it might help to pass TRUE for
only_check_type on svn_io_get_dirents3().
(On Windows that doesn’t change the performance, but I think it does on other
platforms)
Bert
From: Branko Čibej [mailto:br...@wandisco.com]
Sent: dinsdag 1 april 2014 13:10
To: dev@subversion.apache.org
-Original Message-
From: Philip Martin [mailto:philip.mar...@wandisco.com]
Sent: vrijdag 28 maart 2014 20:57
To: Bert Huijben
Cc: dev@subversion.apache.org; comm...@subversion.apache.org
Subject: Re: svn commit: r1582845 -
/subversion/trunk/subversion/libsvn_fs/fs-loader.c
Bert
for svn_fs_lock2?
And it is certainly not our standard pattern for public functions... By default
the output arguments are undefined on error.
(It looks like all the implementations handle things this way...)
Bert
+{
+ svn_fs_lock_result_t *result
. svn:date,
svn:author))},
+ N_(skip revision property ARG (e.g.,
\svn:author\))},
{config-dir,opt_config_dir, 1,
N_(read user configuration files from directory
ARG)},
{username, opt_auth_username, 1,
+1
Bert
-Original Message-
From: Julian Foad [mailto:julianf...@btopenworld.com]
Sent: woensdag 26 maart 2014 12:57
To: Bert Huijben
Cc: dev@subversion.apache.org
Subject: Re: svn commit: r1544182 - in /subversion/trunk/subversion:
include/svn_client.h libsvn_client/cat.c libsvn_client
-Original Message-
From: C. Michael Pilato [mailto:cmpil...@collab.net]
Sent: woensdag 19 maart 2014 16:27
To: Bert Huijben; 'Branko Čibej'; dev@subversion.apache.org
Subject: Re: r1577170 - remove wildcard handling from some .exe's on
Windows
On 03/19/2014 11:12 AM, Bert
either... see this error.
Any BDB experts around that can find a way how to fix this?
Bert
was not *more expensive* than the old check.
But now the code is updated to still wait the old long time... after we spend
time determining whether we should wait at all.
Bert
of the file.
The svn diff didn't make that obvious to me, but the conflict resolution was
as simple as just including both blocks.
Bert
From: Stefan Fuhrmann [mailto:stefan.fuhrm...@wandisco.com]
Sent: zaterdag 22 maart 2014 17:39
To: Bert Huijben
Cc: Subversion Development; Stefan
.
Is the final result still a delta stream against the source in the cases where
it was before?
(Or are there cases where we now get a delta against the empty stream where we
didn't... or the other way around?)
Bert
,
Can you sync this branch with 1.8.x, resolving the conflicts caused by
backporting the 1.8.x-r1574868 branch?
At first sight the text conflicts don't look trivial... But that is probably
because I didn't review this patch yet.
Bert
), not the time of the lock creation.)
Bert
-Original Message-
From: Julian Foad [mailto:julianf...@btopenworld.com]
Sent: woensdag 19 maart 2014 11:44
To: Bert Huijben
Cc: dev@subversion.apache.org
Subject: Re: r1577170 - remove wildcard handling from some .exe's on
Windows
URL: http://svn.apache.org/r1577170
Log
quite surprised that the responses in this thread are all from the common
unix developers on this list; not the Windows developers, while Cygwin (which
is how many unix devs work on Windows) explicitly doesn’t apply this behavior
to the ‘svn’ they provide.
Bert
From: C
-Original Message-
From: Branko Čibej [mailto:br...@wandisco.com]
Sent: woensdag 19 maart 2014 15:50
To: Bert Huijben; 'C. Michael Pilato'; dev@subversion.apache.org
Subject: Re: r1577170 - remove wildcard handling from some .exe's on
Windows
On 19.03.2014 14:59, Bert Huijben
-Original Message-
From: C. Michael Pilato [mailto:cmpil...@collab.net]
Sent: woensdag 19 maart 2014 16:27
To: Bert Huijben; 'Branko Čibej'; dev@subversion.apache.org
Subject: Re: r1577170 - remove wildcard handling from some .exe's on
Windows
On 03/19/2014 11:12 AM, Bert
-Original Message-
From: Ivan Zhakov [mailto:i...@visualsvn.com]
Sent: dinsdag 18 maart 2014 11:07
To: Bert Huijben
Cc: Branko Čibej; Subversion Development
Subject: Re: Subversion checked-out files not indexed in Windows search
On 13 March 2014 17:14, Bert Huijben b
-Original Message-
From: Branko Čibej [mailto:br...@wandisco.com]
Sent: dinsdag 18 maart 2014 11:37
To: 'Subversion Development'
Subject: Re: Subversion checked-out files not indexed in Windows search
On 18.03.2014 11:20, Bert Huijben wrote:
But the reason we set this flag
move bugs in 1.8 (several that can corrupt wc.db) that were fixed
months ago from trunk... Or the ra_serf fixes...
