erally desirable pattern of
> moving towards keeping repos_root and repos_relpath separate.
> Especially in the RA and repository side APIs, I assume, but makes sense
> to do so in other parts of the libraries as well.
>
> Any concerns? If not I'll try to do it.
>
+1, makes sense to me.
--
Ivan Zhakov
VisualSVN Team
ffer of a given type.
> This prevents the need to add casts whenever grabbing bits of allocated
> memory.
It might be useful to add alloc function without argument to allocate
just one element of given type since it could detect size of
parametrized type.
--
Ivan Zhakov
VisualSVN Team
to be installed and running. Some comes with just a zip
archive, others have some installation and configuration scripts.
That's what customers looking for.
3. From my point of view, it is more important for users to explicitly
mark packages that requires registration and email for download.
--
Ivan Zhakov
On Thu, Oct 28, 2010 at 16:42, Daniel Shahaf wrote:
> C. Michael Pilato wrote on Wed, Oct 27, 2010 at 10:11:20 -0400:
>> On 10/27/2010 09:41 AM, Ivan Zhakov wrote:
>> > 2. Let's think about users: they have different needs for Subversion:
>> > just a client,
as the prefix of the returned KEY matches LOOKUP_PATH we
> know it is either LOOKUP_PATH or a decendant thereof. */
> - if (strcmp(path, "/") != 0)
> + if (svn_fspath__is_root(path, strlen(path)))
I think you missed "!" there. It should be "if
(!svn_fspath__is_root(path, strlen(path)))".
> lookup_path = apr_pstrcat(pool, path, "/", (char *)NULL);
--
Ivan Zhakov
VisualSVN Team
its content */
> const char * admdirpath = svn_wc__adm_child(local_abspath, NULL, pool);
> SetFileAttributes(admindirpath, GetFileAttributes(admindirpath)
> |FILE_ATTRIBUTE_NOT_CONTENT_INDEXED);
> #endif
>
>
>
I'm big +1 for marking .svn not indexed by Windows indexing service.
--
Ivan Zhakov
VisualSVN Team
ly mentioned flags, the
FILE_ATTRIBUTE_TEMPORARY attribute does tell the system to hold as
much as possible in the system cache without writing and therefore may
be of concern for certain applications.
]]]
May be we should submit a patch to SQLite to use this attribute?
[1] http://msdn.microsoft.com/en-us/library/aa363858%28v=vs.85%29.aspx
--
Ivan Zhakov
VisualSVN Team
On Wed, Feb 9, 2011 at 17:40, Ivan Zhakov wrote:
> On Tue, Feb 8, 2011 at 12:50, Bert Huijben wrote:
>>> -Original Message-
>>> From: Branko Čibej [mailto:[email protected]] On Behalf Of Branko Cibej
>>> Sent: dinsdag 8 februari 2011 4:39
>>> To: de
On Wed, Feb 9, 2011 at 23:06, Stefan Sperling wrote:
> On Wed, Feb 09, 2011 at 10:56:35PM +0300, Ivan Zhakov wrote:
>> On Wed, Feb 9, 2011 at 17:40, Ivan Zhakov wrote:
>> > On Windows there is FILE_ATTRIBUTE_TEMPORARY [1] attribute, which
>> > prevents
Optionally we can add Apache environment variable to be able
configure compression level based on client IP address.
> integrate-io-improvements
> - changes were already discussed when svn_io_file_read_full2()
> got introduced on the performance branch
Looks fine. I'm +1 to merge.
--
Ivan Zhakov
On Fri, Feb 11, 2011 at 03:51, Stefan Fuhrmann wrote:
> On 10.02.2011 22:15, Ivan Zhakov wrote:
>>
>> On Thu, Feb 10, 2011 at 03:25, Stefan Fuhrmann
>> wrote:
>>>
>>> Hi there,
>>>
>>> Those among you following the commits will have n
update status of "Diff/Blame optimizations" task on
roadmap page [1] since diff-optimizations-bytes branch merged to
trunk.
