On Thu, Aug 2, 2012 at 6:01 PM, Thomas Gummerer wrote:
> Series of patches to introduce the index version 5 file format. This
> series does not include any fancy stuff like partial loading or partial
> writing yet, though it's possible to do that with the new format.
I applied the series on top o
On Wed, Aug 01, 2012 at 10:52:34AM -0700, Junio C Hamano wrote:
> Chris Webb writes:
>
> [summary: this, when 59a8fde does not have any commit log message,
> refuses to commit]
>
Thanks for CC'ing me on this. I'm on vacation currently, but will look at this
in detail as soon as I'm back next we
On 2012.8.2 2:51 PM, Eric Wong wrote:
> svn_path_canonicalize() may be accessible in some versions of SVN,
> but it'll return undef.
Yuck! Good catch!
> I've tested the following on an old CentOS 5.2 chroot with SVN 1.4.2:
Looks good to me.
--
Alligator sandwich, and make it snappy!
--
To u
Here are the topics that have been cooking. Commits prefixed with '-' are
only in 'pu' (proposed updates) while commits prefixed with '+' are in 'next'.
The series to mark strings for i18n in merge(-recursive), am and
rebase have been merged. From now on, until the final release, no
further upda
On 02.08.2012 23:00, Junio C Hamano wrote:
Adam Butcher writes:
+# create a file containing numbers with no newline at
+# the end and modify it such that the starting 10 lines
+# are unchanged, the next 101 are rewritten and the last
+# line differs only in that in is terminated by a newline.
On Thu, Aug 02, 2012 at 03:51:17PM -0700, Junio C Hamano wrote:
> > On Wed, Aug 01, 2012 at 03:10:55PM -0700, Junio C Hamano wrote:
> > ...
> >> When you move porn/0001.jpg in the preimage to naughty/1.jpg in
> >> the postimage, they both can hit "*.jpg contentid=jpeg" line in the
> >> top-lev
Jeff King writes:
> On Wed, Aug 01, 2012 at 03:10:55PM -0700, Junio C Hamano wrote:
> ...
>> When you move porn/0001.jpg in the preimage to naughty/1.jpg in
>> the postimage, they both can hit "*.jpg contentid=jpeg" line in the
>> top-level .gitattribute file, and the contentid driver for jpe
On Thu, Aug 02, 2012 at 07:08:25PM +0700, Nguyen Thai Ngoc Duy wrote:
> > I implemented (1a). Implementing (1b) would be easy, but for a full-on
> > cache (especially for "-C"), I think the resulting size might be
> > prohibitive.
>
> (1a) is good regardless rename overrides. Why don't you polish
On Wed, Aug 01, 2012 at 03:10:55PM -0700, Junio C Hamano wrote:
> Jeff King writes:
>
> > On Tue, Jul 31, 2012 at 11:01:27PM -0700, Junio C Hamano wrote:
> > ...
> >> As we still have the pathname in this codepath, I am wondering if we
> >> would benefit from custom "content hash" that knows the
It hasn't been used since 2006, as of commit 3cd4f5e8
"git-apply --binary: clean up and prepare for --reverse"
Signed-off-by: Paul Gortmaker
diff --git a/builtin/apply.c b/builtin/apply.c
index d453c83..3bf71dc 100644
--- a/builtin/apply.c
+++ b/builtin/apply.c
@@ -188,7 +188,6 @@ struct pa
On 12-08-02 05:20 PM, Junio C Hamano wrote:
> Paul Gortmaker writes:
>
>> The '-D' or '--irreversible-delete' option of format-patch is
>> great for sending out patches to mailing lists, where there
>> is little value in seeing thousands of lines of deleted code.
>> Attention can then be focused
On 02.08.2012 22:33, Jeff King wrote:
On Thu, Aug 02, 2012 at 10:11:02PM +0100, Adam Butcher wrote:
From 01730a741cc5fd7d0a5d8bd0d3df80d12c81fe48 Mon Sep 17 00:00:00
2001
From: Adam Butcher
Date: Wed, 1 Aug 2012 22:25:09 +0100
Subject: [PATCH] Fix 'No newline...' annotation in rewrite diffs.
On Thu, Aug 02, 2012 at 02:52:56PM -0700, Junio C Hamano wrote:
> > Seq is (unfortunately) not portable. I usually use a perl snippet
> > instead, like:
> >
> > perl -le 'print for (1..10)'
> >
> > Though I think we are adjusting that to use $PERL_PATH these days.
