svnserve.exe crashes with "can't create thread"

[Thomas Börkel]

HI!

We have the problem, that the svnserve.exe process crashes when doing a
pretty large commit.

It seems to allocate 700 MB memory and then it crashes with "can't
create thread".

Using svnserve.exe 1.5.5. The machine has 6 GB memory.

Any hints would be greatly appreciated.

Thanks!

Thomas

format: Permission denied even when read permissions are set

[Jeremy Conlin]

I just created a new repository, and loaded it with some initial
files.  However when I go to checkout from my repository I get the
following:

$ svn checkout svn+ssh://jlcon...@myserver.com/path/to/repos/Repo CP
Password:
svn: Can't open file '/path/to/repos/Repo/format': Permission denied

I can log via ssh to check on the permissions of format and this is
what they are:

$ ls -lh
total 9.5K
-rw-rw-r--  1 jlconlin jlconlin 229 Mar 11 06:55 README.txt
drwxrwxr-x  2 jlconlin jlconlin   5 Mar 11 06:55 conf
drwxrwxr-x  2 jlconlin jlconlin   2 Mar 11 06:55 dav
drwxrwsr-x  5 jlconlin jlconlin  10 Mar 11 06:56 db
-rw-rw-r--  1 jlconlin jlconlin   2 Mar 11 06:55 format
drwxrwxr-x  2 jlconlin jlconlin  11 Mar 11 06:55 hooks
drwxrwxr-x  2 jlconlin jlconlin   4 Mar 11 06:55 locks


It seems clear to me that I have both read and write permissions for
the 'format' file.  When I log in, I can view the file and see that
the format says '5'.

Does anyone know what is going wrong here?

Thanks,
Jeremy

RE: format: Permission denied even when read permissions are set

[Giulio Troccoli]

> I just created a new repository, and loaded it with some
> initial files.  However when I go to checkout from my
> repository I get the
> following:
>
> $ svn checkout svn+ssh://jlcon...@myserver.com/path/to/repos/Repo CP
> Password:
> svn: Can't open file '/path/to/repos/Repo/format': Permission denied
>
> I can log via ssh to check on the permissions of format and
> this is what they are:
>
> $ ls -lh
> total 9.5K
> -rw-rw-r--  1 jlconlin jlconlin 229 Mar 11 06:55 README.txt
> drwxrwxr-x  2 jlconlin jlconlin   5 Mar 11 06:55 conf
> drwxrwxr-x  2 jlconlin jlconlin   2 Mar 11 06:55 dav
> drwxrwsr-x  5 jlconlin jlconlin  10 Mar 11 06:56 db
> -rw-rw-r--  1 jlconlin jlconlin   2 Mar 11 06:55 format
> drwxrwxr-x  2 jlconlin jlconlin  11 Mar 11 06:55 hooks
> drwxrwxr-x  2 jlconlin jlconlin   4 Mar 11 06:55 locks
>
>
> It seems clear to me that I have both read and write
> permissions for the 'format' file.  When I log in, I can view
> the file and see that the format says '5'.
>
> Does anyone know what is going wrong here?

Check the user who runs svnserve. Does IT have permission to read and write?

Giulio


Linedata Services (UK) Ltd
Registered Office: Bishopsgate Court, 4-12 Norton Folgate, London, E1 6DB
Registered in England and Wales No 3027851VAT Reg No 778499447

RE: format: Permission denied even when read permissions are set

[Giulio Troccoli]

> > Check the user who runs svnserve. Does IT have permission
> to read and write?
>
> How do I do that?  I've never had a need to do that before,
> so I've never looked up how to do it.

You're probably right actually. I don't have any experience with svnserve, but 
thinking about and remembering from other posts I'd been reading, when using 
svn+ssh a one-off svnserve process is started, which should run under the user 
that authenticates.

As I said thought I don't have experience with svn+ssh.

And please, reply to all so the discussione stays on the list.