[1] http://subversion.apache.org/roadmap.html
--
Ivan Zhakov
PS: See attached patch in case you'd like test performance.
--
Ivan Zhakov
Index: subversion/libsvn_ra_serf/property.c
===
--- subversion/libsvn_ra_serf/property.c(revision 1070442)
+++ subversion/libsvn_ra_serf/proper
On Mon, Feb 14, 2011 at 21:14, C. Michael Pilato wrote:
> On 02/14/2011 01:57 PM, Ivan Zhakov wrote:
>> So I'm propose just to remove DAV properties cache from ra_serf.
>> Objections? Comments?
>>
>> PS: See attached patch in case you'd like test performance
On Tue, Feb 15, 2011 at 11:23, Justin Erenkrantz wrote:
> On Mon, Feb 14, 2011 at 7:57 PM, Ivan Zhakov wrote:
>> Of course these problems can be fixed, but I'm not sure that we need
>> this code. Since in svn 1.7 we have HTTPv2 which doesn't use so many
>> PRO
On Tue, Feb 15, 2011 at 14:26, Mark Phippard wrote:
> On Tue, Feb 15, 2011 at 7:21 AM, Ivan Zhakov wrote:
>
>> Thanks for feedback!
>>
>> I removed DAV properties cache from ra_serf in r1070861.
>
> Does this improve the memory usage on checkout that others have br
On Tue, Feb 15, 2011 at 12:44, Hyrum K Wright wrote:
> On Tue, Feb 15, 2011 at 6:42 AM, Greg Stein wrote:
>> On Mon, Feb 14, 2011 at 14:14, C. Michael Pilato wrote:
>>> On 02/14/2011 01:57 PM, Ivan Zhakov wrote:
>>>> Hi,
>>>>
>>>> Current
roperties in one response? Does it make sense to implement
send-all='true' mode in ra_serf?
--
Ivan Zhakov
and leave a note to change
> them to s.a.o in the 2.0 timeframe (if that ever happens!).
>
+1.
--
Ivan Zhakov
> of cacheable items.
>
Hi Stefan,
Why serialize/deserialize mechanism is better than data copying
function? For me implementing dup function is much easier and better
for performance than serialize/deserialize. Is it possible to keep
both mechanism in inprocess cache?
--
Ivan Zhakov
l *)(void *,apr_off_t *,apr_off_t *,const
svn_diff_datasource_e *,apr_size_t)'
]]
--
Ivan Zhakov
mplementing cache
of open file handles in Subversion.
--
Ivan Zhakov
bit. "Only." I seem to recall that's not even
> hard to do, but my last battle with Windows filesystem internals was
> more than 10 years ago. If you can't find relevant docs, you could try
> asking APR for that functionality. I'm sure Will Rowe will give you a
> dozen reasons why doing that is not a good idea, and also explain how to
> do it. :)
>
Problem of re-installing file over marked for deletion file can be
solved using the following trick:
1. Rename file to temporary name.
2. Delete it
--
Ivan Zhakov
VisualSVN Team
2011/2/26 Branko Čibej :
> On 26.02.2011 07:32, Ivan Zhakov wrote:
>> 2011/2/26 Branko Čibej :
>>> On 25.02.2011 16:53, Julian Foad wrote:
>>>> On Thu, 2011-02-24, Branko Čibej wrote:
>>>>> On 24.02.2011 18:03, Julian Foad wrote:
>>>>&g
2011/2/26 Branko Čibej :
> On 26.02.2011 10:50, Ivan Zhakov wrote:
>> 2011/2/26 Branko Čibej :
>>> On 26.02.2011 07:32, Ivan Zhakov wrote:
>>>> Problem of re-installing file over marked for deletion file can be
>>>> solved using the following trick:
>
svn_error_clear(err);
> + else
> + SVN_ERR(err);
> + file_abspath = temp_abspath;
> +#endif
> +
> + SVN_ERR(svn_io_remove_file2(file_abspath, ignore_enoent, scratch_pool));
> +
> + return SVN_NO_ERROR;
> +}
> +
[...]
