CC += dev@
Victor Sudakov wrote on Mon, Jul 26, 2010 at 12:53:44 +0700:
Daniel Shahaf wrote:
I have the following line in /usr/local/lib/sasl2/svn.conf:
mech_list: gssapi digest-md5 anonymous
How can I guarantee that the subversion client/server will always use
GSSAPI before
Hi Stefan!
A few questions and a WIP patch for making the patch code deal with
property targets being dirs.
* How should we deal with SVN_ERR_ILLEGAL_TARGET? svn_wc_prop_set()
throws one of those if we for instance try to set svn:executable on a
dir or svn:ignore on a file. Should we do some
A merge into a directory with a file external shows the following
problems:
1) mergeinfo at the directory is not inherited (and all files and
subdirectories get their own inheritable mergeinfo).
2) mergeinfo is added to the external file.
The mergeinfo at the external file is just wrong and
Hi Ann,
Patch has not reached the list.
Did you forget attach it?
With regards
Kamesh Jayachandran
On 07/26/2010 08:34 AM, svnusert...@href.com wrote:
[[[
Contributed Feature: enable lazy sync via new export flag,
--skipfilesmatchingsize
* subversion/libsvn_client/export.c
*
Hi!
Posting this one here so I don't forget about it.
The patch code should not change the wc when run with the --dry-run
option but it will if we're adding a file with a parent scheduled for
deletion.
The patch code will add a file with a parent scheduled for deletion to
the filesystem but not
On Mon, Jul 26, 2010 at 09:59:01AM +0200, Daniel Näslund wrote:
[[[
Enable the patch code to use dirs as targets if we only have property
changes.
* subversion/libsvn_client/patch.c
(patch_target_t): Add fields 'has_text_changes' and 'has_prop_changes'
to be able to decide if we
On Mon, Jul 26, 2010 at 01:04:26PM +0200, Daniel Näslund wrote:
Hi!
Posting this one here so I don't forget about it.
The patch code should not change the wc when run with the --dry-run
option but it will if we're adding a file with a parent scheduled for
deletion.
The patch code will
This breaks the 2 tests in svnrdump_tests.
Though the following inline patch may fix it you may have different plans
Index: subversion/tests/cmdline/svnrdump_tests.py
===
--- subversion/tests/cmdline/svnrdump_tests.py (revision
On Mon, Jul 26, 2010 at 01:07:14PM +0200, Stefan Sperling wrote:
On Mon, Jul 26, 2010 at 09:59:01AM +0200, Daniel Näslund wrote:
Index: subversion/libsvn_client/patch.c
===
--- subversion/libsvn_client/patch.c(revision
On Mon, Jul 26, 2010 at 03:04:02AM +, svnusert...@href.com wrote:
[[[
Contributed Feature: enable lazy sync via new export flag,
--skipfilesmatchingsize
Discussed in part here: http://svn.haxx.se/users/archive-2010-07/0282.shtml
Hi Ann,
As I've said before in the discussion you've
Hi,
Ramkumar Ramachandra writes:
The tests in svnsync_tests_data/ are in dumpfile v2 and these are
unsuitable for testing svnrdump. Hence, load all of them into a
repository and re-dump them in dumpfile v3 format before attempting to
add them to svnrdump_tests_data/. I still have to figure
Hi,
Ramkumar Ramachandra writes:
Hi Daniel,
Thanks for the review. Here's another revision.
Could someone review this patch? My subcommand patch to svnrdump
breaks all the tests and I need to fix them.
Thanks!
-- Ram
Hi,
Ramkumar Ramachandra writes:
Hi,
Instead of spawning diff, I used difflib in Python. Here's the
result. I need a +1 to commit this since it's outside my area.
Is this patch alright?
Thanks.
-- Ram
On Mon, Jul 26, 2010 at 05:21:28PM +0530, Ramkumar Ramachandra wrote:
Hi,
Ramkumar Ramachandra writes:
Hi,
Instead of spawning diff, I used difflib in Python. Here's the
result. I need a +1 to commit this since it's outside my area.
Is this patch alright?
Do all tests pass with
Hi Kamesh,
Kamesh Jayachandran writes:
This breaks the 2 tests in svnrdump_tests.
Though the following inline patch may fix it you may have different plans
Thanks! Yes, I'm aware that it breaks tests. Unfortunately, I have too
many uncommitted changes, and I'm still waiting for them to be
On Mon, Jul 26, 2010 at 05:17:07PM +0530, Ramkumar Ramachandra wrote:
Hi,
Ramkumar Ramachandra writes:
The tests in svnsync_tests_data/ are in dumpfile v2 and these are
unsuitable for testing svnrdump. Hence, load all of them into a
repository and re-dump them in dumpfile v3 format
Hi Stefan,
Stefan Sperling writes:
Do all tests pass with it?