>
> t/perf/perf-lib.sh and t/t
"Robin H. Johnson" wrote:
> On Thu, Aug 02, 2012 at 11:58:08AM -0700, Junio C Hamano wrote:
> > > Thanks from me as well. I'm still worried about whether the increased
> > > use of canonicalize_url will introduce regressions for the existing
> > > SVN 1.6 support, and I should have time to look
Eric Wong writes:
> Eric Wong wrote:
>> Junio C Hamano wrote:
>> > Eric Wong writes:
>> > > Thanks for reminding me, I went back to an old chroot 1.4.2 indeed
>> > > does fail canonicalization.
>> > >
>> > > Will bisect and squash a fix in.
>> >
>> > Oops; should I eject this out of next and
Adam Butcher writes:
> +# create a file containing numbers with no newline at
> +# the end and modify it such that the starting 10 lines
> +# are unchanged, the next 101 are rewritten and the last
> +# line differs only in that in is terminated by a newline.
> +seq 1 10 > seq
> +seq 100 +1 200 >>
Eric Wong wrote:
> Junio C Hamano wrote:
> > Eric Wong writes:
> > > Thanks for reminding me, I went back to an old chroot 1.4.2 indeed
> > > does fail canonicalization.
> > >
> > > Will bisect and squash a fix in.
> >
> > Oops; should I eject this out of next and wait for a reroll, then?
>
>
Jeff King writes:
> On Thu, Aug 02, 2012 at 10:11:02PM +0100, Adam Butcher wrote:
>
>> From 01730a741cc5fd7d0a5d8bd0d3df80d12c81fe48 Mon Sep 17 00:00:00 2001
>> From: Adam Butcher
>> Date: Wed, 1 Aug 2012 22:25:09 +0100
>> Subject: [PATCH] Fix 'No newline...' annotation in rewrite diffs.
>
> You
Eric Wong wrote:
> Michael G Schwern wrote:
> > On 2012.7.28 6:55 AM, Jonathan Nieder wrote:
> > > Michael G. Schwern wrote:
> > >> --- a/perl/Git/SVN/Utils.pm
> > >> +++ b/perl/Git/SVN/Utils.pm
> > >> @@ -86,6 +86,27 @@ sub _collapse_dotdot {
> > >>
> > >>
> > >> sub canonicalize_path {
> >
Junio C Hamano wrote:
> Eric Wong writes:
> > Thanks for reminding me, I went back to an old chroot 1.4.2 indeed
> > does fail canonicalization.
> >
> > Will bisect and squash a fix in.
>
> Oops; should I eject this out of next and wait for a reroll, then?
Your call, I doubt anybody on next use
On Thu, Aug 02, 2012 at 10:11:02PM +0100, Adam Butcher wrote:
> From 01730a741cc5fd7d0a5d8bd0d3df80d12c81fe48 Mon Sep 17 00:00:00 2001
> From: Adam Butcher
> Date: Wed, 1 Aug 2012 22:25:09 +0100
> Subject: [PATCH] Fix 'No newline...' annotation in rewrite diffs.
You can drop these lines from the
Eric Wong writes:
> Jonathan Nieder wrote:
>> Thanks from me as well. I'm still worried about whether the increased
>> use of canonicalize_url will introduce regressions for the existing
>> SVN 1.6 support, and I should have time to look it over this weekend.
>>
>> The comment in canonicalize_
Paul Gortmaker writes:
> The '-D' or '--irreversible-delete' option of format-patch is
> great for sending out patches to mailing lists, where there
> is little value in seeing thousands of lines of deleted code.
> Attention can then be focused on the changes relating to
> the binding of the dele
On 08/02/2012 01:02 PM, Andreas Grünbacher wrote:
> Paul and Jim, any thoughts about adding filename quoting to GNU diff anytime
> soon, or about the format?
The git format looks OK. I don't have time right now to add it to GNU diff,
tho.
--
To unsubscribe from this list: send the line "unsubscr
From 01730a741cc5fd7d0a5d8bd0d3df80d12c81fe48 Mon Sep 17 00:00:00 2001
From: Adam Butcher
Date: Wed, 1 Aug 2012 22:25:09 +0100
Subject: [PATCH] Fix 'No newline...' annotation in rewrite diffs.
When operating in --break-rewrites (-B) mode on a file with no newline
terminator (and assuming --break
Paul Eggert wrote:
> On 08/02/2012 01:02 PM, Andreas Grünbacher wrote:
>> Paul and Jim, any thoughts about adding filename quoting to GNU diff anytime
>> soon, or about the format?