Giulio


Linedata Services (UK) Ltd
Registered Office: Bishopsgate Court, 4-12 Norton Folgate, London, E1 6DB
Registered in England and Wales No 3027851VAT Reg No 778499447

Recursive ACLs fail

[Jeremiah Foster]

Hello,

I have a repo that looks like this:

/var/svn-repos/uml-model
/var/svn-repos/uml-model/sandbox/
/var/svn-repos/uml-model/sandbox/trunk/
/var/svn-repos/uml-model/sandbox/tags/
/var/svn-repos/uml-model/sandbox/branch/

And I want to have read-write permissions on the sandbox, but not on 
uml-model. So I have set up an ACLs fie which looks like this:

[uml-model:/]
* = r

[uml-model:/sandbox/]
* = rw


I am using LDAP for authentication which works fine. But what I cannot 
get to work is to permissions to be inheritable. No matter what my 
configuration, Subversion and Apache only allow me to read or write in /sandbox 
depending on the permissions of /.

Is there any way I can specify ACLS per directory? What am I doing 
wrong?

Jeremiah

Unexpected Externals behaviour in working copy

[David Aldrich]

Hi

I have a question about externals. I tried the following scenario in my working 
copy:

1) Add an external definition to directory A, specifying subdirectory B
2) Commit directory A
3) Update directory A

I then see A/B populated

Then:

4) Edit external definition to directory A, specifying subdirectory C
5) Commit directory A
6) Update directory A

I then see A/C populated, but directory A/B remains in the wc.

If I checkout a new working copy, it shows A/C but not A/B, as expected.

So, in short, changing an external definition on a directory seems not to 
remove the previous external content from the committer's working copy.

I hope this makes sense. Please will someone explain?

Best regards

David

Re: Recursive ACLs fail

[Andy Levy]

On Thu, Mar 11, 2010 at 11:06, Jeremiah Foster
 wrote:
> Hello,
>
>        I have a repo that looks like this:
>
>        /var/svn-repos/uml-model
>        /var/svn-repos/uml-model/sandbox/
>        /var/svn-repos/uml-model/sandbox/trunk/
>        /var/svn-repos/uml-model/sandbox/tags/
>        /var/svn-repos/uml-model/sandbox/branch/
>
>        And I want to have read-write permissions on the sandbox, but not on 
> uml-model. So I have set up an ACLs fie which looks like this:
>
>        [uml-model:/]
>        * = r
>
>        [uml-model:/sandbox/]
>        * = rw
>
>
>        I am using LDAP for authentication which works fine. But what I cannot 
> get to work is to permissions to be inheritable. No matter what my 
> configuration, Subversion and Apache only allow me to read or write in 
> /sandbox depending on the permissions of /.
>
>        Is there any way I can specify ACLS per directory? What am I doing 
> wrong?

Change:
   [uml-model:/sandbox/]
To:
   [uml-model:/sandbox]

I don't think the trailing slash is permitted. Every example in the
manual omits the trailing slash on paths.

Re: Recursive ACLs fail

[Jeremiah Foster]

On Mar 11, 2010, at 5:15 PM, Andy Levy wrote:

> On Thu, Mar 11, 2010 at 11:06, Jeremiah Foster
>  wrote:
>> Hello,
>> 
>>I have a repo that looks like this:
>> 
>>/var/svn-repos/uml-model
>>/var/svn-repos/uml-model/sandbox/
>>/var/svn-repos/uml-model/sandbox/trunk/
>>/var/svn-repos/uml-model/sandbox/tags/
>>/var/svn-repos/uml-model/sandbox/branch/
>> 
>>And I want to have read-write permissions on the sandbox, but not on 
>> uml-model. So I have set up an ACLs fie which looks like this:
>> 
>>[uml-model:/]
>>* = r
>> 
>>[uml-model:/sandbox/]
>>* = rw
>> 
>> 
>>I am using LDAP for authentication which works fine. But what I 
>> cannot get to work is to permissions to be inheritable. No matter what my 
>> configuration, Subversion and Apache only allow me to read or write in 
>> /sandbox depending on the permissions of /.
>> 
>>Is there any way I can specify ACLS per directory? What am I doing 
>> wrong?
> 
> Change:
>   [uml-model:/sandbox/]
> To:
>   [uml-model:/sandbox]
> 
> I don't think the trailing slash is permitted. Every example in the
> manual omits the trailing slash on paths.

Thank you Andy. Your suggestion fixed my problem. I have the book open to the 
page mentioning the ACLs format and I can see that, yes, there is no trailing 
slash in the example. Why I added it, I don't know.

Thanks again.