Yes, they do. I ran `make check` and checked. I'm running it once
more, just in case.
Thanks.
-- Ram
On 07/26/2010 05:25 PM, Ramkumar Ramachandra wrote:
Hi Kamesh,
Kamesh Jayachandran writes:
This breaks the 2 tests in svnrdump_tests.
Though the following inline patch may fix it you may have different plans
Thanks! Yes, I'm aware that it breaks tests. Unfortunately, I have too
On Mon, Jul 26, 2010 at 05:25:10PM +0530, Ramkumar Ramachandra wrote:
Hi Kamesh,
Kamesh Jayachandran writes:
This breaks the 2 tests in svnrdump_tests.
Though the following inline patch may fix it you may have different plans
Thanks! Yes, I'm aware that it breaks tests.
Hi,
I've started working on load support too, and `svnrdump load` can be
quite misleading- should we rename the tool? I'm floating this thread
early so that we can minimize code churn due to the rename (in tests
for example).
Thanks.
-- Ram
On Mon, Jul 26, 2010 at 05:33:20PM +0530, Ramkumar Ramachandra wrote:
Hi Stefan,
Stefan Sperling writes:
Do all tests pass with it?
Yes, they do. I ran `make check` and checked. I'm running it once
more, just in case.
That's not neccessary. Your patch and the new diff output look quite
-Original Message-
From: Ramkumar Ramachandra [mailto:artag...@gmail.com]
Sent: maandag 26 juli 2010 14:13
To: Subversion-dev Mailing List
Subject: Proposal: Rename svnrdump
Hi,
I've started working on load support too, and `svnrdump load` can be
quite misleading- should we
Hi Stefan,
Stefan Sperling writes:
What about doing one thing at a time instead?
I'm not trying to be sarcastic or offensive, but you seem to be trying
to juggle too many things at once. Writing good code takes time. Maybe
if you focused more thoroughly on single tasks with a bit more
On Mon, Jul 26, 2010 at 03:16:35PM +0200, Bert Huijben wrote:
But I'm fine with just keeping 'svnrdump' even for loading. It just uses
'dump' for referring that it processes dumpfiles (like svndumpfilter does),
instead of to just dumping the repository.
+1
Stefan
[[[
* subversion/tests/cmdline/svntest/actions.py
(run_and_verify_svnrdump): Make it more extensible by adding extra
arguments to check expected vales. Design it like
run_and_verify_svnadmin2.
* subversion/tests/cmdline/svnrdump_tests.py
(run_test, basic_svnrdump): Track the above change.
On Mon, Jul 26, 2010 at 06:53:26PM +0530, Ramkumar Ramachandra wrote:
Hi Stefan,
Stefan Sperling writes:
What about doing one thing at a time instead?
I'm not trying to be sarcastic or offensive, but you seem to be trying
to juggle too many things at once. Writing good code takes
[[[
* subversion/tests/cmdline/svntest/verify.py
(display_lines): Additionally print unified diff output using Python
difflib.unified_diff.
]]]
Index: subversion/tests/cmdline/svntest/verify.py
===
---
Hi Stefan,
Stefan Sperling writes:
Oh, I know that hacking time is scarce. Mine is, too.
And I'm happy that you are very active. But please forgive me when
I can't commit on every patch you send, or keep comments short.
Don't worry about it at all -- I totally understand. Everyone has
limited
On Thu, 2010-07-22, Ramkumar Ramachandra wrote:
Jonathan Nieder writes:
[...]
+ /* Output name length, then name. */
+ svn_stringbuf_appendcstr(*strbuf,
+ apr_psprintf(pool, K %
APR_SSIZE_T_FMT \n,
+
On Mon, Jul 26, 2010 at 07:07:12PM +0530, Ramkumar Ramachandra wrote:
[[[
* subversion/tests/cmdline/svntest/verify.py
(display_lines): Additionally print unified diff output using Python
difflib.unified_diff.
