oblem of broken
operations when the network layer times out on a long callback invocation.
Bert
Sent from Windows Mail
From: Julian Foad
Sent: Tuesday, March 26, 2013 9:18 PM
To: Stefan Sperling
Cc: Subversion Development; Bert Huijben; Stefan Küng
Stefan Sperling wrote:
>
The openbsd bot explicitly uses the latest https version with its known issues.
I assumed this error was caused by that problem?
Bert
From: Branko Čibej
Sent: Monday, October 21, 2013 8:47 AM
To: 'Subversion Development'
This has been failing on the OpenBSD buil
warning’. There are better ways to fix that problem… and not just for your
favorite RA layer, ignoring others.
These warnings on the fs layer were designed to be logged server side (e.g. in
the apache error log); not to be transferred to the client.
Bert
Sent from Mail for Windows 10
From
tocol.
Bert
Sent from my Windows 10 phone
From: Paul Hammant
Sent: dinsdag 11 oktober 2016 14:09
To: Subversion Development
Subject: Re: New SHA1 property for nodes returned 'svn ls --xml' invocations.
Considering ..
svn info --xml
https://svn.apache.org/repos/asf/subversion/t
I applied your patch on trunk and nominated it for the next 1.9 release. I
created a similar patch for 1.8 as the code has the same bug there. I think the
problem became visible after checking the accepted failures was improved in our
ra-serf library.
Bert
Sent from my Windows 10 phone
From
evision 20.
$
]]
Resuming will now require a –force, which also triggers different behavior in
other parts of checkout.
Which would require me to update dozens of buildscripts that worked just fine
since 1.0, if released.
Bert
Sent from Mail for Windows 10
From: Stefan Sperling
Sent: do
that were setup to be compatible with those old setups don’t see this problem.
Note that I don’t think the unix makefiles build everything either… so they may
have similar problems somewhere.
But +1 on rerolling anyway.
Bert
Sent from Mail for Windows 10
From: Stefan Sperling
Sent: dinsdag 21
That code is in the backing code for svn_ra_replay(), where it also applies to
authz, not on the client.
@Julian Foad Can we use svnsync in this scenario, or does that break in a
similar way?
Bert
Sent from Mail for Windows 10
From: Julian Foad
Sent: woensdag 22 februari 2017 15:11
To
collision found using an insane amount of
CPU/GPU power.
Of course we should fix things to not break, but that is a different story.
Bert
Sent from Mail for Windows 10
From: Stefan Sperling
Sent: vrijdag 24 februari 2017 17:10
To: Andreas Stieger
Cc: Subversion Development
Subject: Re: Files
collision files.
That there is a collision now doesn’t change that we always assumed there would
be collisions, and designed the current behavior with that in mind.
Bert
Sent from Mail for Windows 10
From: Stefan Sperling
Sent: vrijdag 24 februari 2017 21:47
To: Mark Phippard
Cc: Øyvind A
tself fixing is a single right click operation)
Bert
Sent from Mail for Windows 10
From: Stefan
Sent: maandag 10 augustus 2015 04:35
To: dev@subversion.apache.org
Subject: Re: [PATCH] correct ToolsVersion in VS project files
On 10/08/2015 04:01, Stefan wrote:
> Hi,
>
> attached are patch
ils. Currently Subversion
(when using svn+ssh:// on Windows) just goes on trying to commit on Windows and
only notices that svnserve returned an error when done.
Bert
Sent from Mail for Windows 10
From: Thomas Singer
Sent: donderdag 20 augustus 2015 08:23
To: dev@subversion.apache.org
Subj
I don't think we should fix this with a 'revision+1' to explicitly allow many
bad ranges, but I do think we should specifically fix the r0 case for empty
repositories.
Bert
Sent from Mail for Windows 10
From: Julian Foad
Sent: zaterdag 5 september 2015 12:02
To: dev;Bert
I think we currently require APR >= 1.3 on trunk and 1.9, so feel free to
remove outdated references to older versions.
Bert
From: Stefan
Sent: donderdag 17 september 2015 03:08
To: dev@subversion.apache.org
Subject: Dropping APR 0.9.x specifics from INSTALL doc
Hi,
before I'
I created accounts ‘sussman’ and ‘pburba’ on Jira and e-mailed Ben and Paul
about them.