Bert
?
Bert
Modified: subversion/trunk/subversion/libsvn_fs_fs/pack.c
URL:
http://svn.apache.org/viewvc/subversion/trunk/subversion/libsvn_fs_fs/pa
ck.c?rev=1578176r1=1578175r2=1578176view=diff
==
--- subversion
slowed us down on Windows XP and 7 around the time that
we introduced this flag as the indexer opened the file between us creating and
moving the file)
Bert
-Original Message-
From: Bert Huijben [mailto:b...@qqmail.nl]
Sent: vrijdag 14 maart 2014 23:14
To: 'Philip Martin'
Cc: 'Branko Čibej'; 'Subversion Development'
Subject: RE: Mergeinfo overwritten from successive merges
-Original Message-
From: Philip Martin
if the contents doesn't match)
} }
+{ { opt_force, N_(force deletions even if deleted contents don't match)
As far as I can tell it is contents doesn't or content doesn't, not
contents don't.
I like the s/the contents/deleted contents/
Bert
Can you see if this is related to httpv1 vs httpv2.
I'm guessing that in v1 the checkout request would have resulted in the out of
date error.
Bert
-Original Message-
From: Branko Čibej br...@wandisco.com
Sent: 14-3-2014 06:31
To: Subversion Development dev@subversion.apache.org
... And my
guess would be that we forgot to do that for the proppatch.(v1 supplied base
URL). Most tests in our test suite would check for out of date via the root of
the commit editor... So that would explain things.
But that is 100% without looking at the code...
Bert
(cell phone
node
(as that would require describing a prop change on something that doesn't
exist).
The update after that final commit will remove the propchange though.
I can't really think of a simple, cheap way to fix this without causing a
different bug.
Bert
-Original Message-
From: phi
-Original Message-
From: Philip Martin [mailto:philip.mar...@wandisco.com]
Sent: vrijdag 14 maart 2014 15:17
To: Bert Huijben
Cc: Branko Čibej; Subversion Development
Subject: Re: Mergeinfo overwritten from successive merges
Bert Huijben b...@qqmail.nl writes:
The base
the documentation should note that the array values must be pointers as
that is how you pass the comparison function.
There are pretty common cases where we use apr arrays with struct members,
instead of pointer to struct. (E.g. for property changes)
Bert
-Original Message-
From: Philip Martin [mailto:philip.mar...@wandisco.com]
Sent: donderdag 13 maart 2014 12:32
To: Bert Huijben
Cc: dev@subversion.apache.org
Subject: Re: svn commit: r1577082 -
/subversion/trunk/subversion/tests/cmdline/special_tests.py
Bert Huijben b
on !Windows as well instead of
handling it magically in several scripts.
Bert
-Original Message-
From: 'Stefan Sperling' [mailto:s...@elego.de]
Sent: donderdag 13 maart 2014 14:34
To: Bert Huijben
Cc: 'Branko Čibej'; dev@subversion.apache.org
Subject: Re: SVN_INTL_LIBS is not defined
On Thu, Mar 13, 2014 at 01:19:18PM +0100, Bert Huijben wrote
-Original Message-
From: Stefan Sperling [mailto:s...@elego.de]
Sent: maandag 10 maart 2014 14:54
To: Bert Huijben
Cc: comm...@subversion.apache.org; dev@subversion.apache.org
Subject: Re: svn commit: r1575525 - in /subversion/trunk/subversion:
libsvn_client/export.c tests/cmdline
should keep the root (its abspath) in the
baton instead of a boolean. As that is the only proper way to get this working
recursively. (passing force overwrite for externals for the other problem would
be an ugly hack)
Bert
Sent from Windows Mail
From: Stefan Sperling
Sent: Saturday
This 'check via explicit target’ would also fix calling 'svn export’ directly
on a file external, which is probably still broken after this patch.
