+|\
+| * side-2
+|
+| 2
+* Second
+|
+| one
+* sixth
+
+ a/two
+EOF
+
+test_expect_success 'log --graph with --name-only' '
+ git log --graph --format=%s --name-only tangle..reach >actual &&
+ sanitize_output <actual >actual.sanitized &&
+ test_cmp expect actual.sanitized
+'
+
test_expect_success 'dotdot is a parent directory' '
mkdir -p a/b &&
( echo sixth && echo fifth ) >expect &&
--
2.12.0.rc0.371.ga6cf8653b
^ permalink raw reply related
* Re: Trying to use xfuncname without success.
From: Jack Adrian Zappa @ 2017-02-08 20:39 UTC (permalink / raw)
To: René Scharfe; +Cc: git-mailing-list
In-Reply-To: <1aa20b4e-782f-a650-eab8-51218b838337@web.de>
> I'm not so sure about that. With your example I get this diff without
setting diff.natvis.xfuncname:
So, to make sure we are on the same page, I removed the
diff.natvis.xfuncname from the .gitconfig and .git/config. My output
was:
C:\Users\adrianh\Documents\tmp>git diff
diff --git a/test.natvis b/test.natvis
index 93fd5b4..351301f 100644
--- a/test.natvis
+++ b/test.natvis
@@ -18,6 +18,7 @@
<!-- test text -->
+ <Item Name="added var">added_var</Item>
<Item Name="var2">var2</Item>
So I didn't get the default output that your specified.
I've been modifying the .gitconfig file directly, but I tried your command:
git config diff.natvis.xfuncname "^[\t ]*<Type[\t ]+Name=\"([^\"]+)\".*$"
and still had the same results. I.e. NOTHING. :(
On Wed, Feb 8, 2017 at 1:37 PM, René Scharfe <l.s.r@web.de> wrote:
> Am 08.02.2017 um 18:11 schrieb Jack Adrian Zappa:
>> Thanks Rene, but you seem to have missed the point. NOTHING is
>> working. No matter what I put there, it doesn't seem to get matched.
>
> I'm not so sure about that. With your example I get this diff without
> setting diff.natvis.xfuncname:
>
> diff --git a/a.natvis b/a.natvis
> index 7f9bdf5..bc3c090 100644
> --- a/a.natvis
> +++ b/a.natvis
> @@ -19,7 +19,7 @@ xmlns="http://schemas.microsoft.com/vstudio/debugger/natvis/2010">
>
>
> <!-- Non-blank line -->
> - <Item Name="added var">added_var</Item>
> + <Item Name="added var">added_vars</Item>
>
>
> <Item Name="var2">var2</Item>
>
> Note the XML namespace in the hunk header. It's put there by the
> default rule because "xmlns" starts at the beginning of the line. Your
> diff has nothing there, which means the default rule is not used, i.e.
> your user-defined rule is in effect.
>
> Come to think of it, this line break in the middle of the AutoVisualizer
> tab might have been added by your email client unintentionally, so that
> we use different test files, which then of course results in different
> diffs. Is that the case?
>
> Anyway, if I run the following two commands:
>
> $ git config diff.natvis.xfuncname "^[\t ]*<Type[\t ]+Name=\"([^\"]+)\".*$"
> $ echo '*.natvis diff=natvis' >.gitattributes
>
> ... then I get this, both on Linux (git version 2.11.1) and on Windows
> (git version 2.11.1.windows.1):
>
> diff --git a/a.natvis b/a.natvis
> index 7f9bdf5..bc3c090 100644
> --- a/a.natvis
> +++ b/a.natvis
> @@ -19,7 +19,7 @@ test
>
>
> <!-- Non-blank line -->
> - <Item Name="added var">added_var</Item>
> + <Item Name="added var">added_vars</Item>
>
>
> <Item Name="var2">var2</Item>
>
>> Just to be sure, I tested your regex and again it didn't work.
>
> At this point I'm out of ideas, sorry. :( The only way I was able to
> break it was due to mistyping the extension as "netvis" several times
> for some reason.
>
> René
^ permalink raw reply related
* Re: Trying to use xfuncname without success.
From: Jack Adrian Zappa @ 2017-02-08 20:42 UTC (permalink / raw)
To: Samuel Lijin; +Cc: René Scharfe, git-mailing-list
In-Reply-To: <CAJZjrdXjnDMi8gMY6f_UDbMZrZJ=AoPM+g01hqPCO2pB9csoOw@mail.gmail.com>
Thanks Samuel,
So, the question is, what is causing this problem on my system?
Anyone have an idea to help diagnose this problem?
On Wed, Feb 8, 2017 at 3:24 PM, Samuel Lijin <sxlijin@gmail.com> wrote:
> On Windows 7, it works for me in both CMD and Git Bash:
>
> $ git --version
> git version 2.11.0.windows.3
>
> $ git diff HEAD^ --word-diff
> diff --git a/test.natvis b/test.natvis
> index 93396ad..1233b8c 100644
> --- a/test.natvis
> +++ b/test.natvis
> @@ -18,6 +18,7 @@ test
>
>
> <!-- Non-blank line -->
> {+<Item Name="added var">added_var</Item>+}
>
>
> <Item Name="var2">var2</Item>
>
> On Wed, Feb 8, 2017 at 12:37 PM, René Scharfe <l.s.r@web.de> wrote:
>> Am 08.02.2017 um 18:11 schrieb Jack Adrian Zappa:
>>> Thanks Rene, but you seem to have missed the point. NOTHING is
>>> working. No matter what I put there, it doesn't seem to get matched.
>>
>> I'm not so sure about that. With your example I get this diff without
>> setting diff.natvis.xfuncname:
>>
>> diff --git a/a.natvis b/a.natvis
>> index 7f9bdf5..bc3c090 100644
>> --- a/a.natvis
>> +++ b/a.natvis
>> @@ -19,7 +19,7 @@ xmlns="http://schemas.microsoft.com/vstudio/debugger/natvis/2010">
>>
>>
>> <!-- Non-blank line -->
>> - <Item Name="added var">added_var</Item>
>> + <Item Name="added var">added_vars</Item>
>>
>>
>> <Item Name="var2">var2</Item>
>>
>> Note the XML namespace in the hunk header. It's put there by the
>> default rule because "xmlns" starts at the beginning of the line. Your
>> diff has nothing there, which means the default rule is not used, i.e.