Bert
From: Ivan Zhakov
Sent: vrijdag 18 september 2015 14:38
To: Greg Stein
Cc: Subversion Development
Subject: Re: Migrating Subversion issues to ...
On 16 September 2015 at 14:44, Ivan Zhakov wrote
And if somebody wants to see the whiteboard in digital format there is a PDF at
http://1drv.ms/1Mvg4nX
Bert
From: Julian Foad
Sent: vrijdag 18 september 2015 14:28
To: dev
Subject: Whiteboard -- topics list on the white board
It's the last day of the hackathon. The following idea
... That and that I couldn’t think of a simple patch to easily fix
this. Things might be easier today after recent refactorings)
Bert
From: Simon Wilson
Sent: vrijdag 18 september 2015 15:05
To: dev@subversion.apache.org
Subject: Bug in lib_ra_serf ssl_server_cert()
We have found a bug in the
week and do
more handwork.
The users with > 100 comments should already be handled now.
Bert
From: Johan Corveleyn
Sent: vrijdag 18 september 2015 16:32
To: Branko Čibej
Cc: Ivan Zhakov;Subversion Development
Subject: Re: Migrating Subversion issues to ...
On Fri, Sep 18, 2015 at 4:25
diate states,
requires adding even more tests.
Bert
From: Daniel Shahaf
Sent: woensdag 30 september 2015 00:30
To: dev@subversion.apache.org
Subject: Re: [PATCH] mode-changing patches ∩binary patches
Could someone commit this for me, please? It passes tests (as of r1705925).
(I've foll
mentioned as
property change.
Requests to fix this, should now just be forwarded upstream, etc.
I'm not sure if that promise still holds...
Bert
From: Philip Martin
Sent: vrijdag 9 oktober 2015 23:47
To: Bert Huijben
Cc: Mark Phippard;Subversion Development
Subject: Re: Running SVN 1.9
the information
to really obtain where the change originated*. The way this code (build as a
summer of code project if I remember correct) assumed that every revision was
reported as interesting is just one of the many limitations.
Keeping this broken code working is not a goal by itself, that should
nges make a lot less sense in those cases... while
we already have the code to fix those cases.
We should revert the behavior where it makes sense. Reverting it everywhere
'just because' doesn't make sense.
Bert
Sent from Mail for Windows 10
From: Johan Corveleyn
Sent:
I think we should keep the functions in the private namespace until we really
expose this in our api/tools. Otherwise: +1.
Bert
From: Julian Foad
Sent: dinsdag 10 november 2015 16:28
To: dev
Subject: Merge 'svnmover' demo tool to trunk
The work on the 'move-tracking-2&#x
flags might be the hardest part
here, as svnmucc doesn’t have such options on subcommands yet as far as I know.
Bert
Van: 钱海远(Nathan)
Verzonden: dinsdag 28 april 2020 14:45
Aan: dev@subversion.apache.org
Onderwerp: Svnmucc sensitive to order of mkdir
Dear all,
I
’m not sure at which url.
Bert
Van: Graham Bloice
Verzonden: donderdag 23 april 2020 18:17
Aan: dev@subversion.apache.org
Onderwerp: Re: Building SVN (dependencies) on Windows
Apologies for butting in, but as a long time maintainer of the (mainly Windows)
build proce
The Windows buildbots currently produce actual errors, most likely caused by
recent changes to the python test framework. (It is expecting to see \r, but
now sees \n in several text conflicts)
Bert
-Oorspronkelijk bericht-
Van: Stefan Sperling
Verzonden: donderdag 2 april 2020 16:54
This buildbot runs python 2.X (and has for years) and now the tests are
broken. I don't remember seeing changes to the actual Conflict handing code
in Subversion itself, so...
I still think this is related to the changes to the test framework causing
regressions on 2.7.
This buildbot does the test
It looks like you get failures on this when you compile recent SQLite versions
with OMIT_DEPRECATED.
(I just bumped several dependencies on the Windows buildbot in an attempt to
fix the python breakage)
This should probably be backported...
Bert
-Oorspronkelijk bericht-
Van
dependencies, and ARM64 is new as well so most dependencies need
patches (except for those building via CMake)
Bert
Van: Nathan Hartman
Verzonden: woensdag 26 mei 2021 15:00
Aan: dev@subversion.apache.org
Onderwerp: Re: svn commit: r1890223 -
/subversion/trunk/subversion
es that the file is available in the
pristine store as uncompressed file or that we store a copy of the file in a
tempfolder (see the documentation update from Julian in r983593).