--
Ivan Zhakov
VisualSVN Team
On Tue, Mar 1, 2011 at 23:30, Julian Foad wrote:
> On Tue, 2011-03-01 at 21:00 +0100, Stefan Sperling wrote:
>> On Tue, Mar 01, 2011 at 10:07:58PM +0300, Ivan Zhakov wrote:
>> > On Tue, Mar 1, 2011 at 20:47, wrote:
>> > > Author: julianfoad
>> > >
methods REPORT MERGE MKACTIVITY CHECKOUT PROPFIND
> PROXY-AUTHORIZATION OPTIONS
>
> The same trunk Subversion built at home, not going through a proxy works
> fine...
>
> John.
>
Serf doesn't support HTTPS proxies yet. See serf's issue 49:
http://code.google.com/p/serf/issues/detail?id=49
--
Ivan Zhakov
ts especially of ra_serf.
Performance degradation of svn import over ra_serf looks very strange.
Could you please provide more details about your configuration?
--
Ivan Zhakov
On Tue, Mar 8, 2011 at 17:01, John Beranek wrote:
> On 08/03/11 09:34, Ivan Zhakov wrote:
>> On Tue, Mar 8, 2011 at 12:21, John Beranek wrote:
>>> On 08/03/11 05:34, Justin Erenkrantz wrote:
>>>> On Mon, Mar 7, 2011 at 3:26 PM, John Beranek wrote:
>>>&
has not had any exposure in the real world yet.
>
> So I think it would be better if the cache was off by default, and must be
> enabled in a configuration file. I suppose svnserve and mod_dav_svn
> configuration files would be the best location for this, because the
> purpose of t
On Tue, Mar 8, 2011 at 17:11, Mark Phippard wrote:
> On Tue, Mar 8, 2011 at 9:07 AM, Ivan Zhakov wrote:
>> On Tue, Mar 8, 2011 at 17:01, John Beranek wrote:
>>> On 08/03/11 09:34, Ivan Zhakov wrote:
>>>> On Tue, Mar 8, 2011 at 12:21, John Beranek wrote:
&
On Tue, Mar 8, 2011 at 17:20, Mark Phippard wrote:
> On Tue, Mar 8, 2011 at 9:14 AM, Ivan Zhakov wrote:
>>>> [1] http://serf.googlecode.com/svn/trunk/
>>>
>>> Don't the tests show that trunk is slightly faster than 1.6? It seems
>>> like the
On Tue, Mar 8, 2011 at 19:49, Philip Martin wrote:
> Ivan Zhakov writes:
>
>> So, as a comparison, I ran the same tests to a localhost trunk(r1078338)
>> server. Fedora 14 x86_64, Apache 2.2.17.
>>
>> trunk (http-library=neon):
>> real 0m20.785s
>
sed by serf over neon.
>
> I don't know why that system call takes so long, but that's where the
> extra time goes.
>
That's really interesting. Could you provide similiar strace for neon?
--
Ivan Zhakov
On Tue, Mar 8, 2011 at 22:07, Greg Stein wrote:
> On Tue, Mar 8, 2011 at 12:34, Ivan Zhakov wrote:
>>...
>> It seems I found reason why ra_serf is slower than ra_neon. ra_serf
>> sends CHECKOUT request for _each_ folder and file that being imported,
>> while ra_ne
On Thu, Mar 10, 2011 at 05:36, Greg Stein wrote:
> On Wed, Mar 9, 2011 at 15:17, Ivan Zhakov wrote:
>> On Tue, Mar 8, 2011 at 22:07, Greg Stein wrote:
>>> On Tue, Mar 8, 2011 at 12:34, Ivan Zhakov wrote:
>>>>...
>>>> It seems I found reason w
On Thu, Mar 10, 2011 at 01:23, John Beranek wrote:
> On 09/03/2011 20:17, Ivan Zhakov wrote:
>> On Tue, Mar 8, 2011 at 22:07, Greg Stein wrote:
>>> On Tue, Mar 8, 2011 at 12:34, Ivan Zhakov wrote:
>>>> ...