>> your user-defined rule is in effect.
>>
>> Come to think of it, this line break in the middle of the AutoVisualizer
>> tab might have been added by your email client unintentionally, so that
>> we use different test files, which then of course results in different
>> diffs. Is that the case?
>>
>> Anyway, if I run the following two commands:
>>
>> $ git config diff.natvis.xfuncname "^[\t ]*<Type[\t ]+Name=\"([^\"]+)\".*$"
>> $ echo '*.natvis diff=natvis' >.gitattributes
>>
>> ... then I get this, both on Linux (git version 2.11.1) and on Windows
>> (git version 2.11.1.windows.1):
>>
>> diff --git a/a.natvis b/a.natvis
>> index 7f9bdf5..bc3c090 100644
>> --- a/a.natvis
>> +++ b/a.natvis
>> @@ -19,7 +19,7 @@ test
>>
>>
>> <!-- Non-blank line -->
>> - <Item Name="added var">added_var</Item>
>> + <Item Name="added var">added_vars</Item>
>>
>>
>> <Item Name="var2">var2</Item>
>>
>>> Just to be sure, I tested your regex and again it didn't work.
>>
>> At this point I'm out of ideas, sorry. :( The only way I was able to
>> break it was due to mistyping the extension as "netvis" several times
>> for some reason.
>>
>> René
^ permalink raw reply
* Re: Trying to use xfuncname without success.
From: Samuel Lijin @ 2017-02-08 20:46 UTC (permalink / raw)
To: Jack Adrian Zappa; +Cc: René Scharfe, git-mailing-list
In-Reply-To: <CAKepmagwMeky4jPZ-YFgPsZSsyOZZQ-kJSWV8QGg4cUNu-ZS8Q@mail.gmail.com>
I just put this togther: https://github.com/sxlijin/xfuncname-test
Try cloning and then for any of config1 thru 3,
$ cp configX .git/config
$ git diff HEAD^ -- test.natvis
On Wed, Feb 8, 2017 at 2:42 PM, Jack Adrian Zappa <adrianh.bsc@gmail.com> wrote:
> Thanks Samuel,
>
> So, the question is, what is causing this problem on my system?
>
> Anyone have an idea to help diagnose this problem?
>
> On Wed, Feb 8, 2017 at 3:24 PM, Samuel Lijin <sxlijin@gmail.com> wrote:
>> On Windows 7, it works for me in both CMD and Git Bash:
>>
>> $ git --version
>> git version 2.11.0.windows.3
>>
>> $ git diff HEAD^ --word-diff
>> diff --git a/test.natvis b/test.natvis
>> index 93396ad..1233b8c 100644
>> --- a/test.natvis
>> +++ b/test.natvis
>> @@ -18,6 +18,7 @@ test
>>
>>
>> <!-- Non-blank line -->
>> {+<Item Name="added var">added_var</Item>+}
>>
>>
>> <Item Name="var2">var2</Item>
>>
>> On Wed, Feb 8, 2017 at 12:37 PM, René Scharfe <l.s.r@web.de> wrote:
>>> Am 08.02.2017 um 18:11 schrieb Jack Adrian Zappa:
>>>> Thanks Rene, but you seem to have missed the point. NOTHING is
>>>> working. No matter what I put there, it doesn't seem to get matched.
>>>
>>> I'm not so sure about that. With your example I get this diff without
>>> setting diff.natvis.xfuncname:
>>>
>>> diff --git a/a.natvis b/a.natvis
>>> index 7f9bdf5..bc3c090 100644
>>> --- a/a.natvis
>>> +++ b/a.natvis
>>> @@ -19,7 +19,7 @@ xmlns="http://schemas.microsoft.com/vstudio/debugger/natvis/2010">
>>>
>>>
>>> <!-- Non-blank line -->
>>> - <Item Name="added var">added_var</Item>
>>> + <Item Name="added var">added_vars</Item>
>>>
>>>
>>> <Item Name="var2">var2</Item>
>>>
>>> Note the XML namespace in the hunk header. It's put there by the
>>> default rule because "xmlns" starts at the beginning of the line. Your
>>> diff has nothing there, which means the default rule is not used, i.e.
>>> your user-defined rule is in effect.
>>>
>>> Come to think of it, this line break in the middle of the AutoVisualizer
>>> tab might have been added by your email client unintentionally, so that
>>> we use different test files, which then of course results in different
>>> diffs. Is that the case?
>>>
>>> Anyway, if I run the following two commands:
>>>
>>> $ git config diff.natvis.xfuncname "^[\t ]*<Type[\t ]+Name=\"([^\"]+)\".*$"
>>> $ echo '*.natvis diff=natvis' >.gitattributes
>>>
>>> ... then I get this, both on Linux (git version 2.11.1) and on Windows
>>> (git version 2.11.1.windows.1):
>>>
>>> diff --git a/a.natvis b/a.natvis
>>> index 7f9bdf5..bc3c090 100644
>>> --- a/a.natvis
>>> +++ b/a.natvis
>>> @@ -19,7 +19,7 @@ test
>>>
>>>
>>> <!-- Non-blank line -->
>>> - <Item Name="added var">added_var</Item>
>>> + <Item Name="added var">added_vars</Item>
>>>
>>>
>>> <Item Name="var2">var2</Item>
>>>
>>>> Just to be sure, I tested your regex and again it didn't work.
>>>
>>> At this point I'm out of ideas, sorry. :( The only way I was able to
>>> break it was due to mistyping the extension as "netvis" several times
>>> for some reason.
>>>
>>> René
^ permalink raw reply
* Re: [PATCH] push options: fail properly in the stateless case
From: Stefan Beller @ 2017-02-08 20:51 UTC (permalink / raw)
To: Junio C Hamano; +Cc: git@vger.kernel.org, Jonathan Nieder
In-Reply-To: <xmqqpoiswo0l.fsf@gitster.mtv.corp.google.com>
On Wed, Feb 8, 2017 at 10:11 AM, Junio C Hamano <gitster@pobox.com> wrote:
> Stefan Beller <sbeller@google.com> writes:
>
>> When using non-builtin protocols relying on a transport helper
>> (such as http), push options are not propagated to the helper.