]]]
+1
Index: subversion/tests/cmdline/svntest/verify.py
Hi Stefan,
I do hear you. I also question whether RSync, or something, is truly
a good option for Windows users for pulling files from a
repository. Do you know Windows users that use RSync happily?I
fully appreciate that Linux users are happy with RSync. I tried to
find the 2nd
On 07/26/2010 07:25 PM, artag...@apache.org wrote:
Author: artagnon
Date: Mon Jul 26 13:55:25 2010
New Revision: 979282
URL: http://svn.apache.org/viewvc?rev=979282view=rev
Log:
svnrdump: Add an a no-op load editor
* subversion/svnrdump/svnrdump.c
(load_revisions): New function to fetch and
Ann,
I am a Windows user, and I agree that such a feature shouldn't go into
the codebase. Many such features are traditionally solved using
scripting, and while scripting on unix used to be much more powerful
than on windows, this is no longer true for many years now. Instead of
creating a
On Mon, Jul 26, 2010 at 01:40:01PM +, svnusert...@href.com wrote:
Hi Stefan,
I do hear you. I also question whether RSync, or something, is
truly a good option for Windows users for pulling files from a
repository. Do you know Windows users that use RSync happily?I
fully
Kamesh Jayachandran kam...@collab.net writes:
On 07/26/2010 07:25 PM, artag...@apache.org wrote:
Author: artagnon
Date: Mon Jul 26 13:55:25 2010
New Revision: 979282
URL: http://svn.apache.org/viewvc?rev=979282view=rev
Log:
svnrdump: Add an a no-op load editor
*
On Mon, Jul 26, 2010 at 6:16 AM, Bert Huijben b...@qqmail.nl wrote:
(A much older suggestion was moving the code into svnsync as they are
related tools)
I think either that or folding it into svnadmin itself. But, we don't
need another tool program, IMO. -- justin
On Mon, Jul 26, 2010 at 10:08 AM, C. Michael Pilato cmpil...@collab.net wrote:
On 07/26/2010 10:54 AM, Justin Erenkrantz wrote:
On Mon, Jul 26, 2010 at 6:16 AM, Bert Huijben b...@qqmail.nl wrote:
(A much older suggestion was moving the code into svnsync as they are
related tools)
I think
On Mon, Jul 26, 2010 at 07:54:30AM -0700, Justin Erenkrantz wrote:
On Mon, Jul 26, 2010 at 6:16 AM, Bert Huijben b...@qqmail.nl wrote:
(A much older suggestion was moving the code into svnsync as they are
related tools)
I think either that or folding it into svnadmin itself. But, we don't
On Mon, Jul 26, 2010 at 8:10 AM, Hyrum K. Wright
hyrum_wri...@mail.utexas.edu wrote:
+1. The charter for svnadmin is currently, and should remain, solely
focused on local access to a repository.
Why?
svnadmin dump --remote url
seems the most intuitive approach for this - as well as load,
-Original Message-
From: Stefan Sperling [mailto:s...@elego.de]
Sent: maandag 26 juli 2010 17:15
To: Justin Erenkrantz
Cc: Bert Huijben; Ramkumar Ramachandra; Subversion-dev Mailing List
Subject: Re: Proposal: Rename svnrdump
On Mon, Jul 26, 2010 at 07:54:30AM -0700, Justin
On Mon, Jul 26, 2010 at 8:30 AM, Justin Erenkrantz
jus...@erenkrantz.com wrote:
On Mon, Jul 26, 2010 at 8:10 AM, Hyrum K. Wright
hyrum_wri...@mail.utexas.edu wrote:
+1. The charter for svnadmin is currently, and should remain, solely
focused on local access to a repository.
Why?
svnadmin
For people that are interested, I've recently changed the location of
the trunk nightlies. They are now hosted here:
http://people.apache.org/~hwright/svn/nightly/
The old site on orac should now redirect to people.apache.org.
-Hyrum
-Original Message-
From: Stefan Sperling [mailto:s...@elego.de]
Sent: Monday, July 26, 2010 3:34 PM
To: svnusert...@href.com
Cc: dev@subversion.apache.org
Subject: Re: [PATCH] new feature, lazy sync via export --
skipfilesmatchingsize
[snip]
I don't know if there is interest
Hi Bert,
rhuij...@apache.org writes:
Author: rhuijben
Date: Mon Jul 26 14:24:43 2010
New Revision: 979303
URL: http://svn.apache.org/viewvc?rev=979303view=rev
Log:
Add a simple property verifyer to the upgrade tests to test if the upgrade
code correctly handles property upgrades. This
Hi Stefan,
Stefan Sperling writes:
I am a bit worried that a loader is being worked on before the dumper is
completely done. I think we should finish the dumper first.