>>>> It seems I found reason why ra_serf
t; trunk(r1080029), 10 iterations, from a localhost trunk(r1080029) server
> to a local disc:
> ra_neon: 14.94
> ra_serf: 15.76
>
Hi John,
It better to separate network layer tests from WC layer. The following
operations are mostly depends on network layer and server performance:
1. svn ls / svn ls -R
2. svn export
3. svn log
The following operations are client side:
1. svn st
2. svn diff (when working copy is modified)
--
Ivan Zhakov
using the actual network stack... -- justin
>
Unfortunately ra_serf will be worse than ra_neon, because it uses
skelta style update editor with many GET/PROPFIND requests. While
ra_neon just uses one REPORT request with large response. I have task
in my todo list to implement non-skelta (send-all) mode update editor
in ra_serf to make performance comparable.
--
Ivan Zhakov
when working with pre-1.7 servers and improve
performance over WAN.
]]]
http://svn.apache.org/viewvc?rev=1080245
--
Ivan Zhakov
On Thu, Mar 10, 2011 at 23:02, Greg Stein wrote:
> On Thu, Mar 10, 2011 at 11:50, Ivan Zhakov wrote:
>> On Thu, Mar 10, 2011 at 19:42, Justin Erenkrantz
>> wrote:
>>...
>>> Heh, while we've got you testing stuff...what happens if the server is
>>>
he.org/repos/': XML parse error at
line 1: no element
found (https://svn.apache.org)
[[[
Improve RA neon error message when non XML response received.
* subversion/libsvn_ra_neon/util.c
(wrapper_reader_cb): Wrap neon XML parser error.
]]]
PS: Tests are still running.
--
Ivan Zhakov
Visu
On Mon, Jan 11, 2010 at 1:29 PM, Philip Martin
wrote:
> Ivan Zhakov writes:
>
>> - return ne_xml_parse(pwb->parser, data, len);
>> + parser_status = ne_xml_parse(pwb->parser, data, len);
>> + if (parser_status)
>> + {
>> + /* Pass XML parser
of us that signed for Windows tested Serf with no
> errors. I used 1.6.8 mod_dav_svn for this.
>
I recall some issues that Paul reported that assertions during tests
are ignored on Windows.
--
Ivan Zhakov
VisualSVN Team
l is already in my task queue, so don't worry I discuss it
with Julian and commit it myself.
--
Ivan Zhakov
VisualSVN Team
found in this
> manner. Can I back-port the updates just to accommodate this quirk
> because it has a real practical benefit, even though technically
> speaking it's a stupid thing?
>
Posting actual list of commiters somewhere on web site will solve this
problem since it will have much better PageRank than COMMITERS file
somewhere in repository. Also this saves us from rewriting our
history.
--
Ivan Zhakov
VisualSVN Team
On Wed, Jan 13, 2010 at 6:04 PM, Julian Foad wrote:
> (I've tweaked the subject line to distinguish this part of the thread
> (and this patch) from the first part (which had its own patch).)
>
> On Mon, 2010-01-11, Ivan Zhakov wrote:
>> On Mon, Jan 11, 2010 at 1:29 PM,
On Mon, Jan 25, 2010 at 6:21 PM, Julian Foad wrote:
> On Fri, 2010-01-22, Ivan Zhakov wrote:
>> On Wed, Jan 13, 2010 at 6:04 PM, Julian Foad
>> wrote:
>> > On Mon, 2010-01-11, Ivan Zhakov wrote:
>> > Your new patch now has a more descriptive error message,
od, pwb->parser,
>> + pwb->req->url));
>
> There should be no white space before the function-call open paren.
>
Hi Greg,
Sorry I didn't understand what white space do you mean. There is no
white spaces after SVN_RA_NEON__REQ_ERR macro and arguments are
wrapped on next line.