>>
>> Fix this by propagating the push options to the transport helper and
>
> The description up to this point is VERY readable and sensible. But
> that makes the title sound a bit strange.
Yes, the title was there first and then I massaged the commit message
until it was readable. Originally I thought the issue is with stateless
protocols, but that is wrong. The underlying issue is just the transport
helper communication lacking.
> I read it as if it were
> saying "stateless case can never support push-options so fail if the
> caller attempts to use one", but that does not seem to be what is
> going on.
Indeed the subject is wrong.
>
>> adding a test that push options using http fail properly.
>
> Sounds sensible. What end-user visible effect does this fix have?
> IOW, what feature do we use "push-option" for?
The Gerrit world started using push options for having a better git
experience, not needing to navigate to the web UI, e.g.:
# push for review and set me as a reviewer:
git push --push-option reviewer=sbeller@google.com ssh://gerrit...
# same with http, it looks like it worked, but the push option
# never made it to the server
git push --push-option reviewer=sbeller@google.com http://gerrit...
# This patch will make the second command fail, reporting
# the http helper not supporting push options.
>
> Ahh, OK, so you need to describe that there are two issues in order
> to be understood by the readers:
>
> (1) the helper protocol does not propagate push-option
> (2) the http helper is not prepared to handle push-option
>
> You fix (1), and you take advantage of the fact (2) to ensure that
> (1) is fixed in the new test.
>
> With such an understanding, the title makes (sort of) sense and you
> wouldn't have to be asked "what end-user visible effect/benefit does
> this have?"
Your analysis is correct.
>
>> +'option push-option <c-string>::
>> + Transmit this push option.
>> +
>
> There is no "c-string" in the current documentation used or
> defined. The closest thing I found is
>
> ... that field will be quoted in the manner of a C string ...
>
> in git-status page, but I do not think you send the value for an
> push-option after running quote_c_style(), so I am puzzled.
When implementing push options, we discussed that and according to
Documentation/git-push:
The given string must not contain a NUL or LF character.
>
> I'd rather see 'option push-option <string>' as the bullet item, and
> in its description say how arbitrary values (if you allow them, that
> is) can be used, e.g. "Transmit <string> encoded in such and such
> way a the value of the push-option".
okay.
^ permalink raw reply
* [PATCH v2 0/11] reducing resource usage of for_each_alternate_ref
From: Jeff King @ 2017-02-08 20:52 UTC (permalink / raw)
To: git; +Cc: Junio C Hamano
In-Reply-To: <20170124003729.j4ygjcgypdq7hceg@sigill.intra.peff.net>
This is a minor re-roll of the patches from:
http://public-inbox.org/git/20170124003729.j4ygjcgypdq7hceg@sigill.intra.peff.net/
(which got some review, but I don't think was picked up for even 'pu').
I won't repeat the numbers and background from that message, but the
gist of it is that this reduces memory usage significantly when your
alternate has a lot of refs in it.
This version makes two minor changes:
- it drops the save_commit_buffer patch to clone; it's redundant with
what fetch_pack() is doing internally, and I wasn't able to measure
any improvement
- it adds a missing "static" to an internal function
The only other possible change from the review would be sorting the
expected output in the test of the final script. I'm on the fence
whether it is a feature that we expect a particular ordering. It's not
set in stone ,but it _is_ deterministic, and if we change the order, it
might be worth somebody actually noticing.
[01/11]: for_each_alternate_ref: handle failure from real_pathdup()
[02/11]: for_each_alternate_ref: stop trimming trailing slashes
[03/11]: for_each_alternate_ref: use strbuf for path allocation
[04/11]: for_each_alternate_ref: pass name/oid instead of ref struct
[05/11]: for_each_alternate_ref: replace transport code with for-each-ref
[06/11]: fetch-pack: cache results of for_each_alternate_ref
[07/11]: add oidset API
[08/11]: receive-pack: use oidset to de-duplicate .have lines
[09/11]: receive-pack: fix misleading namespace/.have comment
[10/11]: receive-pack: treat namespace .have lines like alternates
[11/11]: receive-pack: avoid duplicates between our refs and alternates
Makefile | 1 +
builtin/receive-pack.c | 41 +++++++++++++++-------------
fetch-pack.c | 48 ++++++++++++++++++++++++++++-----
object.h | 2 +-
oidset.c | 49 ++++++++++++++++++++++++++++++++++
oidset.h | 45 +++++++++++++++++++++++++++++++
t/t5400-send-pack.sh | 38 ++++++++++++++++++++++++++
transport.c | 72 +++++++++++++++++++++++++++++++++++---------------
transport.h | 2 +-
9 files changed, 249 insertions(+), 49 deletions(-)
create mode 100644 oidset.c
create mode 100644 oidset.h
^ permalink raw reply
* [PATCH v2 01/11] for_each_alternate_ref: handle failure from real_pathdup()
From: Jeff King @ 2017-02-08 20:52 UTC (permalink / raw)
To: git; +Cc: Junio C Hamano
In-Reply-To: <20170208205219.twgm5rggovqbepte@sigill.intra.peff.net>
In older versions of git, if real_path() failed to resolve
the alternate object store path, we would die() with an
error. However, since 4ac9006f8 (real_path: have callers use
real_pathdup and strbuf_realpath, 2016-12-12) we use the
real_pathdup() function, which may return NULL. Since we
don't check the return value, we can segfault.
This is hard to trigger in practice, since we check that the
path is accessible before creating the alternate_object_database
struct. But it could be removed racily, or we could see a
transient filesystem error.
We could restore the original behavior by switching back to
xstrdup(real_path()). However, dying is probably not the
best option here. This whole function is best-effort
already; there might not even be a repository around the
shared objects at all. And if the alternate store has gone
away, there are no objects to show.
So let's just quietly return, as we would if we failed to
open "refs/", or if upload-pack failed to start, etc.
