Hi Dave,
On Sun, 3 Jul 2016, Johannes Schindelin wrote:
> On Sun, 3 Jul 2016, David Turner wrote:
>
> > This addresses comments on v13:
> > removed unnecessary no_mmap ifdef
> > add an ifdef in unix-socket
> > OS X fix for select()
> > test improvement
>
> Thanks.
>
> Would you mind re-sending
Signed-off-by: David Turner
---
.mailmap | 1 +
1 file changed, 1 insertion(+)
diff --git a/.mailmap b/.mailmap
index e5b4126..edcae89 100644
--- a/.mailmap
+++ b/.mailmap
@@ -46,6 +46,7 @@ David D. Kilzer
David Kågedal
David Reiss
David S. Miller
+David Turner
Deskin Miller
Dirk Sü
From: David Turner
Add a configuration variable, indexhelper.exitAfter, which provides a
default time to keep the index-helper alive. This is useful with
indexhelper.autorun; some users will want to keep the
automatically-run index-helper alive across their lunch break and will
thus set indexhel
On Mon, Jul 04, 2016 at 08:44:39AM +0900, Mike Hommey wrote:
> + /* Ensure a pack window on the data before that, otherwise,
> + * use_pack() may try to create a window that overlaps with
> + * this one, and that won't work because it won't be complete.
> */
The are many ways in which fast-import ends up calling gfi_unpack_entry,
and fery few work-arounds. I've patched fast-import for it to be smarter
in corner cases, allowing some additional work-arounds, but it's just
too easy to fall on gfi_unpack_entry again, so I abandonned that path.
The problem
= SUMMARY ==
My public git.git is available here[1]. I regularly keep pushing my work so
anyone interested can track me there. Feel free to participate in the
discussions going on PRs with my mentors. Your comments are valuable.
On Fri, Jul 1, 2016 at 1:54 AM, Ben Peart wrote:
> I've found (at least on Windows) that as the repo size gets larger, the
> time to read the index becomes a much smaller percentage of the overall
> time. I just captured some perf traces of git status on a large repo we
> have. Of that, 92.5% wa
Hi Dave,
On Sun, 3 Jul 2016, David Turner wrote:
> This addresses comments on v13:
> removed unnecessary no_mmap ifdef
> add an ifdef in unix-socket
> OS X fix for select()
> test improvement
Thanks.
Would you mind re-sending 20 & 21, they seem to have gotten lost. Or is
there a public reposito
On 29 June 2016 at 08:49, Andy Falanga (afalanga) wrote:
> Is there some sort of strange file caching that happening when
> make starts that, although the local db is updated, I don't get what I'm
> after?
I don't have time to look at your git issue, but I write this quick
note just in case it
From: Nguyễn Thái Ngọc Duy
The previous patch has the logic to clear bits in 'WAMA' bitmap. This
patch has logic to set bits as told by watchman. The missing bit,
_using_ these bits, are not here yet.
A lot of this code is written by David Turner originally, mostly from
[1]. I'm just copying and
From: David Turner
Add a new command (and command-line arg) to allow index-helpers to
exit cleanly.
This is mainly useful for tests.
Signed-off-by: David Turner
Signed-off-by: Junio C Hamano
---
Documentation/git-index-helper.txt | 3 +++
index-helper.c | 31
From: Nguyễn Thái Ngọc Duy
The extension contains a bitmap, one bit for each entry in the
index. If the n-th bit is zero, the n-th entry is considered
unchanged, we can ce_mark_uptodate() it without refreshing. If the bit
is non-zero and we found out the corresponding file is clean after
refresh,
From: Nguyễn Thái Ngọc Duy
Signed-off-by: Nguyễn Thái Ngọc Duy
Signed-off-by: David Turner
Signed-off-by: Junio C Hamano
---
Documentation/git-index-helper.txt | 3 +++
Documentation/git-update-index.txt | 6 ++
builtin/update-index.c | 15 +++
3 files changed, 2
From: Nguyễn Thái Ngọc Duy
All the known heavy code blocks are measured (except object database
access). This should help identify if an optimization is effective or
not. An unoptimized git-status would give something like below (92% of
time is accounted). To sum up the effort of making git scale
From: Nguyễn Thái Ngọc Duy
Watchman is hidden behind index-helper. Before git tries to read the
index from shm, it notifies index-helper through the socket and waits
for index-helper to prepare a file for sharing memory (with
MAP_SHARED). index-helper then contacts watchman, updates 'WAMA'
extens
From: David Turner
Make git checkout (and other unpack_tree operations) preserve the
untracked cache and watchman status. This is valuable for two reasons:
1. Often, an unpack_tree operation will not touch large parts of the
working tree, and thus most of the untracked cache will continue to be
From: David Turner
For installations that have centrally-managed configuration, it's
easier to set a config once than to run update-index on every
repository.