--
Ivan Zhakov
VisualSVN Team
On Tue, Jan 26, 2010 at 3:01 PM, Greg Stein wrote:
> On Tue, Jan 26, 2010 at 06:59, Ivan Zhakov wrote:
>> On Tue, Jan 26, 2010 at 2:41 PM, Greg Stein wrote:
>>> On Tue, Jan 26, 2010 at 05:12, wrote:
>>>> Author: ivan
>>>> Date: Tue Jan 26 10:12:05 20
html
> Fork from "getting.html", heavily reformat, and add some JavaScript
> cuteness to allow quick jumps to packages.
>
Nice stuff, but this commits makes "Packages" heading on
"packages.html" page redundant and can be removed.
--
Ivan Zhakov
VisualSVN Team
On Tue, Jan 26, 2010 at 3:23 PM, Greg Stein wrote:
> On Tue, Jan 26, 2010 at 07:13, Ivan Zhakov wrote:
>> On Tue, Jan 26, 2010 at 3:01 PM, Greg Stein wrote:
>>> On Tue, Jan 26, 2010 at 06:59, Ivan Zhakov wrote:
>>>> On Tue, Jan 26, 2010 at 2:41 PM, Greg Stein wro
;d be helpful if you
> could fix that (be it inlining, etc, etc.).
>
> Other than that, looks right...
>
Tip: You may add .txt extension to the patch file name to display it
inline. Like fuzzy_patch_notifications.diff.txt
--
Ivan Zhakov
VisualSVN Team
; Would you, or anyone you know, be interested in working on it? I would
> be glad to give you some help and guidance.
>
Agreed. svn_client_* functions spend a lot time on creating RA
connection. My idea was to introduce "connection pool" to hold and
reuse RA connections. This pool can be stored in svn_client_ctx_t or
in separate object.
--
Ivan Zhakov
VisualSVN Team
On Tue, Feb 9, 2010 at 7:51 PM, C. Michael Pilato wrote:
> Ivan Zhakov wrote:
>> On Tue, Feb 9, 2010 at 7:09 PM, Julian Foad wrote:
>>> There are several places where the client should/could re-use a
>>> connection. During a multi-target update command like "svn
to
repository root [1]. I.e.:
[[
svn_ra_svn_open3(repos_url)
svn_ra_svn_reparent(svn_ra_svn_repos_root());
]]
[1] http://svn.haxx.se/dev/archive-2009-10/0315.shtml
--
Ivan Zhakov
VisualSVN Team
t; new session (since as Stefan suggested, this is an optimization so it
> doesn't need to break anything).
>
Yes, you can avoid issues like #3242 if you open RA session to URL and
then reparent to the root.
--
Ivan Zhakov
VisualSVN Team
hentication on that connection, you likely won't need to re-auth if
> you reuse the TCP connection via Serf or Neon. (All hail stupid
> authentication systems like Kerberos which are connection-oriented.
> Not.) -- justin
>
You can do that only if you assume that connections from dif
translated
pristine size to quickly answer that file modified without comparison
file contents at all.
--
Ivan Zhakov
VisualSVN Team
On Wed, Feb 17, 2010 at 10:12 PM, Greg Stein wrote:
> On Wed, Feb 17, 2010 at 14:03, Ivan Zhakov wrote:
>> On Wed, Feb 17, 2010 at 9:57 PM, Greg Stein wrote:
>>> Also: PRISTINE.size is the size on-disk, which may be different from
>>> the uncompressed size.
>>
html+xml". They
interpret them as _invalid_ HTML. [1,2]
Recommendation is to use HTML 4.01, which I suggest. In our experience
this simple change resolves a lot IE rendering problems.
[1] http://webkit.org/blog/68/understanding-html-xml-and-xhtml/
[2] http://hixie.ch/advocacy/xhtml
--
Ivan Zhakov
VisualSVN Team
functions to not forcibly operate against
> the repository root URL, but to operate instead against the RA
> session's current URL (which is generally going to be something that
> is readable by the accessing user).