Signed-off-by: Jeff King <peff@peff.net>
---
transport.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/transport.c b/transport.c
index d72e08948..9ce0ee96b 100644
--- a/transport.c
+++ b/transport.c
@@ -1222,6 +1222,8 @@ static int refs_from_alternate_cb(struct alternate_object_database *e,
struct alternate_refs_data *cb = data;
other = real_pathdup(e->path);
+ if (!other)
+ return 0;
len = strlen(other);
while (other[len-1] == '/')
--
2.12.0.rc0.371.ga6cf8653b
^ permalink raw reply related
* [PATCH v2 02/11] for_each_alternate_ref: stop trimming trailing slashes
From: Jeff King @ 2017-02-08 20:52 UTC (permalink / raw)
To: git; +Cc: Junio C Hamano
In-Reply-To: <20170208205219.twgm5rggovqbepte@sigill.intra.peff.net>
The real_pathdup() function will have removed extra slashes
for us already (on top of the normalize_path() done when we
created the alternate_object_database struct in the first
place).
Incidentally, this also fixes the case where the path is
just "/", which would read off the start of the array.
That doesn't seem possible to trigger in practice, though,
as link_alt_odb_entry() blindly eats trailing slashes,
including a bare "/".
Signed-off-by: Jeff King <peff@peff.net>
---
transport.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/transport.c b/transport.c
index 9ce0ee96b..6fba9e95b 100644
--- a/transport.c
+++ b/transport.c
@@ -1226,8 +1226,6 @@ static int refs_from_alternate_cb(struct alternate_object_database *e,
return 0;
len = strlen(other);
- while (other[len-1] == '/')
- other[--len] = '\0';
if (len < 8 || memcmp(other + len - 8, "/objects", 8))
goto out;
/* Is this a git repository with refs? */
--
2.12.0.rc0.371.ga6cf8653b
^ permalink raw reply related
* Re: [PATCH] dir: avoid allocation in fill_directory()
From: Brandon Williams @ 2017-02-08 19:54 UTC (permalink / raw)
To: Duy Nguyen; +Cc: René Scharfe, Git List
In-Reply-To: <CACsJy8CE-cyTZHZZhvhdsNau7iSqBci1BdUqDYvxoE5odV2SBA@mail.gmail.com>
On 02/08, Duy Nguyen wrote:
> On Wed, Feb 8, 2017 at 5:04 AM, René Scharfe <l.s.r@web.de> wrote:
> > Pass the match member of the first pathspec item directly to
> > read_directory() instead of using common_prefix() to duplicate it first,
> > thus avoiding memory duplication, strlen(3) and free(3).
>
> How about killing common_prefix()? There are two other callers in
> ls-files.c and commit.c and it looks safe to do (but I didn't look
> very hard).
>
> > diff --git a/dir.c b/dir.c
> > index 65c3e681b8..4541f9e146 100644
> > --- a/dir.c
> > +++ b/dir.c
> > @@ -174,20 +174,19 @@ char *common_prefix(const struct pathspec *pathspec)
> >
> > int fill_directory(struct dir_struct *dir, const struct pathspec *pathspec)
> > {
> > - char *prefix;
> > + const char *prefix;
> > size_t prefix_len;
> >
> > /*
> > * Calculate common prefix for the pathspec, and
> > * use that to optimize the directory walk
> > */
> > - prefix = common_prefix(pathspec);
> > - prefix_len = prefix ? strlen(prefix) : 0;
> > + prefix_len = common_prefix_len(pathspec);
> > + prefix = prefix_len ? pathspec->items[0].match : "";
>
> There's a subtle difference. Before the patch, prefix[prefix_len] is
> NUL. After the patch, it's not always true. If some code (incorrectly)
> depends on that, this patch exposes it. I had a look inside
> read_directory() though and it looks like no such code exists. So, all
> good.
Yeah I had the exact same thought when looking at this, but I agree
everything looks fine. And if something does indeed depend on prefix
having a \0 at prefix_len then this will allow us to more easily find
the bug and fix it.
>
> >
> > /* Read the directory and prune it */
> > read_directory(dir, prefix, prefix_len, pathspec);
> >
> > - free(prefix);
> > return prefix_len;
> > }
> --
> Duy
--
Brandon Williams
^ permalink raw reply
* [PATCH v2 04/11] for_each_alternate_ref: pass name/oid instead of ref struct
From: Jeff King @ 2017-02-08 20:52 UTC (permalink / raw)
To: git; +Cc: Junio C Hamano
In-Reply-To: <20170208205219.twgm5rggovqbepte@sigill.intra.peff.net>
Breaking down the fields in the interface makes it easier to
change the backend of for_each_alternate_ref to something
that doesn't use "struct ref" internally.
The only field that callers actually look at is the oid,
anyway. The refname is kept in the interface as a plausible
thing for future code to want.