Bert
Sent from Windows Mail
From: Bert Huijben
Sent: Saturday, March 8, 2014 8:55 PM
To: comm...@subversion.apache.org, dev
the description on *nix, the only
reason to fill in the description was to produce Win32 resources.
I'm trying to find where we produce the pkg-config files, but I can't find
it... Do you know where we write these?
Bert
UnknownDependency(Exception):
We don't know how to deal with the dependent to link it in.
pass
Bert
, but I was using a nightly build from last
week.
This should be fixed by r1575018.
I think this regression was introduced a few months ago (around December
2013), while reducing the memory used by the svn_ra_get_dir() api for serf.
Bert
| svnserve | serf] and confirmed
that the buildscripts are compatible with the different build variants for
SlikSvn, SharpSvn, etc.
apr 1.5.0
apr-util 1.5.3
db 4.4.20
expat 2.0.1
httpd 2.4.7
java-sdk 1.7.0_45
mod_dav 2.4.7
openssl 1.0.1f
sasl 2.1.25
sqlite 3.8.3.1
zlib 1.2.8
Bert
-Original Message-
From: Stefan Sperling [mailto:s...@apache.org]
Sent: donderdag 27 februari 2014 13:27
To: Bert Huijben
Cc: dev@subversion.apache.org
Subject: Re: svn commit: r1572105 - in /subversion/trunk/subversion:
svn/auth-cmd.c svn/cl.h svn/svn.c tests/cmdline
the next openssl crashes harder on another platform.
Bert
I would guess that the actual keys are one byte longer than the length passed
to the hash function: a final '\0' to make them a c string.
Does using a strdup instead of that memdup fix the problem? (Or memdup with +1)
Bert
-Original Message-
From: Philip Martin philip.mar
here.
Bert
-no-log -c R:\
To run the http tests as a non administrator.
(Using 127.0.0.1 instead of localhost and --httpd-no-log make things
noticeably faster)
Bert
++ 2010 compiler.
I used:
Apr 1.5.0
Apr-Util 1.5.3
BDB 4.4.20
ZLib 1.2.8
Serf 1.2.0
Cyrus Sasl 2.1.23
Swig 2.0.4
Httpd 2.2.26
Sqlite 3.7.17
For this test run,
Bert
if they want to test fsfs or fsx
+if fs_type != 'bdb':
all_tests = gen_obj.test_progs + gen_obj.scripts
Applied a similar fix in r1571028.
Thanks for finding this problem. (I always build with BDB, so I didn't know
about this list)
Bert
change while the repository form didn’t.
Blame, especially with whitespace ignores might not be interested in these
cases, but there are api users that use the file revs api that are interested.
Bert
Sent from Windows Mail
From: Stefan Fuhrmann
Sent: Saturday, February 22, 2014
but don't have a certificate you would get the question over and over again.
Bert
Thanks,
Thomas Å.
-Original Message-
From: Julian Foad [mailto:julianf...@btopenworld.com]
Sent: woensdag 19 februari 2014 13:58
To: Bert Huijben
Cc: dev@subversion.apache.org
Subject: Re: svn commit: r1569697 - create a diff processor in one
place...
Hi Bert. I like what you're doing here
running tests on a fix that detects the case where the parent
directory is not readable, and in that case restores the old behavior (of
not being able to show the directory replacement) while keeping the improved
behavior when you do have the necessary rights.
Bert
[fsfs | bdb] x [local | svn | dav]
Signatures are already committed.
Bert
knowledge about both the fs
internals and this very specific api that drives blame (and several third
party tools that want to obtain all versions of a file in the most efficient
way possible).
Bert
From: Stefan Fuhrmann [mailto:stefan.fuhrm...@wandisco.com]
Sent: zaterdag 15
with a few calls to
apr_atomic_inc32().