>
> The git format looks OK. I don't have time right now to add it to GNU diff,
> tho.
I agree, and have no more fr
Jonathan Nieder wrote:
> Thanks from me as well. I'm still worried about whether the increased
> use of canonicalize_url will introduce regressions for the existing
> SVN 1.6 support, and I should have time to look it over this weekend.
>
> The comment in canonicalize_url "There wasn't a 1.6 way
The '-D' or '--irreversible-delete' option of format-patch is
great for sending out patches to mailing lists, where there
is little value in seeing thousands of lines of deleted code.
Attention can then be focused on the changes relating to
the binding of the deleted code (Makefiles, etc).
However
Junio,
2012/8/1 Junio C Hamano :
> Andreas Grünbacher writes:
>
>> * Support for double-quoted filenames in the "diff --git" format: when a
>> filename starts with a double quote, it is interpreted as a C string
>> literal. The escape sequences \\, \", \a, \b, \f, \n, \r, \t, \v, and \ooo
>>
On Thu, Aug 02, 2012 at 11:58:08AM -0700, Junio C Hamano wrote:
> > Thanks from me as well. I'm still worried about whether the increased
> > use of canonicalize_url will introduce regressions for the existing
> > SVN 1.6 support, and I should have time to look it over this weekend.
>
> Likewise
Jonathan Nieder writes:
> Hi,
>
> Eric Wong wrote:
>> "Michael G. Schwern" wrote:
>
>>> This patch series fixes git-svn for SVN 1.7 tested against SVN 1.7.5 and
>>> 1.6.18. Patch 7/8 is where SVN 1.7 starts passing.
>>
>> Thanks Michael. I've made minor editorial changes (mostly rewording
>> c
On Wed, Aug 1, 2012 at 7:09 PM, Ralf Thielow wrote:
> $(eval_gettext 'It seems that there is already a $state_dir_base directory,
> and
> -I wonder if you ware in the middle of another rebase. If that is the
> +I wonder if you are in the middle of another rebase. If that is the
> case, please
Translate 76 new messages came from git.pot update
in 3b6137f (l10n: Update git.pot (76 new, 4 removed
messages)).
Signed-off-by: Ralf Thielow
---
Hi German l10n team,
please review this update of German translation.
Please note that this patch is based on a preparation
commit which just contai
Hi,
Eric Wong wrote:
> "Michael G. Schwern" wrote:
>> This patch series fixes git-svn for SVN 1.7 tested against SVN 1.7.5 and
>> 1.6.18. Patch 7/8 is where SVN 1.7 starts passing.
>
> Thanks Michael. I've made minor editorial changes (mostly rewording
> commit titles to fit the larger project
Alright, I have revised the patch and fixed up the nits that were
picked and made a quick modification. I've added a setting for
grep.patternType for "default" which can restore the default grep
pattern matching behaviour and restores the functionality back to
grep.extendedRegexp. I added this func
On 08/02, Nguyen Thai Ngoc Duy wrote:
> On Thu, Aug 2, 2012 at 6:01 PM, Thomas Gummerer wrote:
> > @@ -1443,7 +1452,6 @@ void read_index_v2(struct index_state *istate, void
> > *mmap, int mmap_size)
> > src_offset += consumed;
> > }
> > strbuf_release(&previous_nam
On 08/02, Nguyen Thai Ngoc Duy wrote:
> On Thu, Aug 2, 2012 at 6:01 PM, Thomas Gummerer wrote:
> > @@ -1785,7 +1785,7 @@ void update_index_if_able(struct index_state *istate,
> > struct lock_file *lockfile
> > rollback_lock_file(lockfile);
> > }
> >
> > -int write_index(struct i
On 08/02, Nguyen Thai Ngoc Duy wrote:
> General note. I wonder if we should create a separate source file for
> v5 (at least the low level handling part). Partial reading/writing
> will come (hopefully soon) and read-cache.c on master is already close
> to 2000 lines.
To me it would make sense, bu
On 08/02, Nguyen Thai Ngoc Duy wrote:
> On Thu, Aug 2, 2012 at 6:02 PM, Thomas Gummerer wrote:
> > Add a performance test for index version [23]/4/5 by using
> > git update-index --force-rewrite, thus testing both the reader
> > and the writer speed of all index formats.