Signed-off-by: Jeff King <peff@peff.net>
---
builtin/receive-pack.c | 6 ++++--
fetch-pack.c | 12 ++++++++----
transport.c | 2 +-
transport.h | 2 +-
4 files changed, 14 insertions(+), 8 deletions(-)
diff --git a/builtin/receive-pack.c b/builtin/receive-pack.c
index 1dbb8a069..d21332d9e 100644
--- a/builtin/receive-pack.c
+++ b/builtin/receive-pack.c
@@ -277,10 +277,12 @@ static int show_one_alternate_sha1(const unsigned char sha1[20], void *unused)
return 0;
}
-static void collect_one_alternate_ref(const struct ref *ref, void *data)
+static void collect_one_alternate_ref(const char *refname,
+ const struct object_id *oid,
+ void *data)
{
struct sha1_array *sa = data;
- sha1_array_append(sa, ref->old_oid.hash);
+ sha1_array_append(sa, oid->hash);
}
static void write_head_info(void)
diff --git a/fetch-pack.c b/fetch-pack.c
index 601f0779a..54f84c573 100644
--- a/fetch-pack.c
+++ b/fetch-pack.c
@@ -253,9 +253,11 @@ static void send_request(struct fetch_pack_args *args,
write_or_die(fd, buf->buf, buf->len);
}
-static void insert_one_alternate_ref(const struct ref *ref, void *unused)
+static void insert_one_alternate_ref(const char *refname,
+ const struct object_id *oid,
+ void *unused)
{
- rev_list_insert_ref(NULL, ref->old_oid.hash);
+ rev_list_insert_ref(NULL, oid->hash);
}
#define INITIAL_FLUSH 16
@@ -619,9 +621,11 @@ static void filter_refs(struct fetch_pack_args *args,
*refs = newlist;
}
-static void mark_alternate_complete(const struct ref *ref, void *unused)
+static void mark_alternate_complete(const char *refname,
+ const struct object_id *oid,
+ void *unused)
{
- mark_complete(ref->old_oid.hash);
+ mark_complete(oid->hash);
}
static int everything_local(struct fetch_pack_args *args,
diff --git a/transport.c b/transport.c
index 6b7847131..c87147046 100644
--- a/transport.c
+++ b/transport.c
@@ -1238,7 +1238,7 @@ static int refs_from_alternate_cb(struct alternate_object_database *e,
for (extra = transport_get_remote_refs(transport);
extra;
extra = extra->next)
- cb->fn(extra, cb->data);
+ cb->fn(extra->name, &extra->old_oid, cb->data);
transport_disconnect(transport);
out:
strbuf_release(&path);
diff --git a/transport.h b/transport.h
index e597b31b3..bc5571574 100644
--- a/transport.h
+++ b/transport.h
@@ -255,6 +255,6 @@ int transport_refs_pushed(struct ref *ref);
void transport_print_push_status(const char *dest, struct ref *refs,
int verbose, int porcelain, unsigned int *reject_reasons);
-typedef void alternate_ref_fn(const struct ref *, void *);
+typedef void alternate_ref_fn(const char *refname, const struct object_id *oid, void *);
extern void for_each_alternate_ref(alternate_ref_fn, void *);
#endif
--
2.12.0.rc0.371.ga6cf8653b
^ permalink raw reply related
* [PATCH v2 03/11] for_each_alternate_ref: use strbuf for path allocation
From: Jeff King @ 2017-02-08 20:52 UTC (permalink / raw)
To: git; +Cc: Junio C Hamano
In-Reply-To: <20170208205219.twgm5rggovqbepte@sigill.intra.peff.net>
We have a string with ".../objects" pointing to the
alternate object store, and overwrite bits of it to look at
other paths in the (potential) git repository holding it.
This works because the only path we care about is "refs",
which is shorter than "objects".
Using a strbuf to hold the path lets us get rid of some
magic numbers, and makes it more obvious that the memory
operations are safe.
Signed-off-by: Jeff King <peff@peff.net>
---
transport.c | 28 ++++++++++++++--------------
1 file changed, 14 insertions(+), 14 deletions(-)
diff --git a/transport.c b/transport.c
index 6fba9e95b..6b7847131 100644
--- a/transport.c
+++ b/transport.c
@@ -1214,34 +1214,34 @@ struct alternate_refs_data {
static int refs_from_alternate_cb(struct alternate_object_database *e,
void *data)
{
- char *other;
- size_t len;
+ struct strbuf path = STRBUF_INIT;
+ size_t base_len;
struct remote *remote;
struct transport *transport;
const struct ref *extra;
struct alternate_refs_data *cb = data;
- other = real_pathdup(e->path);
- if (!other)
- return 0;
- len = strlen(other);
-
- if (len < 8 || memcmp(other + len - 8, "/objects", 8))
+ if (!strbuf_realpath(&path, e->path, 0))
+ goto out;
+ if (!strbuf_strip_suffix(&path, "/objects"))
goto out;
+ base_len = path.len;
+
/* Is this a git repository with refs? */
- memcpy(other + len - 8, "/refs", 6);
- if (!is_directory(other))
+ strbuf_addstr(&path, "/refs");
+ if (!is_directory(path.buf))
goto out;
- other[len - 8] = '\0';
- remote = remote_get(other);
- transport = transport_get(remote, other);
+ strbuf_setlen(&path, base_len);
+
+ remote = remote_get(path.buf);
+ transport = transport_get(remote, path.buf);
for (extra = transport_get_remote_refs(transport);
extra;
extra = extra->next)
cb->fn(extra, cb->data);
transport_disconnect(transport);
out:
- free(other);
+ strbuf_release(&path);
return 0;
}
--
2.12.0.rc0.371.ga6cf8653b
^ permalink raw reply related
* [PATCH v2 05/11] for_each_alternate_ref: replace transport code with for-each-ref
From: Jeff King @ 2017-02-08 20:53 UTC (permalink / raw)
To: git; +Cc: Junio C Hamano
In-Reply-To: <20170208205219.twgm5rggovqbepte@sigill.intra.peff.net>
The current method for getting the refs from an alternate is
to run upload-pack in the alternate and parse its output
using the normal transport code. This works and is
reasonably short, but it has a very bad memory footprint
when there are a lot of refs in the alternate. There are two
problems:
1. It reads in all of the refs before passing any back to
us. Which means that our peak memory usage has to store
every ref (including duplicates for peeled variants),
even if our callback could determine that some are not
interesting (e.g., because they point to the same sha1
as another ref).
2. It allocates a "struct ref" for each one. Among other
things, this contains 3 separate 20-byte oids, along
with the name and various pointers. That can add up,
especially if the callback is only interested in the
sha1 (which it can store in a sha1_array as just 20
bytes).
On a particularly pathological case, where the alternate had
over 80 million refs pointing to only around 60,000 unique
objects, the peak heap usage of "git clone --reference" grew
to over 25GB.
This patch instead calls git-for-each-ref in the alternate
repository, and passes each line to the callback as we read
it. That drops the peak heap of the same command to 50MB.
I considered and rejected a few alternatives.
We could read all of the refs in the alternate using our own
ref code, just as we do with submodules. However, as memory
footprint is one of the concerns here, we want to avoid
loading those refs into our own memory as a whole.
It's possible that this will be a better technique in the
future when the ref code can more easily iterate without
loading all of packed-refs into memory.
Another option is to keep calling upload-pack, and just
parse its output ourselves in a streaming fashion. Besides
for-each-ref being simpler (we get to define the format
ourselves, and don't have to deal with speaking the git
protocol), it's more flexible for possible future changes.