Bert
.
If you just drop that one (that you only used for debugging your code. There
is no single read of that variable) the rest can be an atomic increment.
Bert
From: Stefan Fuhrmann [mailto:stefan.fuhrm...@wandisco.com]
Sent: donderdag 13 februari 2014 23:57
To: Ivan Zhakov
Cc
Sorry for the big number of small replies.
And the 'cache-total_reads++;' calls?
Doesn't that need a similar handling?
Bert
From: Bert Huijben [mailto:b...@qqmail.nl]
Sent: vrijdag 14 februari 2014 00:58
To: 'Stefan Fuhrmann'; 'Ivan Zhakov'
Cc: 'Subversion
.
(It is a known issue in 1.8.x since 1.8.0 that I recently uncovered though)
I don't see these errors when I test with SQLite 3.8.3.
Bert
Sent from Windows Mail
From: Branko Čibej
Sent: Saturday, February 8, 2014 11:42 PM
To: 'Subversion Development'
Many tests
and comment are on most cases the only thing users are
really interested in. The token itself doesn't give you any privileges…
Bert
Sent from Windows Mail
From: Philip Martin
Sent: Thursday, February 6, 2014 8:41 PM
To: 'Subversion Development'
Philip Martin philip.mar
should change the scope of the date lookup as the first step.
Perhaps we should try to add a path to the resolve function and see where that
helps for specific cases?
After that we can look at the problems this will introduce in the backend.
Bert
From: Stefan
-Original Message-
From: Greg Stein [mailto:gst...@gmail.com]
Sent: dinsdag 28 januari 2014 12:12
To: Bert Huijben
Cc: Branko Čibej; dev@subversion.apache.org
Subject: Re: 1.9 issues
On Fri, Jan 24, 2014 at 3:17 PM, Bert Huijben b...@qqmail.nl wrote:
...
And in many
-Original Message-
From: Greg Stein [mailto:gst...@gmail.com]
Sent: dinsdag 28 januari 2014 14:20
To: Bert Huijben
Cc: Branko Čibej; dev@subversion.apache.org
Subject: Re: 1.9 issues
On Tue, Jan 28, 2014 at 6:02 AM, Bert Huijben b...@qqmail.nl wrote:
-Original Message
Serf callback (through Subversion) we can just implement a proper
store (either in Subversion or in the clients that implement this on Windows)
as we have access to the uri of the repository. In the OpenSSL layer we don't
have access to that.
Bert
but would be
a useful extra check if we can implement it efficiently it without a
table scan. Is sqlite3_last_insert_rowid() the function we want?
Bert pointed out that is the wrong function and there isn't really a
suitable
function. So to do this check we need something like
Julian's
are continuously growing, but it is more disk anyway.
I would say these users should just delete the sqlite database (or update the
sqlite file manually) in the case where they are changing the repository
structure in unsupported ways.
So +1 on the suggestion of just removing the dead code.
Bert
://svn.apache.org/viewvc?view=revisionrevision=r1367683
[3] http://svn.apache.org/viewvc?view=revisionrevision=r1503742
[4] http://svn.apache.org/viewvc?view=revisionrevision=r1404675
[5] http://svn.apache.org/viewvc?view=revisionrevision=r1404708
Thanks,
Bert
Thanks and regards,
Evgeny
I don't think locks.c was removed with this commit. Only some code was removed
from it.
Bert
Sent from Windows Mail
From: Ben Reser
Sent: Tuesday, January 21, 2014 12:23 AM
To: comm...@subversion.apache.org
Author: breser
Revision: 1551910
Modified property: svn:log
for this patch, but svn_xml_make_open_tag() allows adding xml
arguments optionally. When the value is passed as NULL (or probably
SVN_VA_NULL... I see a TODO in ra_serf), the argument will be skipped, but the
processing will continue on the next argument.
Bert
/* authorxxx/author
of your (and our) time.
So, perhaps it would help if you provide the errors you see and as much
information as possible to allow us to help you.
Bert
From: Oikonomou Ioannis [mailto:ioannis.oikono...@trasysgroup.com]
Sent: donderdag 16 januari 2014 12:23
To: dev
the status of the properties and entries of
the directory are tied to that.