The deprecation of this api is most likely part of removing this dependency.
Bert
in(eb->root_url,
> > + path,
> > + pool),
> > + pool);
svn_uri_join expects it's arguments to be canonical, so path should be fixed
before passing. And as path may contain existing sequences like %20, that
need further escaping instead of just passing the real fix would be to call
svn_path_uri_encode().
Bert
path,
> > + pool),
> > + pool);
> >
>
> See svn_path_url_add_component2().
This is a better approach then the one I suggested in my other mail as it
handles all in one call.
Bert
your Apache
> > > configuration per the instructions in the
> > > subversion/tests/cmdline/README file.
>
> In r984365 I've updated davautocheck.sh to load mod_alias which is
> required for these RedirectMatch directives.
And r984348 applied the redirects and module load in win-tests.py to fix the
tests on Windows.
Bert
ng LDAP other than when you sign into your windows/domain account.
LDAP is a directory protocol; it allows to verify and get information about
tokens.
SSPI (as implemented by mod_auth_sspi, neon and serf) handles the logon without
asking for credentials.
If you want to check group membership and other details you can combine these
two systems. (That is how I configured it at TCG).
Bert
me additional tracing to detect where in
ra_serf we leaked these cancel errors)
Bert
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
Bert
h pointer is
available. (We just create the new hash and that value happens to be valid as
the old hash)
We have had a similar construct in svn_io_* for years, where we passed a
svn_node_kind* as value from a hash, then some struct and then later a larger
struct.. but you always get the same pointers, because the layout matches.
Bert
o -r).
We can't break this common use case without breaking almost every user of
the local diff feature.
svn cat uses a different definition.
Bert
-ng/count-progress.py, which are nice examples of files with
multiple revisions after the copy that you will still get as copied from the
old location... Exactly those that missed their actual properties before
r985254)
Bert
> -Original Message-
> From: Vincent Lefevre [mailto:vincent-...@vinc17.net]
> Sent: zondag 15 augustus 2010 16:50
> To: dev@subversion.apache.org
> Subject: Re: svn diff - Is this behaviour expected?
>
> On 2010-08-15 05:34:08 -0700, Bert Huijben wrote:
> >
property differences are not interesting, but for copied node
the fact that a property changed (after the copy) is.
If not for 'svn', it certainly is for library users.
Bert
ASE): Mark for split up.
>
> Modified:
> subversion/trunk/subversion/libsvn_wc/wc-queries.sql
> subversion/trunk/subversion/libsvn_wc/wc_db.c
Your log message doesn't describe any changes in wc_db.c
Bert
the dump in svnadmin has to create diffs from
fulltexts itself, while svnrdump has diffs and must apply these itself to
get full texts. The checksums have a similar mangling. The FS has access to
some of the checksums and recalculates others for you. (See the performance
drop in 1.6 of svnadmin dump)
There is a similar case at the import side. Applying commits can't check all
the checksums, but the really important ones are already handled. Svnrdump
dump and svnrdump load are a nice match.
Bert
scussions about this. But that could easily be because of the
> not-very-good search feature of the mailing list archives.
>
> Or is this a bug?
This looks like a bug to me. It should show all nodes with -v.
Bert
ifferent way.
The difference is between:
What would you get with
* svn up --depth infinity
and
* svn up --set-depth infinity
svn status --depth infinity used to show the last variant, but shows the
first variant now.
I think we should add an option to choose between those two variants. (By
enabling --set-depth on 'svn status' and a similar change to libsvn_client)
Bert
> -Original Message-
> From: Stefan Küng [mailto:tortoise...@gmail.com]
> Sent: dinsdag 17 augustus 2010 12:45
> To: Bert Huijben
> Cc: 'Subversion Development'
> Subject: Re: svn_client_status5() and depth
>
> On 17.08.2010 21:37, Bert Huijben wrote:
pose(compose_path, pool);
> + copyfrom_path = svn_relpath_join(svn_relpath_basename(path, pool),
> + pb->copyfrom_path, pool);
>copyfrom_rev = pb->copyfrom_rev;
> }
copyfrom_path is probably relative from the repository or session root (not
sure which in your case). You are now appending the basename of
a/paths/basename and then joining the copyfrom_path after it. Shouldn't that be
the other way around?