For instance, it might be useful for the caller to be able
to limit the set of "interesting" alternate refs. The
motivating example is one where many "forks" of a particular
repository share object storage, and the shared storage has
refs for each fork (which is why so many of the refs are
duplicates; each fork has the same tags). A plausible
future optimization would be to ask for the alternate refs
for just _one_ fork (if you had some out-of-band way of
knowing which was the most interesting or important for the
current operation).
Similarly, no callbacks actually care about the symref value
of alternate refs, and as before, this patch ignores them
entirely. However, if we wanted to add them, for-each-ref's
"%(symref)" is going to be more flexible than upload-pack,
because the latter only handles the HEAD symref due to
historical constraints.
There is one potential downside, though: unlike upload-pack,
our for-each-ref command doesn't report the peeled value of
refs. The existing code calls the alternate_ref_fn callback
twice for tags: once for the tag, and once for the peeled
value with the refname set to "ref^{}".
For the callers in fetch-pack, this doesn't matter at all.
We immediately peel each tag down to a commit either way (so
there's a slight improvement, as do not bother passing the
redundant data over the pipe). For the caller in
receive-pack, it means we will not advertise the peeled
values of tags in our alternate. However, we also don't
advertise peeled values for our _own_ tags, so this is
actually making things more consistent.
It's unclear whether receive-pack advertising peeled values
is a win or not. On one hand, giving more information to the
other side may let it omit some objects from the push. On
the other hand, for tags which both sides have, they simply
bloat the advertisement. The upload-pack advertisement of
git.git is about 30% larger than the receive-pack
advertisement due to its peeled information.
This patch omits the peeled information from
for_each_alternate_ref entirely, and leaves it up to the
caller whether they want to dig up the information.
Signed-off-by: Jeff King <peff@peff.net>
---
transport.c | 48 ++++++++++++++++++++++++++++++++++++++----------
1 file changed, 38 insertions(+), 10 deletions(-)
diff --git a/transport.c b/transport.c
index c87147046..48864b3a9 100644
--- a/transport.c
+++ b/transport.c
@@ -1206,6 +1206,42 @@ char *transport_anonymize_url(const char *url)
return xstrdup(url);
}
+static void read_alternate_refs(const char *path,
+ alternate_ref_fn *cb,
+ void *data)
+{
+ struct child_process cmd = CHILD_PROCESS_INIT;
+ struct strbuf line = STRBUF_INIT;
+ FILE *fh;
+
+ cmd.git_cmd = 1;
+ argv_array_pushf(&cmd.args, "--git-dir=%s", path);
+ argv_array_push(&cmd.args, "for-each-ref");
+ argv_array_push(&cmd.args, "--format=%(objectname) %(refname)");
+ cmd.env = local_repo_env;
+ cmd.out = -1;
+
+ if (start_command(&cmd))
+ return;
+
+ fh = xfdopen(cmd.out, "r");
+ while (strbuf_getline_lf(&line, fh) != EOF) {
+ struct object_id oid;
+
+ if (get_oid_hex(line.buf, &oid) ||
+ line.buf[GIT_SHA1_HEXSZ] != ' ') {
+ warning("invalid line while parsing alternate refs: %s",
+ line.buf);
+ break;
+ }
+
+ cb(line.buf + GIT_SHA1_HEXSZ + 1, &oid, data);
+ }
+
+ fclose(fh);
+ finish_command(&cmd);
+}
+
struct alternate_refs_data {
alternate_ref_fn *fn;
void *data;
@@ -1216,9 +1252,6 @@ static int refs_from_alternate_cb(struct alternate_object_database *e,
{
struct strbuf path = STRBUF_INIT;
size_t base_len;
- struct remote *remote;
- struct transport *transport;
- const struct ref *extra;
struct alternate_refs_data *cb = data;
if (!strbuf_realpath(&path, e->path, 0))
@@ -1233,13 +1266,8 @@ static int refs_from_alternate_cb(struct alternate_object_database *e,
goto out;
strbuf_setlen(&path, base_len);
- remote = remote_get(path.buf);
- transport = transport_get(remote, path.buf);
- for (extra = transport_get_remote_refs(transport);
- extra;
- extra = extra->next)
- cb->fn(extra->name, &extra->old_oid, cb->data);
- transport_disconnect(transport);
+ read_alternate_refs(path.buf, cb->fn, cb->data);
+
out:
strbuf_release(&path);
return 0;
--
2.12.0.rc0.371.ga6cf8653b
^ permalink raw reply related
* [PATCH v2 08/11] receive-pack: use oidset to de-duplicate .have lines
From: Jeff King @ 2017-02-08 20:53 UTC (permalink / raw)
To: git; +Cc: Junio C Hamano
In-Reply-To: <20170208205219.twgm5rggovqbepte@sigill.intra.peff.net>
If you have an alternate object store with a very large
number of refs, the peak memory usage of the sha1_array can
grow high, even if most of them are duplicates that end up
not being printed at all.
The similar for_each_alternate_ref() code-paths in
fetch-pack solve this by using flags in "struct object" to
de-duplicate (and so are relying on obj_hash at the core).
But we don't have a "struct object" at all in this case. We
could call lookup_unknown_object() to get one, but if our
goal is reducing memory footprint, it's not great:
- an unknown object is as large as the largest object type
(a commit), which is bigger than an oidset entry
- we can free the memory after our ref advertisement, but
"struct object" entries persist forever (and the
receive-pack may hang around for a long time, as the
bottleneck is often client upload bandwidth).
So let's use an oidset. Note that unlike a sha1-array it
doesn't sort the output as a side effect. However, our
output is at least stable, because for_each_alternate_ref()
will give us the sha1s in ref-sorted order.
In one particularly pathological case with an alternate that
has 60,000 unique refs out of 80 million total, this reduced
the peak heap usage of "git receive-pack . </dev/null" from
13GB to 14MB.