Bert
-Original Message-
From: stef...@apache.org stef...@apache.org
Sent: 14-1-2014 23:09
To: comm...@subversion.apache.org comm...@subversion.apache.org
Subject: svn commit: r1558224 - in
/subversion/trunk
Hi Mojca,
Thanks for checking out trunk. I'm guessing I found the same problem last week
when I tried running our test suite against a pre 1.7 server. I fixed the
problem I found in r1556974.
Bert
Sent from Windows Mail
From: Mojca Miklavec
Sent: Monday, January 13
(E.g
../!svn/rvr..) for something that is not really a repository.
Your segfault should be properly fixed with r1557094, but we should probably
also run some tests on a repository that is at the server root (instead of
only a parent path).
Bert
-Original Message-
From: Ivan Zhakov [mailto:i...@visualsvn.com]
Sent: donderdag 9 januari 2014 10:18
To: Bert Huijben
Cc: Subversion Development
Subject: Re: svn commit: r1555716 -
/subversion/trunk/subversion/libsvn_ra_serf/update.c
On 7 January 2014 18:29, Bert Huijben b
is empty).
If you want to look at improving it now, feel free. I'm first working on
switching the update logic to the new xml parser. (And will perform a further
cleanup including these points round after that).
Thanks for your reviews!
Bert
Sent from Windows Mail
From: Ivan Zhakov
of SharpSvn we would just have found the fsfs symptoms months from now
when releasing 1.9)
Bert
-Original Message-
From: Branko Čibej br...@wandisco.com
Sent: 05/01/2014 04:51
To: Subversion Development dev@subversion.apache.org
Subject: Re: svn commit:
r1555350-/subversion/trunk/subversion
,
Subclipse, etc. Etc.
That all our command line tools had to be patched shows a pattern to a bigger
problem of new pool lifetime issues that we should really fix before 1.9 or we
will have many segfaults to fix for the 1.9 patch releases.
Bert
-Original Message-
From: Bert Huijben b
,
but especially generated bindings (like swig) have strange pool handling much
longer.
Bert
-Original Message-
From: stef...@apache.org stef...@apache.org
Sent: 4-1-2014 15:05
To: comm...@subversion.apache.org comm...@subversion.apache.org
Subject: svn commit: r1555350 -/subversion/trunk
in a 1.9 or 1.10 release.
As is the cache logic is flawed if it has all these new pool usage requirements
that we have to apply all through our code. (Not just here but svnserve and a
few other applications were changed for similar things)
Bert
-Original Message-
From: Stefan Fuhrmann
there might be
another bug that you fixed… but by reading your log message and change this
looks like covering up for a design problem)
We can't fix all code using our APIs until 2.0. Until then our existing API
surface should cover for requirements changes.
Bert
Sent from Windows Mail
be very hard (if need
impossible) to implement on the automatically generated swig bindings. But
perhaps it can be done in a higher layer abstraction.
It would be very nice if somebody invested some time to take a look at this
someday...
Bert
We have a baton in the callback, so no reason for bottle magic.
Bert
-Original Message-
From: Gabriela Gibson gabriela.gib...@gmail.com
Sent: 23-12-2013 20:14
To: Subversion Development dev@subversion.apache.org
Subject: prop edit: lost user edit bug
Hi,
I found the other day
libsvn_client is safe as we always construct these
sessions based on the state in svn_client_ctx_t, which must outlive both Ra
sessions.
Bert
Sent from Windows Mail
From: Julian Foad
Sent: Friday, December 20, 2013 1:12 PM
To: Bert Huijben
Cc: Subversion Development
URL: http
All those options work, but you could also just follow the Ra code and
implement a callback with the result. The Ra layers can then batch or stream
the result as they wish.
Bert
Sent from Windows Mail
From: Philip Martin
Sent: Thursday, December 19, 2013 3:28 PM
svnserve based servers where we always obtain all details.
Bert
arguments, parsing the format string, etc.).
In this case I'm really wondering why we change working code...?
Bert
:
// A already defined as variable
static APR_INLINE int A()
{
return A();
}
Luckily you just get a compiler error for the duplicated symbol, otherwise
it would just implement endless recursion instead of an actual function
call.