Jeremiah

Re: format: Permission denied even when read permissions are set

[David Weintraub]

The way svn+ssh works is very strange. svnserve is executed by each
user that uses the svn+ssh protocol. That means that you have to make
sure that all of these users are in the same group, and that group has
read/write permission on the entire Subversion archive. Plus, you have
to set umask for each user, so when they create a new file in the
Subversion repository, it is read/writeable by everyone in the group.

Take a look at the Tunneling Over SSH chapter in the Subversion on line book:
.

-- 
David Weintraub
qazw...@gmail.com

Re: Recursive ACLs fail

[Andrey Repin]

Greetings, Jeremiah Foster!

> Is there any way I can specify ACLS per directory? What am I doing 
> wrong?

Just mention everyone at every point.
Like,

[]
~someone = rw
someone = r


--
WBR,
 Andrey Repin (anrdae...@freemail.ru) 11.03.2010, <19:46>

Sorry for my terrible english...

Re: format: Permission denied even when read permissions are set

[Jeremy Conlin]

On Thu, Mar 11, 2010 at 9:27 AM, David Weintraub  wrote:
> The way svn+ssh works is very strange. svnserve is executed by each
> user that uses the svn+ssh protocol. That means that you have to make
> sure that all of these users are in the same group, and that group has
> read/write permission on the entire Subversion archive. Plus, you have
> to set umask for each user, so when they create a new file in the
> Subversion repository, it is read/writeable by everyone in the group.
>
> Take a look at the Tunneling Over SSH chapter in the Subversion on line book:
> .

Thanks for the reminder.  Currently it is just one user (me) that is
accessing this repository and I know I have access to the files and
folder in the directory as I showed in an earlier email and because I
can go inside and see everything.  What else could be wrong?

Thanks,
Jeremy

Re: format: Permission denied even when read permissions are set

[David Weintraub]

On Thu, Mar 11, 2010 at 12:07 PM, Jeremy Conlin  wrote:
> On Thu, Mar 11, 2010 at 9:27 AM, David Weintraub  wrote:
> Thanks for the reminder.  Currently it is just one user (me) that is
> accessing this repository and I know I have access to the files and
> folder in the directory as I showed in an earlier email and because I
> can go inside and see everything.  What else could be wrong?

Is this Subversion repository on your system or on another system? It
could very well be that when you login as SSH, you are not exactly the
user you think you are.

Can you SSH directly into that system, check out your user ID and
verify that its the ID you think it is?

(Of course, you SHOULDN'T be able to SSH into the Subversion
repository because that would allow users to munge the repository
directly which they shouldn't be doing. But if you can, you can at
least verify that you are who you think you are.)


-- 
David Weintraub
qazw...@gmail.com

Re: format: Permission denied even when read permissions are set

[Jeremy Conlin]

On Thu, Mar 11, 2010 at 10:22 AM, David Weintraub  wrote:
> On Thu, Mar 11, 2010 at 12:07 PM, Jeremy Conlin  wrote:
>> On Thu, Mar 11, 2010 at 9:27 AM, David Weintraub  wrote:
>> Thanks for the reminder.  Currently it is just one user (me) that is
>> accessing this repository and I know I have access to the files and
>> folder in the directory as I showed in an earlier email and because I
>> can go inside and see everything.  What else could be wrong?
>
> Is this Subversion repository on your system or on another system? It
> could very well be that when you login as SSH, you are not exactly the
> user you think you are.
>
> Can you SSH directly into that system, check out your user ID and
> verify that its the ID you think it is?

Yes I can do this and I have verified that I am a member of the group
that owns the files and folders of the repository.  The owning group
has read and write permissions.

>
> (Of course, you SHOULDN'T be able to SSH into the Subversion
> repository because that would allow users to munge the repository
> directly which they shouldn't be doing. But if you can, you can at
> least verify that you are who you think you are.)

How can you prevent this and still allow the users to read and write
to the repository?

Thanks,
Jeremy

Re: format: Permission denied even when read permissions are set

[Ryan Schmidt]

On Mar 11, 2010, at 11:27, Jeremy Conlin wrote:

> On Thu, Mar 11, 2010 at 10:22 AM, David Weintraub wrote:
>> 
> 
>> (Of course, you SHOULDN'T be able to SSH into the Subversion
>> repository because that would allow users to munge the repository
>> directly which they shouldn't be doing. But if you can, you can at
>> least verify that you are who you think you are.)
> 
> How can you prevent this and still allow the users to read and write
> to the repository?