Signed-off-by: David Turner
Signed-off-by: Junio C Hamano
---
.gitignore| 1 +
Documentation/config.txt | 4
From: Nguyễn Thái Ngọc Duy
We detach after creating and opening the socket, because otherwise
we might return control to the shell before index-helper is ready to
accept commands. This might lead to flaky tests.
Signed-off-by: Nguyễn Thái Ngọc Duy
Signed-off-by: David Turner
Signed-off-by: Ju
From: David Turner
Instead of writing warnings to stderr, write them to a log. Later, we'll
probably be daemonized, so writing to stderr will be pointless.
Signed-off-by: David Turner
Signed-off-by: Junio C Hamano
---
Documentation/git-index-helper.txt | 3 +++
index-helper.c
From: David Turner
Introduce a new config option, indexhelper.autorun, to automatically
run git index-helper before starting up a builtin git command. This
enables users to keep index-helper running without manual
intervention.
Signed-off-by: David Turner
Signed-off-by: Junio C Hamano
---
Do
From: David Turner
Signed-off-by: David Turner
Signed-off-by: Junio C Hamano
---
index-helper.c | 7 +++
t/t7900-index-helper.sh | 9 +
2 files changed, 16 insertions(+)
diff --git a/index-helper.c b/index-helper.c
index 943197e..feb09fb 100644
--- a/index-helper.c
+++ b/
From: David Turner
Soon, we'll want to automatically start index-helper, so we need
a mode that silently exits if it can't start up (either because
it's not in a git dir, or because another one is already running).
Signed-off-by: David Turner
Signed-off-by: Junio C Hamano
---
Documentation/gi
From: Nguyễn Thái Ngọc Duy
Later, we will introduce git index-helper to share this memory with
other git processes.
We only unmap it when we discard the index (although the kernel may of
course choose to page it out).
Signed-off-by: Nguyễn Thái Ngọc Duy
Signed-off-by: David Turner
Signed-off-
From: Nguyễn Thái Ngọc Duy
Instead of reading the index from disk and worrying about disk
corruption, the index is cached in memory (memory bit-flips happen
too, but hopefully less often). The result is faster read. Read time
is reduced by 70%.
The biggest gain is not having to verify the traili
From: David Turner
packet_write calls write_or_die, which dies with a sigpipe even if
calling code has explicitly blocked that signal.
Add packet_write_gently and packet_flush_gently, which don't. Soon,
we will use this for communication with git index-helper, which, being
merely an optimizatio
From: Nguyễn Thái Ngọc Duy
This allows signal handlers and atexit functions to realize this
situation and not clean up.
Signed-off-by: Nguyễn Thái Ngọc Duy
Signed-off-by: David Turner
Signed-off-by: Junio C Hamano
---
builtin/gc.c | 2 +-
cache.h | 2 +-
daemon.c | 2 +-
setup.c
From: Nguyễn Thái Ngọc Duy
There are "holes" in the index-helper approach because the shared
memory is not verified again by git. If $USER is compromised, shared
memory could be modified. But anyone who could do this could already
modify $GIT_DIR/index. A more realistic risk is some bugs in
index
This addresses comments on v13:
removed unnecessary no_mmap ifdef
add an ifdef in unix-socket
OS X fix for select()
test improvement
Thanks to all for suggestions.
David Turner (10):
pkt-line: add gentle version of packet_write
index-helper: log warnings
unpack-trees: preserve index extensi
From: Nguyễn Thái Ngọc Duy
This keeps #ifdef at the callee instead of caller, it's less messier.
The caller in question is in read-cache.c which, unlike other
unix-socket callers so far, is always built regardless of unix socket
support. No extra handling (for ENOSYS) is needed because in this
b
29 matches
Mail list logo