>
Wow, Michael you are reading my mind! I've started backporting
issue-3242 fix just yesterday evening :) Thanks!
--
Ivan Zhakov
VisualSVN Team
On Fri, Feb 26, 2010 at 8:11 PM, C. Michael Pilato wrote:
> Ivan Zhakov wrote:
>> On Thu, Feb 25, 2010 at 5:23 AM, wrote:
>>> Author: cmpilato
>>> Date: Thu Feb 25 02:23:24 2010
>>> New Revision: 916089
>>>
>>> URL: http://svn.apache.org
oveFileEx with
MOVEFILE_REPLACE_EXISTING flag, so it should replace existing file.
http://msdn.microsoft.com/en-us/library/aa365240%28VS.85%29.aspx
--
Ivan Zhakov
VisualSVN Team
ileEx with
>> MOVEFILE_REPLACE_EXISTING flag, so it should replace existing file.
>
> The text above talks about the case that another application has the file
> open.
>
Ok, didn't get it.
--
Ivan Zhakov
VisualSVN Team
ny improvement is lost in
>> the noise when I run the command line client on Linux. It doesn't
>> change the number of system calls reported by 'strace -c'.
>
> Do our servers (mod_dav_svn and svnserve) parse authz files using this
> codepath?
Yes, they use the same code.
> How often? On a per-connection basis?
mod_dav_svn caches authz file per connection.
--
Ivan Zhakov
VisualSVN Team
n. Client
perform a lot of requests to decide what revisions and files to merge.
Just an idea: move this logic to server side and use slightly extended
reporter/editor to apply changes on client.
--
Ivan Zhakov
VisualSVN Team
On Tue, Mar 23, 2010 at 00:53, Mark Phippard wrote:
> On Mon, Mar 22, 2010 at 5:14 PM, Ivan Zhakov wrote:
>> On Mon, Mar 22, 2010 at 20:37, wrote:
>>> +Some other random stuff Hyrum would like to talk about:
>>> + * Why is merge slow (compared to $OTHER_SYSTEM)?
dev to be logged. All that needs to be done is that we get a
> bunch of +1s here and then tell the guy on #irclogger to please log
> #svn-dev, including a reference to this mail thread. The channel will
> then be logged exactly like #svn is logged right now.
>
+1.
--
Ivan Zhakov
VisualSVN Team
'))
>> + + glob.glob(os.path.join(serf_path, 'auth',
>> '*.c'))
>
> *.h maybe? :-)
>
You're right as usual! Fixed in r953456.
--
Ivan Zhakov
VisualSVN Team
te request body to file, to able resend it if needed.
>>
>> * subversion/libsvn_ra_serf/update.c
>> (): Include apr_version.h.
>
> You need private/svn_dep_compat.h since APR_VERSION_AT_LEAST was only
> introduced in 1.3.
>
I see already you fixed it r955164. Thanks a lot!
--
Ivan Zhakov
VisualSVN Team
g for this new function.
>
Hi Julian,
This is standard serf callback and we never documented them. Anyway
I've added simple doc string for create_update_report_body in r955317.
--
Ivan Zhakov
VisualSVN Team
le working copy, from its root down?
> Should they simply complain that you've not targeted the root of the working
> copy, and prescribe doing exactly that?
My opinion that best way is forbidding relocating subdirectory working
copy, since it's proven to broke parent working copy.
As far I remember we have the same behavior for svn upgrade now.
--
Ivan Zhakov
VisualSVN Team
webdav-acl-notes
* webdav-general-summary
* webdav-protocol
* webdav-proxy
* webdav-usage.html
--
Ivan Zhakov
VisualSVN Team
gt;> What do you think Subversion should do when it encounters a temporary
>> redirect, then?
>>
> Follow the redirect but not record the new URL in the working copy metadata.
+1.
--
Ivan Zhakov
VisualSVN Team
; + if (b->propchanges->nelts > 0)
>> + {
>> + remove_non_prop_changes(pristine_props, b->propchanges);
>> + }
>
> Shouldn't the remove_non_prop_changes() be moved into the else block? That
> would be the same as before, and there is no need to try and remove stuff
> from an empty hash.