Bert
From: daniel.lescoh
them but... I
don't see this as an excuse to use a bad design for the fs layer until 2.0.
Bert
-Original Message-
From: Stefan Fuhrmann stefan.fuhrm...@wandisco.com
Sent: 8-12-2013 11:09
To: Bert Huijben b...@qqmail.nl
Cc: Subversion Development dev@subversion.apache.org
Subject: Re
it properly handles alignment within the structure. (It
probably works now, but the old code was far less sensitive for future changes
to this struct)
Bert
@@ -356,6 +358,7 @@ svn_fs_x__id_txn_create_root(const svn_f
id-generic_id.vtable = id_vtable;
id-generic_id.fsap_data
-Original Message-
From: Bert Huijben [mailto:b...@qqmail.nl]
Sent: vrijdag 6 december 2013 19:14
To: 'William A. Rowe Jr.'; 'Stefan Fuhrmann'
Cc: 'APR Developer List'; 'Stefan Fuhrman'; 'Philip Martin'; 'Subversion
Development'
Subject: RE: Race condition
.
Bert
parts of this patch are to make sure some move behavior patches I
have
kept local for some time don't have to keep in sync with more files than
necessary.
* subversion/libsvn_wc/wc_db_update_move.c
(replace_moved_layer): Properly extend parent delete.
Hi Bert,
Can you please
-Original Message-
From: William A. Rowe Jr. [mailto:wr...@rowe-clan.net]
Sent: vrijdag 6 december 2013 18:24
To: Stefan Fuhrmann
Cc: Bert Huijben; APR Developer List; Stefan Fuhrman; Philip Martin;
Subversion Development
Subject: Re: Race condition in APR_DECLARE_LATE_DLL_FUNC
to be alsways there instead of loading them
dynamically.
Bert
the command 'skeleton', but perhaps
you should leave that here as an (undocumented) alias for strip?
Bert
this isn't supported for 1.5 repositories?
The get_file_revs api is much older than 1.5. Just the mergeinfo awareness was
added in 1.5.
Did you just mask a regression or is there some other problem?
Bert
in a
repository (not a path) and it would be very inefficient in our current design
to update it to walk history for a binary search.
Bert
From: Branko Čibej [mailto:br...@wandisco.com]
Sent: vrijdag 29 november 2013 16:50
To: dev@subversion.apache.org
Subject: Re
the new in 1.8 'EPHEMERAL txn' properties for some of
the client to server communication.
Bert
Future thought (triggered by this log message): How should we handle merge info
containing not normalized paths?
Bert
-Original Message-
From: br...@apache.org br...@apache.org
Sent: 24-11-2013 05:11
To: comm...@subversion.apache.org comm...@subversion.apache.org
Subject: svn commit
Strange I debugged this on Windows and I'm sure it worked there. On which
platform was it broken and does it still work on Windows?
Bert
-Original Message-
From: stef...@apache.org stef...@apache.org
Sent: 23-11-2013 01:04
To: comm...@subversion.apache.org comm
Ok, reviewed apr’s usage of these two variables.
Using fname is the better case on all platforms. name is only set on the CP/M
based platforms, not on the posix platforms.
Bert
From: Bert Huijben [mailto:b...@qqmail.nl]
Sent: zaterdag 23 november 2013 13:00
To: dev
a stringbuffer in this way is not really the appropriate
operation?
Bert
Sent from Windows Mail
From: astie...@apache.org
Sent: Saturday, November 23, 2013 9:46 PM
To: comm...@subversion.apache.org
Author: astieger
Date: Sat Nov 23 20:46:50 2013
New Revision: 1544878
handle these things. A really nice case where the ACL's make our
code easier, by not letting user code worry about this.
(There is also a per user temporary directory instead of one for all users)
Bert
-Original Message-
From: Bert Huijben [mailto:b...@qqmail.nl]
Sent: vrijdag 22 november 2013 15:21
To: 'Philip Martin'; 'Branko Čibej'
Cc: dev@subversion.apache.org
Subject: RE: Fwd: [PATCH] Fix for `svn patch` changing permissions of
patched files
Branko: On Windows we never
501 - 600 of 1808 matches
Mail list logo