Don't worry about it- I won't leave the dumper pending. It's just that
there are some changes that I'm waiting for before
Hi,
David Glasser writes:
But we do. 'svn' is a wrapper around svn_client/svn_ra. 'svnadmin'
is a wrapper around svn_repos/svn_fs. 'svn' always refers to
repositories via URLs. 'svnadmin' always (I think) refers to
repositories via paths. 'svnadmin' does not have a dependency on Neon
or
On Mon, Jul 26, 2010 at 10:09 AM, Ramkumar Ramachandra
artag...@gmail.com wrote:
+1
I don't see how it's possible to merge, atleast at this
stage. `svnadmin` is inherently dependent on the filesytem- we should
be careful not to stuff too much unrelated functionality into one
tool.
I
We used to be brutal against adding flags let alone separate
binaries,
so why are we so willing to add new binaries willy-nilly? To me,
as
an admin, it just makes *so* much sense that a remote dump feature
be
tied into svnadmin dump. -- justin
I agree this makes sense.
You could also
On Mon, Jul 26, 2010 at 12:30 PM, Bob Archer bob.arc...@amsi.com wrote:
We used to be brutal against adding flags let alone separate
binaries,
so why are we so willing to add new binaries willy-nilly? To me,
as
an admin, it just makes *so* much sense that a remote dump feature
be
tied into
Hi Julian,
Julian Foad writes:
On Thu, 2010-07-22, Ramkumar Ramachandra wrote:
Jonathan Nieder writes:
[...]
+ /* Output name length, then name. */
+ svn_stringbuf_appendcstr(*strbuf,
+
Hi,
artag...@apache.org writes:
-def revision0(sbox):
+def revision_0(sbox):
dump revision zero
- run_test(sbox, dumpfile_name = revision0.dump)
+ run_test(sbox, dumpfile_name = revision-0.dump)
+def copy_and_modify(sbox):
+ copy and modify
+ run_test(sbox, copy-and-modify.dump)
Hi,
just wanted to let you guys know that in the coming weeks I won't be able to
get the mac buildbot back online.
It'll probably be mid-september or so, when my exams and other work is
finished.
Lieven
Hi,
It's a little late here, and I don't have the patience the debug this
segmentation fault right now. Here's the backtrace from a simple `svn
diff`: if it helps, revision-0.dump had just been `svn mv`ed from
revision0.dump. I think it has something to do with the file's
svn:mime-type being
Hi Stefan,
I imported and ran the copy_and_modify test from svnsync. The diff I
get clearly indicates what more needs to be done, and I think this
test is very valuable. Unfortunately, `svn diff` refuses to show me
the dumpfile itself (application/octet-stream), so here's the rest of
the diff
Stefan Sperling wrote on Mon, Jul 26, 2010 at 16:05:23 +0200:
On Mon, Jul 26, 2010 at 07:07:12PM +0530, Ramkumar Ramachandra wrote:
[[[
* subversion/tests/cmdline/svntest/verify.py
(display_lines): Additionally print unified diff output using Python
difflib.unified_diff.
]]]
+1
[ VERY briefly, since it's middle of the night here ]
Stefan Sperling wrote on Mon, Jul 26, 2010 at 16:33:45 +0200:
In this case though, I'd like to explicitly ask any readers who have silently
been following the thread to voice their opinion, especially if they
differ with my view. Maybe
On Mon, Jul 26, 2010 at 05:30:56PM +0100, Bolstridge, Andrew wrote:
If you have a directory full of source files,
and it contains a sub-directory full of almost-never-changing, large
binary files, then it would be great to initiate an export and have it
recognise the binaries haven't changed
Ramkumar Ramachandra wrote on Mon, Jul 26, 2010 at 22:17:44 +0530:
rhuij...@apache.org writes:
+ def diff_props(dict1, dict2, name, match):
I wrote some 'dict diffing' code a long time ago in run_and_verify_info().
Maybe it's relevant to this thread...
Hi Bert,
rhuij...@apache.org writes:
Author: rhuijben
Date: Mon Jul 26 19:35:44 2010
New Revision: 979416
URL: http://svn.apache.org/viewvc?rev=979416view=rev
Log:
Remove a left over reference to revert properties. New code should talk
about base_properties or pristine_properties. This
Hi Daniel,
Daniel Shahaf writes:
Ramkumar Ramachandra wrote on Tue, Jul 27, 2010 at 00:02:07 +0530:
Hi,
It's a little late here, and I don't have the patience the debug this
segmentation fault right now. Here's the backtrace from a simple `svn
diff`: if it helps, revision-0.dump had
What does WC-NG stand for? Working Copy ... with only 1 .svn folder
instead of nested .svn folders at every level, is that correct?
I'll tell you the one way in which that might not completely solve
the use-case that I have in mind, and that is doing a series of
partial exports from multiple
61 matches
Mail list logo