>
remove_non_prop_changes() removes the changes from propchanges list,
not from pristine cache.
--
Ivan Zhakov
servers but our own), so we just see where it names
> the property that was missing.
>
> Basically: a backwards-compat bug where we failed to work against old
> servers. Fixed. (and no, I won't point fingers at Ivan... oh. oops.
> :-P )
>
Hi Greg,
Thanks for fixing this stupid bug that I introduced.
--
Ivan Zhakov
with a caveat of a few issues here
> and there).
>
> Can anybody think of additional high-level things that would keep us
> from shipping trunk *today*?
>
I think that we should reduce number of PROPFINDs in checkout/update
over ra_serf somehow before releasing 1.8.
--
Ivan Zhakov
loaded on his behalf!
To prevent such issues we should valdiate fs-type to be only file name
with only alphanumeric characters. No dots, spaces or slashes. We also
should only load DSO module from directory where Subversion installed
for better protection.
--
Ivan Zhakov
On Tue, Jul 17, 2012 at 5:21 PM, Philip Martin
wrote:
> Ivan Zhakov writes:
>
>> On Tue, Jul 17, 2012 at 2:14 PM, Philip Martin
>> wrote:
>>>
>>> [email protected] writes:
>>>
>>> > Author: philip
>>> > Date: Tue J
2009-10/0334.shtml
--
Ivan Zhakov
did not have this
> problem: console output would just fly by very quickly. At least
> that's how I remember it :-), I haven't run CVS for some time now
> (it's also possible that I am misremembering, or that it is because I
> was running the client on WinXP back then).
>
One difference that Subversion uses direct unbuffered WriteFile to
write to stdout instead
of buffered printf()/puts(). You may try to tweak
svn_stream_for_stdout() to return buffered stream and check
performance difference.
Other possible reason that Subversion converts all texts from utf8 to
console encoding before writing text to stdout.
--
Ivan Zhakov
slow on modern systems (compared to other bits).
>
It seems this performance issue is not related to locking in CRT: I've
replaced svn_cmdline_puts with direct WriteFile API to stdout and got
the same results. It seems the problem in other side that reads and
display svn.exe output.
--
Ivan Zhakov
lient caching the properties
> for an entire tree in memory, just waiting for a place to put them.
It should not happen in current ra_serf editor implementation: parsing
REPORT will be paused when active_fetchs + active_propfinds greater
than 1000. See REQUEST_COUNT_TO_PAUSE/REQUEST_COUNT_TO_RESUME
constants in libsvn_ra_serf/update.c
--
Ivan Zhakov
On Thu, Aug 30, 2012 at 10:43 PM, C. Michael Pilato wrote:
> On 08/30/2012 10:45 AM, Ivan Zhakov wrote:
>> On Thu, Aug 30, 2012 at 4:56 PM, C. Michael Pilato
>> wrote:
>>> Theoretically, though, it seems reasonable that my approach would have the
>>> distinct
"peg" not merely
> the current HEAD, but any arbitrary revision as the youngest revision the RA
> session will be allowed to recognize. But I can't quickly come up with a
> good reason for doing that.
>
For me it looks like abstraction leakage. Better solution would be
create new object like svn_client_ra_session_t() and store
cached/negotiate latest revision here, while extend
svn_ra_session_open to return youngest revision if requested.
--
Ivan Zhakov
On Fri, Aug 31, 2012 at 5:58 PM, Mark Phippard wrote:
> On Fri, Aug 31, 2012 at 8:03 AM, Ivan Zhakov wrote:
>>
>> On Thu, Aug 30, 2012 at 10:43 PM, C. Michael Pilato
>> wrote:
>> > On 08/30/2012 10:45 AM, Ivan Zhakov wrote:
>> >> On Thu, Aug 30, 2012
SVN. Instead of reading data a byte
> at a time, read 128 byte chunks and scan those.