Bert
> -Original Message-
> From: Stefan Küng [mailto:tortoise...@gmail.com]
> Sent: woensdag 18 augustus 2010 11:39
> To: Hyrum K. Wright
> Cc: C. Michael Pilato; Bert Huijben; Subversion Development
> Subject: Re: svn_client_status5() and depth
>
> On 17.08.201
_cache, file_external
> +
> +When op_depth == 0 the node-rev columns represent the checked-out
> +repository node, otherwise they represent the copyfrom node.
> +
> +Presence has the same six values as BASE_NODE/WORKING_NODE:
> normal,
> +incomplete, absent, excluded, not-present, base-deleted. There are
> +some presence/op_depth constraints, e.g. base-deleted is not valid for
> +op_depth 0 and absent is not valid for op_depth != 0.
> +
Bert
> -Original Message-
> From: Stefan Küng [mailto:tortoise...@gmail.com]
> Sent: woensdag 18 augustus 2010 13:37
> To: Bert Huijben
> Cc: 'Hyrum K. Wright'; 'C. Michael Pilato'; 'Subversion Development'
> Subject: Re: svn_client_status5() an
ixed that, but I'm not too sure if
that's ok.
Should be ok.
> After these changes, I can get svn.exe to build. For svnadmin and
> svnserve, I still get linker errors like:
>
> main.obj : error LNK2019: unresolved external symbol
> _svn_fs_fs__set_cache_config referenced in function _main
> main.obj : error LNK2019: unresolved external symbol
> _svn_fs_fs__get_cache_config referenced in function _main
> ..\..\..\Debug\subversion\svnadmin\svnadmin.exe : fatal error LNK1120:
> 2 unresolved externals
I think this requires adding an additional dependency to svnadmin in
build.conf.
(My guess would be libsvn_fs_fs; but I can't look at the code right now)
@Stefan, can you apply these changes?
Thanks,
Bert
> -Original Message-
> From: Philip Martin [mailto:philip.mar...@wandisco.com]
> Sent: donderdag 19 augustus 2010 3:39
> To: Greg Stein
> Cc: Bert Huijben; dev@subversion.apache.org; phi...@apache.org
> Subject: Re: svn commit: r986865 - /subversion/trunk/notes/
tion and you get an easy to
understand result. (Make sure you have a recent version; the older versions
only allowed attaching to an already running process)
Bert
k"));
> +
The capability of committing revprops is unrelated to whether we can change
revision properties.
In 1.5 we added the option to add a list of revision properties to set directly
at commit, and this is what you check here. If you are committing to older
repositories you can just set the revision properties after the commit (like
you do for svn:author and svn:date).
And if you get a failure on doing that, then you should show this error.
Bert
ovide
too much overhead.
We are not going to provide long-time support for these dev only
intermediate formats anyway.
Bert
d"
> state is only meaningful with respect to the expected location as
> determined from the node's parent directory in the same WC. (The
> "svnversion" command has an option to specify an expected location
> so that it can report the status of the whole WC as "switched"
> relative to that specified location.)
>
> [4] A possible exception is that "update" should obey the "pinned
> revision number" if present. But that need not necessarily be an
> exception: we might choose to define that functionality on all WC
> nodes but only use it on 'external' nodes.
See my answer on [2]. I like this generalized pinned idea as option to replace
our checks on file externals.
(First thing that comes to mind is how it would interact with 'svn switch on
some parent')
Bert
> -Original Message-
> From: Stefan Sperling [mailto:s...@elego.de]
> Sent: vrijdag 20 augustus 2010 3:29
> To: dev@subversion.apache.org
> Subject: Reduce the 1.7 release feature set a bit?
>
> I think that the recent NODE_DATA developments in wc-ng show that wc-ng
> is still in flux, o
> -Original Message-
> From: Stefan Sperling [mailto:s...@elego.de]
> Sent: zaterdag 21 augustus 2010 3:32
> To: Bert Huijben
> Cc: dev@subversion.apache.org
> Subject: Re: Reduce the 1.7 release feature set a bit?
>
> On Fri, Aug 20, 2010 at 11:55:10AM -
> -Original Message-
> From: Bert Huijben [mailto:b...@qqmail.nl] On Behalf Of Bert Huijben
> Sent: zaterdag 21 augustus 2010 9:18
> To: 'Stefan Sperling'
> Cc: dev@subversion.apache.org
> Subject: RE: Reduce the 1.7 release feature set a bit?
> &
and add the mapping in the
pristines table) in the upgrade step. I expected that this was already handled?
Bert
> -Original Message-
> From: Philip Martin [mailto:philip.mar...@wandisco.com]
> Sent: dinsdag 24 augustus 2010 10:40
> To: Bert Huijben
> Cc: dev@subversion.apache.org
> Subject: Re: svn commit: r988074 - in
> /subversion/trunk/subversion/tests/cmdline: svntest/wc.
> -Original Message-
> From: Julian Foad [mailto:julian.f...@wandisco.com]
> Sent: dinsdag 24 augustus 2010 13:04
> To: Bert Huijben
> Cc: 'Philip Martin'; dev@subversion.apache.org
> Subject: RE: svn commit: r988074 - in
> /subversion/trunk/subversi
orded as entry->deleted in
the parent dir). But if you set the directory to incomplete the 'not-present'
status gets lost.
And this can't be fixed from the directory itself, as it doesn't know that it
originally existed as not-present, because we only started propagating this
information into the subdir in the WC-NG work.
Bert
> -Original Message-
> From: Philip Martin [mailto:philip.mar...@wandisco.com]
> Sent: dinsdag 24 augustus 2010 16:58
> To: Bert Huijben
> Cc: 'Julian Foad'; 'Bert Huijben'; dev@subversion.apache.org
> Subject: Re: svn commit: r988074 - in
> /
wc__db_status_normal;
> > > + conflicted = FALSE;
> > >
> > >versioned_locally_and_present = FALSE;
> > > }
> >
> > Could someone confirm that FALSE is correct here?
>
> Looks fine to me.
Answered on IRC: No, this is not the correct value.
A node can still be in conflict while _db_read_info() returns path not
found. Currently delete-delete tree conflicts can be the exception.
So in this case it's safer to set the conflicted variable to TRUE and
thereby perform the more expensive checks, to verify if the node is really
in conflict.
Bert
in only a
few weeks.
A slightly easier variant (that needs less hacks) might be to create the
wc.db in a temp directory, like you recently implemented for 'svn copy
wc-dir wc-dir2'.
Bert
> -Original Message-
> From: Philip Martin [mailto:philip.mar...@wandisco.com]
> Sent: donderdag 26 augustus 2010 21:33
> To: Greg Stein
> Cc: Bert Huijben; dev@subversion.apache.org
> Subject: Re: Two svn_wc__db_t for single-db upgrade
>
> Philip Martin writes
> -Original Message-
> From: Philip Martin [mailto:philip.mar...@wandisco.com]
> Sent: vrijdag 27 augustus 2010 11:50
> To: Bert Huijben
> Cc: 'Greg Stein'; dev@subversion.apache.org
> Subject: Re: Two svn_wc__db_t for single-db upgrade
>
> "Bert
> -Original Message-
> From: Philip Martin [mailto:philip.mar...@wandisco.com]
> Sent: vrijdag 27 augustus 2010 14:57
> To: Bert Huijben
> Cc: 'Bert Huijben'; 'Greg Stein'; dev@subversion.apache.org
> Subject: Re: Two svn_wc__db_t for single-
most every case a user error.
So this triggers a maintainer/debug only assertion]
This allows easier debugging of these errors.
SVN_ERR_MALFUNCTION() and SVN_ERR_ASSERT() are for tests that also need
evaluation in release mode. The fatal error is avoidable by installing a
error callback.
The SVN_ERR_MALFUNCTION_NO_RETURN and SVN_ERR_ASSERT_NO_RETURN() function
should be avoided (real fix: Make the function return svn_error_t*) as these
will trigger an error and then an abort at runtime. So it will CRASH third
party application when this error occurs there.
Bert
dn't
feel like reinventing a complete lock store system like the old access
batons just to fix these issues for multi-db. (Single-db has a real
recursive lock, so it only has to delete the lock from the root)
Bert
> ClientContext.h
>
> "Bert Huijben" writes:
>
> > (I see a completely different test failure on the Windows bot)
>
> On Linux I get the same as Hyrum.
>
> > Can you run this same test with single-db?