Signed-off-by: Jeff King <peff@peff.net>
---
builtin/receive-pack.c | 26 ++++++++++++--------------
1 file changed, 12 insertions(+), 14 deletions(-)
diff --git a/builtin/receive-pack.c b/builtin/receive-pack.c
index d21332d9e..a4926fcfb 100644
--- a/builtin/receive-pack.c
+++ b/builtin/receive-pack.c
@@ -21,6 +21,7 @@
#include "sigchain.h"
#include "fsck.h"
#include "tmp-objdir.h"
+#include "oidset.h"
static const char * const receive_pack_usage[] = {
N_("git receive-pack <git-dir>"),
@@ -271,27 +272,24 @@ static int show_ref_cb(const char *path_full, const struct object_id *oid,
return 0;
}
-static int show_one_alternate_sha1(const unsigned char sha1[20], void *unused)
+static void show_one_alternate_ref(const char *refname,
+ const struct object_id *oid,
+ void *data)
{
- show_ref(".have", sha1);
- return 0;
-}
+ struct oidset *seen = data;
-static void collect_one_alternate_ref(const char *refname,
- const struct object_id *oid,
- void *data)
-{
- struct sha1_array *sa = data;
- sha1_array_append(sa, oid->hash);
+ if (oidset_insert(seen, oid))
+ return;
+
+ show_ref(".have", oid->hash);
}
static void write_head_info(void)
{
- struct sha1_array sa = SHA1_ARRAY_INIT;
+ static struct oidset seen = OIDSET_INIT;
- for_each_alternate_ref(collect_one_alternate_ref, &sa);
- sha1_array_for_each_unique(&sa, show_one_alternate_sha1, NULL);
- sha1_array_clear(&sa);
+ for_each_alternate_ref(show_one_alternate_ref, &seen);
+ oidset_clear(&seen);
for_each_ref(show_ref_cb, NULL);
if (!sent_capabilities)
show_ref("capabilities^{}", null_sha1);
--
2.12.0.rc0.371.ga6cf8653b
^ permalink raw reply related
* [PATCH v2 07/11] add oidset API
From: Jeff King @ 2017-02-08 20:53 UTC (permalink / raw)
To: git; +Cc: Junio C Hamano
In-Reply-To: <20170208205219.twgm5rggovqbepte@sigill.intra.peff.net>
This is similar to many of our uses of sha1-array, but it
overcomes one limitation of a sha1-array: when you are
de-duplicating a large input with relatively few unique
entries, sha1-array uses 20 bytes per non-unique entry.
Whereas this set will use memory linear in the number of
unique entries (albeit a few more than 20 bytes due to
hashmap overhead).
Signed-off-by: Jeff King <peff@peff.net>
---
Makefile | 1 +
oidset.c | 49 +++++++++++++++++++++++++++++++++++++++++++++++++
oidset.h | 45 +++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 95 insertions(+)
create mode 100644 oidset.c
create mode 100644 oidset.h
diff --git a/Makefile b/Makefile
index 8e4081e06..a5433978e 100644
--- a/Makefile
+++ b/Makefile
@@ -781,6 +781,7 @@ LIB_OBJS += notes-cache.o
LIB_OBJS += notes-merge.o
LIB_OBJS += notes-utils.o
LIB_OBJS += object.o
+LIB_OBJS += oidset.o
LIB_OBJS += pack-bitmap.o
LIB_OBJS += pack-bitmap-write.o
LIB_OBJS += pack-check.o
diff --git a/oidset.c b/oidset.c
new file mode 100644
index 000000000..ac169f05d
--- /dev/null
+++ b/oidset.c
@@ -0,0 +1,49 @@
+#include "cache.h"
+#include "oidset.h"
+
+struct oidset_entry {
+ struct hashmap_entry hash;
+ struct object_id oid;
+};
+
+static int oidset_hashcmp(const void *va, const void *vb,
+ const void *vkey)
+{
+ const struct oidset_entry *a = va, *b = vb;
+ const struct object_id *key = vkey;
+ return oidcmp(&a->oid, key ? key : &b->oid);
+}
+
+int oidset_contains(const struct oidset *set, const struct object_id *oid)
+{
+ struct hashmap_entry key;
+
+ if (!set->map.cmpfn)
+ return 0;
+
+ hashmap_entry_init(&key, sha1hash(oid->hash));
+ return !!hashmap_get(&set->map, &key, oid);
+}
+
+int oidset_insert(struct oidset *set, const struct object_id *oid)
+{
+ struct oidset_entry *entry;
+
+ if (!set->map.cmpfn)
+ hashmap_init(&set->map, oidset_hashcmp, 0);
+
+ if (oidset_contains(set, oid))
+ return 1;
+
+ entry = xmalloc(sizeof(*entry));
+ hashmap_entry_init(&entry->hash, sha1hash(oid->hash));
+ oidcpy(&entry->oid, oid);
+
+ hashmap_add(&set->map, entry);
+ return 0;
+}
+
+void oidset_clear(struct oidset *set)
+{
+ hashmap_free(&set->map, 1);
+}
diff --git a/oidset.h b/oidset.h
new file mode 100644
index 000000000..b7eaab5b8
--- /dev/null
+++ b/oidset.h
@@ -0,0 +1,45 @@
+#ifndef OIDSET_H
+#define OIDSET_H
+
+/**
+ * This API is similar to sha1-array, in that it maintains a set of object ids
+ * in a memory-efficient way. The major differences are:
+ *
+ * 1. It uses a hash, so we can do online duplicate removal, rather than
+ * sort-and-uniq at the end. This can reduce memory footprint if you have
+ * a large list of oids with many duplicates.
+ *
+ * 2. The per-unique-oid memory footprint is slightly higher due to hash
+ * table overhead.
+ */
+
+/**
+ * A single oidset; should be zero-initialized (or use OIDSET_INIT).
+ */
+struct oidset {
+ struct hashmap map;
+};
+
+#define OIDSET_INIT { { NULL } }
+
+/**
+ * Returns true iff `set` contains `oid`.
+ */
+int oidset_contains(const struct oidset *set, const struct object_id *oid);
+
+/**
+ * Insert the oid into the set; a copy is made, so "oid" does not need
+ * to persist after this function is called.
+ *
+ * Returns 1 if the oid was already in the set, 0 otherwise. This can be used
+ * to perform an efficient check-and-add.
+ */
+int oidset_insert(struct oidset *set, const struct object_id *oid);
+
+/**
+ * Remove all entries from the oidset, freeing any resources associated with
+ * it.