>
Hi Stefan,
Why do not use APR internal buffering? I meant APR_BUFFERED flag.
--
Ivan Zhakov
time; /**< Compilation time */
> + const char *build_host; /**< Build canonical host name */
Hi, Branko.
What's is the purpose of build_host?
I think it's not good idea to include build host information in
svn.exe. It could leak private or corporate network details to public.
--
Ivan Zhakov
On Tue, Sep 11, 2012 at 12:05 AM, Branko Čibej wrote:
> On 10.09.2012 21:52, Ivan Zhakov wrote:
>> On Mon, Sep 10, 2012 at 9:19 PM, wrote:
>>> Author: brane
>>> Date: Mon Sep 10 17:19:31 2012
>>> New Revision: 1382990
>>>
>>> URL: htt
vide hints
> (read_window): seek before read as the file object may be shared
Hi Stefan,
Did you consider using apr_file_dup2() to save disk access, but do not
worry about seek to right position when reading data?
> (svn_fs_fs__get_file_delta_stream,
> verify_walker): adapt callers not using hints
>
--
Ivan Zhakov
dding pool parameter to authz_svn__subreq_bypass_func_t is not enough
to fix all memory issues. Request's actively pool used in other places
like get_username_to_authorize() or get_access_conf()
--
Ivan Zhakov
h temporary pool
3. Use scratch_pool in subreq_bypass2 instead of r->pool in all places.
4. Bump AUTHZ_SVN__SUBREQ_BYPASS_PROV_VER and export subreq_bypass2
5. Call AUTHZ_SVN__SUBREQ_BYPASS_PROV_VER2 if available in mod_dav_svn.
Backport changes 1-3 to svn 1.7.x, 4-5 leave on trunk.
--
Ivan Zhakov
On Wed, Sep 19, 2012 at 4:07 PM, Philip Martin
wrote:
> Ivan Zhakov writes:
>
>> On Wed, Sep 19, 2012 at 3:07 PM, Roderich Schupp
>> wrote:
>>> On Wed, Sep 19, 2012 at 10:24 AM, Roderich Schupp
>>> wrote:
>>>> Change the definition of aut
On Wed, Sep 19, 2012 at 3:29 PM, Ivan Zhakov wrote:
> On Wed, Sep 19, 2012 at 3:22 PM, Philip Martin
> wrote:
>> Roderich Schupp writes:
>>
>>>> Since this definition is in include/mod_authz_svn.h (i.e. it's a public
>>>> interfac
On Wed, Sep 19, 2012 at 6:06 PM, Philip Martin
wrote:
> Ivan Zhakov writes:
>
>> -get_access_conf(request_rec *r, authz_svn_config_rec *conf)
>> +get_access_conf(request_rec *r, authz_svn_config_rec *conf, apr_pool_t
>> *scratch_pool)
>
> That lin
On Wed, Sep 19, 2012 at 6:24 PM, Ivan Zhakov wrote:
> On Wed, Sep 19, 2012 at 6:06 PM, Philip Martin
> wrote:
>> Ivan Zhakov writes:
>>
>>> -get_access_conf(request_rec *r, authz_svn_config_rec *conf)
>>> +get_access_conf(request_rec *r, authz_svn_config_rec
On Thu, Sep 20, 2012 at 3:27 PM, Roderich Schupp
wrote:
> On Thu, Sep 20, 2012 at 12:19 PM, Ivan Zhakov wrote:
>> Commited in r1387943.
>
> Thanks! And with this patch, memory consumption for my test case
> for "SVNPathAuthz short_circuit" is now even lower than fo
.
> It it supposed to help users with gazillions of working copies that
> they have lying around and want to use with a newly installed svn
> client, and who don't have enough time to type 'svn upgrade' when
> they want to start using one of their old working copies. [*]
> And it annoys everyone else, as far as I can tell.
>
> [*] That's the only reason for its existence that I've heard, to date.
--
Ivan Zhakov
1 - 100 of 1096 matches
Mail list logo