>
> On the testing side,
On Thu, Aug 2, 2012 at 8:47 PM, Thomas Gummerer wrote:
> On 08/02, Nguyen Thai Ngoc Duy wrote:
>> On Thu, Aug 2, 2012 at 6:01 PM, Thomas Gummerer wrote:
>> > Documentation/technical/index-file-format-v5.txt | 281
>> > ++
>> > builtin/update-index.c
On 08/02, Nguyen Thai Ngoc Duy wrote:
> On Thu, Aug 2, 2012 at 6:01 PM, Thomas Gummerer wrote:
> > Documentation/technical/index-file-format-v5.txt | 281
> > ++
> > builtin/update-index.c |5 +-
> > cache-tree.c
On Thu, Aug 2, 2012 at 6:02 PM, Thomas Gummerer wrote:
> Add a performance test for index version [23]/4/5 by using
> git update-index --force-rewrite, thus testing both the reader
> and the writer speed of all index formats.
On the testing side, it may be an interesting idea to force the whole
t
General note. I wonder if we should create a separate source file for
v5 (at least the low level handling part). Partial reading/writing
will come (hopefully soon) and read-cache.c on master is already close
to 2000 lines.
On Thu, Aug 2, 2012 at 6:01 PM, Thomas Gummerer wrote:
> +static struct ca
On Thu, Aug 2, 2012 at 6:01 PM, Thomas Gummerer wrote:
> @@ -1785,7 +1785,7 @@ void update_index_if_able(struct index_state *istate,
> struct lock_file *lockfile
> rollback_lock_file(lockfile);
> }
>
> -int write_index(struct index_state *istate, int newfd)
> +int write_index_v2(
On Thu, Aug 2, 2012 at 6:01 PM, Thomas Gummerer wrote:
> @@ -1443,7 +1452,6 @@ void read_index_v2(struct index_state *istate, void
> *mmap, int mmap_size)
> src_offset += consumed;
> }
> strbuf_release(&previous_name_buf);
> -
> while (src_offset <= mmap_si
On Thu, Aug 2, 2012 at 6:01 PM, Thomas Gummerer wrote:
> diff --git a/cache.h b/cache.h
> index 6e9a243..d4028ef 100644
> --- a/cache.h
> +++ b/cache.h
> @@ -99,9 +99,12 @@ unsigned long git_deflate_bound(git_zstream *, unsigned
> long);
> */
>
> #define CACHE_SIGNATURE 0x44495243 /* "DIRC
On Thu, Aug 2, 2012 at 6:01 PM, Thomas Gummerer wrote:
> Documentation/technical/index-file-format-v5.txt | 281
> ++
> builtin/update-index.c |5 +-
> cache-tree.c | 145 ++
> cache-t
On Thu, Aug 2, 2012 at 4:27 AM, Jeff King wrote:
> Yes, if you go with a commit-based approach, you can do either notes or
> in-commit messages. In other words, I would break the solutions down as:
>
> 1. Store sha1+sha1 -> score mapping (i.e., what I suggested). This is
> fundamentally a g
Make git read the index file version 5 without complaining.
This version of the reader doesn't read neither the cache-tree
nor the resolve undo data, but doesn't choke on an index that
includes such data.
Signed-off-by: Thomas Gummerer
---
cache.h | 79 +
read-cache.c | 570 ++
Make git read the resolve-undo data from the index.
Since the resolve-undo data is joined with the conflicts in
the ondisk format of the index file version 5, conflicts and
resolved data is read at the same time, and the resolve-undo
data is then converted to the in-memory format.
Signed-off-by:
Add a force-rewrite option to update-index, which allows the user
to rewrite the index, even if there are no changes. This can be used
to do performance tests of both the reader and the writer.
Signed-off-by: Thomas Gummerer
---
builtin/update-index.c | 5 -
1 file changed, 4 insertions(+),
Add a performance test for index version [23]/4/5 by using
git update-index --force-rewrite, thus testing both the reader
and the writer speed of all index formats.
Signed-off-by: Thomas Gummerer
---
t/perf/p0002-index.sh | 33 +
1 file changed, 33 insertions(+)
Write the cache-tree data for the index version 5 file format. The
in-memory cache-tree data is converted to the ondisk format, by adding
it to the directory entries, that were compiled from the cache-entries
in the step before.
Signed-off-by: Thomas Gummerer
---
cache-tree.c | 52 ++
Write the resolve undo data to the ondisk format, by joining the data
in the resolve-undo string-list with the already existing conflicts
that were compiled before, when searching the directories and add
them to the corresponding directory entries.
Signed-off-by: Thomas Gummerer
---
read-cache.c
Write the index version 5 file format to disk. This version doesn't
write the cache-tree data and resolve-undo data to the file.