+ */
+void oidset_clear(struct oidset *set);
+
+#endif /* OIDSET_H */
--
2.12.0.rc0.371.ga6cf8653b
^ permalink raw reply related
* [PATCH v2 06/11] fetch-pack: cache results of for_each_alternate_ref
From: Jeff King @ 2017-02-08 20:53 UTC (permalink / raw)
To: git; +Cc: Junio C Hamano
In-Reply-To: <20170208205219.twgm5rggovqbepte@sigill.intra.peff.net>
We may run for_each_alternate_ref() twice, once in
find_common() and once in everything_local(). This operation
can be expensive, because it involves running a sub-process
which must freshly load all of the alternate's refs from
disk.
Let's cache and reuse the results between the two calls. We
can make some optimizations based on the particular use
pattern in fetch-pack to keep our memory usage down.
The first is that we only care about the sha1s, not the refs
themselves. So it's OK to store only the sha1s, and to
suppress duplicates. The natural fit would therefore be a
sha1_array.
However, sha1_array's de-duplication happens only after it
has read and sorted all entries. It still stores each
duplicate. For an alternate with a large number of refs
pointing to the same commits, this is a needless expense.
Instead, we'd prefer to eliminate duplicates before putting
them in the cache, which implies using a hash. We can
further note that fetch-pack will call parse_object() on
each alternate sha1. We can therefore keep our cache as a
set of pointers to "struct object". That gives us a place to
put our "already seen" bit with an optimized hash lookup.
And as a bonus, the object stores the sha1 for us, so
pointer-to-object is all we need.
There are two extra optimizations I didn't do here:
- we actually store an array of pointer-to-object.
Technically we could just walk the obj_hash table
looking for entries with the ALTERNATE flag set (because
our use case doesn't care about the order here).
But that hash table may be mostly composed of
non-ALTERNATE entries, so we'd waste time walking over
them. So it would be a slight win in memory use, but a
loss in CPU.
- the items we pull out of the cache are actual "struct
object"s, but then we feed "obj->sha1" to our
sub-functions, which promptly call parse_object().
This second parse is cheap, because it starts with
lookup_object() and will bail immediately when it sees
we've already parsed the object. We could save the extra
hash lookup, but it would involve refactoring the
functions we call. It may or may not be worth the
trouble.
Signed-off-by: Jeff King <peff@peff.net>
---
fetch-pack.c | 52 ++++++++++++++++++++++++++++++++++++++++++----------
object.h | 2 +-
2 files changed, 43 insertions(+), 11 deletions(-)
diff --git a/fetch-pack.c b/fetch-pack.c
index 54f84c573..e0f5d5ce8 100644
--- a/fetch-pack.c
+++ b/fetch-pack.c
@@ -35,6 +35,7 @@ static const char *alternate_shallow_file;
#define COMMON_REF (1U << 2)
#define SEEN (1U << 3)
#define POPPED (1U << 4)
+#define ALTERNATE (1U << 5)
static int marked;
@@ -67,6 +68,41 @@ static inline void print_verbose(const struct fetch_pack_args *args,
fputc('\n', stderr);
}
+struct alternate_object_cache {
+ struct object **items;
+ size_t nr, alloc;
+};
+
+static void cache_one_alternate(const char *refname,
+ const struct object_id *oid,
+ void *vcache)
+{
+ struct alternate_object_cache *cache = vcache;
+ struct object *obj = parse_object(oid->hash);
+
+ if (!obj || (obj->flags & ALTERNATE))
+ return;
+
+ obj->flags |= ALTERNATE;
+ ALLOC_GROW(cache->items, cache->nr + 1, cache->alloc);
+ cache->items[cache->nr++] = obj;
+}
+
+static void for_each_cached_alternate(void (*cb)(struct object *))
+{
+ static int initialized;
+ static struct alternate_object_cache cache;
+ size_t i;
+
+ if (!initialized) {
+ for_each_alternate_ref(cache_one_alternate, &cache);
+ initialized = 1;
+ }
+
+ for (i = 0; i < cache.nr; i++)
+ cb(cache.items[i]);
+}
+
static void rev_list_push(struct commit *commit, int mark)
{
if (!(commit->object.flags & mark)) {
@@ -253,11 +289,9 @@ static void send_request(struct fetch_pack_args *args,
write_or_die(fd, buf->buf, buf->len);
}
-static void insert_one_alternate_ref(const char *refname,
- const struct object_id *oid,
- void *unused)
+static void insert_one_alternate_object(struct object *obj)
{
- rev_list_insert_ref(NULL, oid->hash);
+ rev_list_insert_ref(NULL, obj->oid.hash);
}
#define INITIAL_FLUSH 16
@@ -300,7 +334,7 @@ static int find_common(struct fetch_pack_args *args,
marked = 1;
for_each_ref(rev_list_insert_ref_oid, NULL);
- for_each_alternate_ref(insert_one_alternate_ref, NULL);
+ for_each_cached_alternate(insert_one_alternate_object);
fetching = 0;
for ( ; refs ; refs = refs->next) {
@@ -621,11 +655,9 @@ static void filter_refs(struct fetch_pack_args *args,
*refs = newlist;
}
-static void mark_alternate_complete(const char *refname,
- const struct object_id *oid,
- void *unused)
+static void mark_alternate_complete(struct object *obj)
{
- mark_complete(oid->hash);
+ mark_complete(obj->oid.hash);
}
static int everything_local(struct fetch_pack_args *args,
@@ -661,7 +693,7 @@ static int everything_local(struct fetch_pack_args *args,
if (!args->deepen) {
for_each_ref(mark_complete_oid, NULL);
- for_each_alternate_ref(mark_alternate_complete, NULL);
+ for_each_cached_alternate(mark_alternate_complete);
commit_list_sort_by_date(&complete);
if (cutoff)
mark_recent_complete_commits(args, cutoff);
diff --git a/object.h b/object.h
index 614a00675..f52957dcb 100644
--- a/object.h
+++ b/object.h
@@ -29,7 +29,7 @@ struct object_array {
/*
* object flag allocation:
* revision.h: 0---------10 26
- * fetch-pack.c: 0---4
+ * fetch-pack.c: 0---5
* walker.c: 0-2
* upload-pack.c: 4 11----------------19
* builtin/blame.c: 12-13
--
2.12.0.rc0.371.ga6cf8653b
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox