Philip Martin philip.mar...@wandisco.com writes:
The txn_dir_cache gets cleared when the pool associated with the
txn_root is cleared
The net result is that the txn_dir_cache doesn't really work with
apache/mod_dav_svn. Import a directory containing a large number of
small files and the rate
Alexey Neyman sti...@att.net writes:
Indeed, calling svn_fs_txn_root() after execution of the pre-commit
hook in svn_repos_fs_commit_txn() makes Subversion behave in the same
way as it did in 1.6.
Just so I can be clear about what is happening: can you confirm that you
are not using
Philip Martin philip.mar...@wandisco.com writes:
1. client sends MKCOL
2. apache process MKCOL and populates txn_dir_cache
3. apache sends MKCOL response
4. apache clears MKCOL request pool
5. client sends MERGE
6. apache process MERGE
5 can happen before 4 since two processes
On Tuesday, August 12, 2014 09:49:08 AM Philip Martin wrote:
Alexey Neyman sti...@att.net writes:
Indeed, calling svn_fs_txn_root() after execution of the pre-commit
hook in svn_repos_fs_commit_txn() makes Subversion behave in the
same
way as it did in 1.6.
Just so I can be clear about
On 12.08.2014 17:28, s...@apache.org wrote:
Author: stsp
Date: Tue Aug 12 15:28:49 2014
New Revision: 1617507
URL: http://svn.apache.org/r1617507
Log:
* subversion/libsvn_subr/debug.c
(debug_vprintf): Increase local buffer size so clients built with
SQLITE3_DEBUG don't assert in
On Tue, Aug 12, 2014 at 12:59 PM, Philip Martin philip.mar...@wandisco.com
wrote:
Philip Martin philip.mar...@wandisco.com writes:
1. client sends MKCOL
2. apache process MKCOL and populates txn_dir_cache
3. apache sends MKCOL response
4. apache clears MKCOL request pool
5.
On 7 August 2014 00:47, Stefan Fuhrmann stefan.fuhrm...@wandisco.com wrote:
On Fri, Aug 1, 2014 at 7:37 PM, Ivan Zhakov i...@visualsvn.com wrote:
On 22 July 2014 01:48, stef...@apache.org wrote:
Author: stefan2
Date: Mon Jul 21 21:48:35 2014
New Revision: 1612405
URL:
Stefan,
I've noticed another severe issue in named atomic infrastructure used
by revprop caching code while reviewing r1611379 fix:
svn_atomic_namespace__create() doesn't release file lock and
process-wide mutex (!) on error in libsvn_subr\named_atomic.c:446.
Which is basically mean that server
I'd like to always make update/switch target revisions available to
the conflict resolver. We currently don't record this information
for tree conflict victims which are not present in the target revision.
For instance, 'svn info' might show:
Tree conflict: local file edit, incoming file delete
On 11 August 2014 20:29, Stefan Fuhrmann stefan.fuhrm...@wandisco.com wrote:
On Mon, Aug 4, 2014 at 4:23 PM, Ivan Zhakov i...@visualsvn.com wrote:
On 4 August 2014 17:48, Branko Čibej br...@wandisco.com wrote:
On 04.08.2014 15:22, Ivan Zhakov wrote:
On 2 August 2014 23:13,
On 11 August 2014 20:51, Ben Reser b...@reser.org wrote:
On 8/11/14 1:59 AM, Ivan Zhakov wrote:
My primary concerns was that with svn_checksum_to_cstring_display2()
we have to care at every place to use proper flags to get some
canonical representation for protocol/storage.
How about
I don't see a problem with not allowing a repository location, but it doesn't
make sense to me to have a revision in that case. You either have both, or have
none.
If you want to tell that the node didn't exist in a revision it makes more
sense to me to tell that it did’t exist at a specific
On 12.08.2014 21:56, Ivan Zhakov wrote:
On 11 August 2014 20:51, Ben Reser b...@reser.org wrote:
On 8/11/14 1:59 AM, Ivan Zhakov wrote:
My primary concerns was that with svn_checksum_to_cstring_display2()
we have to care at every place to use proper flags to get some
canonical representation
On Sun, Aug 10, 2014 at 3:56 PM, Evgeny Kotkov evgeny.kot...@visualsvn.com
wrote:
Hi Stefan,
Thanks for having a the (missing) instance ID issue.
From initial review, I have 1 objection and 2 issues
that your patch does not address, yet.
Thank you for reviewing this patch. I did my
On 13 August 2014 01:24, Branko Čibej br...@wandisco.com wrote:
On 12.08.2014 21:56, Ivan Zhakov wrote:
On 11 August 2014 20:51, Ben Reser b...@reser.org wrote:
On 8/11/14 1:59 AM, Ivan Zhakov wrote:
My primary concerns was that with svn_checksum_to_cstring_display2()
we have to care at
Arrgs!! Random key combo made me sent the mail before completing it ...
On Tue, Aug 12, 2014 at 11:29 PM, Stefan Fuhrmann
stefan.fuhrm...@wandisco.com wrote:
On Sun, Aug 10, 2014 at 3:56 PM, Evgeny Kotkov
evgeny.kot...@visualsvn.com wrote:
Hi Stefan,
Thanks for having a the (missing)
On 8/12/14 12:56 PM, Ivan Zhakov wrote:
My concerns are the following:
1. Avoid unrelated branch changes
2. Have some function for converting checksum to canonical form:
a) one option is just leave svn_checksum_to_cstring_display() and
add svn_checksum_to_cstring_display_ex()
b)
Since WC-NG we tried not to introduce new functions with flag arguments as in
general functions like that are hard to maintain, while it is easy to rev
functions to add another separate argument. (Another less preferred option is
using a struct with separate args)
I remember arguments from
On 8/12/14 3:55 PM, Bert Huijben wrote:
Since WC-NG we tried not to introduce new functions with flag arguments as in
general functions like that are hard to maintain, while it is easy to rev
functions to add another separate argument. (Another less preferred option is
using a struct with
On 12.08.2014 23:47, Ben Reser wrote:
On 8/12/14 12:56 PM, Ivan Zhakov wrote:
My concerns are the following:
1. Avoid unrelated branch changes
2. Have some function for converting checksum to canonical form:
a) one option is just leave svn_checksum_to_cstring_display() and
add
On Tue, Aug 12, 2014 at 11:47 PM, Stefan Fuhrmann
stefan.fuhrm...@wandisco.com wrote:
Arrgs!! Random key combo made me sent the mail before completing it ...
On Tue, Aug 12, 2014 at 11:29 PM, Stefan Fuhrmann
stefan.fuhrm...@wandisco.com wrote:
On Sun, Aug 10, 2014 at 3:56 PM, Evgeny
On 8/12/14 4:39 PM, Branko Čibej wrote:
This thing against flags and booleans that change API behaviour does have a
very good argument going for it: it makes the code less obvious for the sake
of
saving a few function names.
Consider the standard example, which (these days) would be an API
22 matches
Mail list logo