>
> In single-db on Linux I ge
> ClientContext.h
>
> "Bert Huijben" writes:
>
See original thread
>
> > The current recursive lock code for multi-db doesn't have stable
> behavior
> > over added and removed directories, while the single-db code has. I
> didn't
> > feel l
og file next to wc.db. But then it finds that
this directory no longer exists, so it can't create the file.
On Windows the directory is not deleted because we still have that file
open. (File.delete() returns false in SVNTest.removeDirOrFile(), but we
don't check the result there).
But as this doesn't give the expected result (X is unmodified there), the
test fails a few lines later.
This specific issue is fixed once we move to SingleDB, but the problem that
we can delete databases that we have open will stay (on non Windows
systems).
Bert
ter the stat tests are fixed, so I intend to enable
> it then. Any objections?
+1
All tests Pass or XPass on Windows with single-db, except for the upgrade
tests. (Philip is working on this one)
Bert
e that pattern for these
functions.
In all the cases I knew, this revision moved the flushing outside the #ifdef
SINGLE_DB blocks. (There might be more wc_db operations that need additional
flushing, but a simple scan of all the db operation didn't show any of
these. But an extra pair of eyes on this would certainly help)
Bert
blic interface anyway. (Resulting in a linker error)
For a few specific cases (like mac only function) we applied some exceptions
to the extractor, but in this case I think we should just start removing
multi-db support.
Bert
> -Original Message-
> From: pbu...@apache.org [mailto:pbu...@apache.org]
> Sent: donderdag 2 september 2010 20:51
> To: comm...@subversion.apache.org
> Subject: svn commit: r992050 -
> /subversion/trunk/subversion/tests/cmdline/merge_tests.py
>
> Author: pburba
> Date: Thu Sep 2 18:51:
commit process: They would
only have to do restructuring changes(add/delete) when there is a new
op_depth.
So the commit harvester would be as simple as walking the operations (for
restructuring operations) and checking present nodes for text and prop mods.
If multiple levels of adds would be a single operation, we would have to
perform more checks to see if it is an operation that must be committed (or
requires restructuring on revert), while we could just compare the op_depth
with the number of '/'-s in the local_relpath if every local add is its own
op_depth.
Bert
> Cheers,
> -g
> -Original Message-
> From: Greg Stein [mailto:gst...@gmail.com]
> Sent: vrijdag 3 september 2010 17:23
> To: Bert Huijben
> Cc: Julian Foad; Erik Huelsmann; dev@subversion.apache.org
> Subject: Re: [PROPOSAL] WC-NG: merge NODE_DATA, WORKING_NODE and
> BASE_NOD
ransaction.
(On Windows this specific stat to check for other processes operating on the
same db is the performance killer for svn status: Just this stat takes more
than 50% of the total processing).
Bert
> -Original Message-
> From: Greg Stein [mailto:gst...@gmail.com]
> Sent: zaterdag 4 september 2010 6:28
> To: Bert Huijben
> Cc: Julian Foad; Erik Huelsmann; dev@subversion.apache.org
> Subject: Re: [PROPOSAL] WC-NG: merge NODE_DATA, WORKING_NODE and
> BASE_NOD
d
also safely speed up 'svn status -u' which does multiple small local status
walks).
It would be interesting to see what combining the SAVEPOINT and
read_uncommitted pragma would do here. (And maybe we should also look at the
Write Ahead Logging here, but that only works for local disks)
Bert
d_token(stmt, 5, presence_map,
> + svn_wc__db_status_normal));
> + SVN_ERR(svn_sqlite__bind_token(stmt, 6, kind_map,
> + svn_wc__db_kind_subdir));
Why don't you use _bindf("isistt", ...) here?
That would include all the other fields. (Other option: separate binds of all
values)
Bert
> -Original Message-
> From: Matthew Bentham [mailto:mj...@artvps.com]
> Sent: maandag 6 september 2010 15:07
> To: Justin Erenkrantz
> Cc: Bert Huijben; Greg Stein; Johan Corveleyn; Subversion Development
> Subject: Re: Worried about single-db performance
>
tion) is
handled by your ssh client instead of by svn (or in this case svnsync).
You could try using svn+ssh://u...@server instead of svn+ssh://server, as
this usually makes your ssh client pass this 'user'.