The main work is done when filtering out the directories from the
current in-memory format, where in the same turn also the conflicts
and the file data is calculated.
S
Since the cache-tree data is saved as part of the directory data,
we have already read it, when we want to read the cache-tree. The
cache-tree then only has to be converted from the directory data.
The cache-tree isn't lexically sorted, but after the pathlen at
each level, therefore the directorie
The new git racy code uses the mtime of cache-entries to smudge
a racy clean entry, and loads the work, of checking the file-system
if the entry has really changed, off to the reader. This interferes
with this test, because the entry is racily smudged and thus has
mtime 0. We wait 1 second to avoid
Make the in-memory format aware of the stat_crc used by index-v5.
It is simply ignored by index version prior to v5.
Signed-off-by: Thomas Gummerer
---
cache.h | 1 +
read-cache.c | 27 +++
2 files changed, 28 insertions(+)
diff --git a/cache.h b/cache.h
index 3aa7
Add a documentation of the index file format version 5 to
Documentation/technical.
Signed-off-by: Thomas Gummerer
---
Documentation/technical/index-file-format-v5.txt | 281 +++
1 file changed, 281 insertions(+)
create mode 100644 Documentation/technical/index-file-format-v5
Modify match_stat_basic, into one function that handles the
general case, which is the same for all index formats, and
a function that handles the specific parts for each index
file version.
Signed-off-by: Thomas Gummerer
---
read-cache.c | 77 +++-
Modify the read_index_from function, splitting it up into
one function that stays the same for every index format,
doing the basic operations such as verifying the header,
and a function which is specific for each index version,
which does the real reading of the index.
Signed-off-by: Thomas Gumme
Modify the cache_header such that other index file formats
can be added and reusing the common part of each index format.
The signature and version have to be present in every
version of the index file format, to check if it can be read
by a specific version of git, while other entries (eg. number
The test t2104 currently checks if the index version is correctly
reduced to 2/increased to 3, when an entry need extended flags,
or doesn't use them anymore. Since index-v5 doesn't have extended
flags (the extended flags are part of the normal flags), we simply
add a check if the index version is
Modify the write_index function to add the possibility to add
other index formats, that are written in a different way. Also
mark all functions, which shall only be used with v2-v4 as v2
functions.
Signed-off-by: Thomas Gummerer
---
read-cache.c | 40
1 f
Series of patches to introduce the index version 5 file format. This
series does not include any fancy stuff like partial loading or partial
writing yet, though it's possible to do that with the new format.
There was already a POC for partial loading, which gave pretty good
results, which was howe
Robert Luberda wrote:
> dcommit didn't handle errors returned by SVN and coped very
> poorly with concurrent commits that appear in SVN repository
> while dcommit was running. In both cases it left git repository
> in inconsistent state: index (which was reset with `git reset
> --mixed' after a su
Scripts such as git rebase -i cannot currently cherry-pick commits which
have an empty commit message, as git cherry-pick calls git commit
without the --allow-empty-message option.
Add an --allow-empty-message option to git cherry-pick which is passed
through to git commit, so this behaviour can b
"Michael G. Schwern" wrote:
> This patch series fixes git-svn for SVN 1.7 tested against SVN 1.7.5 and
> 1.6.18. Patch 7/8 is where SVN 1.7 starts passing.
Thanks Michael. I've made minor editorial changes (mostly rewording
commit titles to fit the larger project).
Junio:
The following change
On 1 August 2012 23:26, Junio C Hamano wrote:
> I've read your entire response three times, and I am having a hard
> time deciding if you are against my suggestion, or you misread my
> suggestion.
My apologies, I can see how my message wasn't as clear as it could have been.
> I guess by "perpetu
Junio C Hamano writes:
> My recommendation, backed by the above line of thought, is to add
> support for the "--allow-empty-message" option to both "rebase [-i]"
> and "cherry-pick", defaulting to false.
Thanks for the very detailed analysis and advice Junio. I like your
suggested --allow-empty-
This fix is used to return the svn reference of the remote svn upstream
branch when the git repository is a clone of a svn repository that was
created with the --stdlayout and --prefix options of git svn command.
* completion/git-prompt.sh: add function to resolve svn branches
into git remote r
Hi,
It seems that git gui doesn't like multiple remotes without a "fetch =
" line. Having just one remote with no "fetch =" line is OK. Having
more than one bombs, though adding "fetch =" to just one of them seems
to work. It's easy to reproduce:
lasse@plystrofyf:/tmp$ mkdir test.git
lasse@plystr
72 matches
Mail list logo