Bert
> -Original Message-
> From: Matthew Bentham [mailto:mj...@artvps.com]
> Sent: dinsdag 7 september 2010 13:48
> To: Bert Huijben
> Cc: 'Justin Erenkrantz'; 'Greg Stein'; 'Johan Corveleyn'; 'Subversion
> Development'
> Subject:
Because a simple test with using this lock with status shows a 70% performance
increase on Windows.
You can't get to this speed with just removing 50% of the info queries and
Philips tests showed that this is probably the only way to get back at 1.6
status performance.
> -Original Message-
> From: style...@apache.org [mailto:style...@apache.org]
> Sent: donderdag 9 september 2010 13:47
> To: comm...@subversion.apache.org
> Subject: svn commit: r995388 -
> /subversion/trunk/subversion/tests/cmdline/update_tests.py
>
> Author: stylesen
> Date: Thu Sep 9
key_or_val(&keybuf, actual_length,
> - stream, atoi(buf + 2), proppool));
> + stream, (apr_size_t)len, proppool));
What would this do if you have 4GB + 1 byte?
I expected that we would use the new svn_cstring conversion variants to check
for that kind of errors (for overflows, etc.), but now we just ignore the error
at a different level.
Bert
e it makes sense to call it base (vs mine vs theirs) or maybe older,
but svn_client_*merge* uses the libsvn_wc merge code in a different way.
Bert
> -Original Message-
> From: Stefan Sperling [mailto:s...@elego.de]
> Sent: vrijdag 10 september 2010 15:26
> To: Bert Huijben
> Cc: dev@subversion.apache.org
> Subject: Re: svn commit: r995475 -
> /subversion/trunk/subversion/libsvn_repos/load.c
>
> On Thu,
at the original intent was..
so we ask the caller to fix it for us by returning an error.
(And this old function used to behave this way before we 'revved' it to
integrate an optional repair functionality)
Bert
t; access to the paths it touches.
Is it possible to port this back to 1.6.x?
Bert
NULL);
This doesn't look right (in the old and new code). If you use a client session
multiple time, you shouldn't recreate the auth baton on every invocation, but
only if there is a reason to recreate.
The auth baton also handles some caching for the auth providers. (At least it
stores the used passwords that weren't saved to disk, but I think it also
stores some context state for a few providers)
Bert
actly right!)
>
> These rules are not documented as far as I can see, so the question is
> whether we have to maintain this arbitrary behaviour.
>
> Perhaps we should simply notify for all nodes? Perhaps each
> individual client should be deciding which notifications to suppress?
+1 on moving the choice to the clients. (svn has different requirements then
clients like Subclipse).
Personally (and for AnkhSVN) I don't have an issue with notifying adds and
deletes for the root only. (The fact that it is an add or delete tells that
it's operation changes the state of all descendants)
Bert
hich (when not NULL) can contain
those error codes, or can be NULL/SVN_NO_ERROR. SVN_NO_ERROR and the specific
error codes should not be in the same list without some separation.
Bert
ge governing permissions and limitations
> + *under the License.
> + *
> ====
> + * @endcopyright
> + *
> + * @file Client.h
> + * @brief Interface of the class Client
> + */
> +
> +#ifndef CLIENT_H
> +#define CLIENT_H
If this is going to be some public API, this should be SVN__CLIENT_H
or alike and the name of the file should also be prefixed with 'svn_'
> +
> +#include "Version.h"
> +
> +#include "svn_client.h"
> +
> +namespace SVN
> +{
> +
> + class Client
Same problem. We define it in the SVN namespace here, but the name of the class
is far to generic to be sure not to clash with other classes.
SVNClient?
(SVNVersion?)
Bert
claims that it is 'switched', and inserts a repos-id
> and repos_relpath.
>
> I'm not yet clear on what this really means. Sure it's a 'hack' but I
> guess this is the way we do it for the time being, so, the statement
> that not-present implies repos-id/relpath/revnum are all null is false.
I don't think they are really NULL, but that doesn't tell that the values are
relevant.
Bert
ation on
the highest op_depth version of the node itself.
The update behavior op_depth should be explicit instead of just taking the
highest current op_depth. (Updating the highest existing op_depth doesn't
verify if the node is deleted/replaced (where you can still elide lower layers)
or maybe just a base_node (error in caller))
Bert
1 - 100 of 2022 matches
Mail list logo