It gets a bit involved but it should be described here:

http://svnbook.red-bean.com/en/1.5/svn.serverconfig.svnserve.html#svn.serverconfig.svnserve.sshtricks

If you're the only user here, presumably you won't mess up your own repository 
(at least not intentionally) so I wouldn't necessarily worry about this at this 
point.

Re: STDOUT and Hook Scripts

[Stein Somers]

The post-commit hook starts a single process, and eats its STDOUT.

But you can make that single process into whatever you want. If you hook 
script is for instance:


#!/bin/sh
perl watch-file.pl | perl notify.pl

then your hook consists of a shell process and two perl processors 
working together. SVN doesn't know anything about the STDOUT produced by 
watch-file.pl. It still eats the STDOUT that notify.pl produces - 
someone has to. The commit is finished when notify.pl is done (more or 
less).


--
Stein

Re: STDOUT and Hook Scripts

[Les Mikesell]

On 3/11/2010 12:56 PM, Stein Somers wrote:

The post-commit hook starts a single process, and eats its STDOUT.

But you can make that single process into whatever you want. If you hook
script is for instance:

#!/bin/sh
perl watch-file.pl | perl notify.pl

then your hook consists of a shell process and two perl processors
working together. SVN doesn't know anything about the STDOUT produced by
watch-file.pl. It still eats the STDOUT that notify.pl produces -
someone has to. The commit is finished when notify.pl is done (more or
less).


Likewise, perl is just as good as the shell at starting programs on 
pipes and manipulating file descriptors.  One program could read from or 
write to a pipe within your perl code.


--
  Les Mikesell
   lesmikes...@gmail.com

unable to reintegrate branch

[Manny DaSilva]

Server at v1.6.9
CollabNet command-line client at v1.6.9
TortoiseSVN client at v1.6.7
1. Trunk copied to branches/X
2. Shallow checkout of branches/X creates sparse working-copy:
X/A
X/B
3. update --set-depth infinity X/A downloads additional folders from X/A tree:
X/A/1
X/A/2
X/A/3
3. merge --depth infinity /trunk
The merge works correctly but causes svn:mergeinfo properties like
the following. Note the asterisked rev number in svn:mergeinfo:
/trunk/A:199-202*
/trunk/B:199-202*
4. Using E:\X41\trunk fully recursive working copy of trunk, run:
merge --reintegrate /branches/X
Get this error:
Command: Reintegrate merge /branches/X into E:\X41\trunk  
Error: Reintegrate can only be used if revisions 199 through 204 were 
previously   
Error: merged from /trunk to the reintegrate source, but this is   
Error: not the case:  
Error:   branches/X/A  
Error: Missing ranges: /trunk/A:199-202  
Error:   branches/X/B  
Error: Missing ranges: /trunk/B:199-202  

Any ideas as to what may be causing these errors?


  

Re: RedHat 5.4 / CentOS 5.4 upgrade to Subversion 1.6.9 - SQLite compiled for 3.6.13, but running with 3.3.6

[Marcus Schultheiss]

Hello,

does anybody do have any hints on this issue regarding CentOS 5.4.
How to upgrade to 1.6.9 but not break Distribution - SQLite 3.3.6 => 3.6.13.

A work-a-round (e.g. get LD_LIBRARY_PATH to work would help too)

any help is appreciated!

- Marcus

On Fri, Mar 5, 2010 at 11:12 AM, Marcus Schultheiss
 wrote:
> Hello,
>
> having problems to upgrade to Subversion 1.6.9 on CentOS 5.4 (x86_64)
> system.
> I am able to compile subversion (agains sqlite 3.6.13) and install.
> svn command looks fine but using svnadmin does not work
> # /usr/local/bin/svnadmin create /tmp/test1
> svnadmin: SQLite compiled for 3.6.13, but running with 3.3.6
>
> # ldd /usr/local/bin/svnadmin| grep sqlite
>     libsqlite3.so.0 => /usr/lib64/libsqlite3.so.0 (0x003f7a40)
>
> trying to fix:
> # export LD_LIBRARY_PATH=/usr/local/lib64
> # ls -al /usr/local/lib64/libsqlite3.so.0
> lrwxrwxrwx 1 root root 19 Mar  4 18:16 /usr/local/lib64/libsqlite3.so.0 ->
> libsqlite3.so.0.8.6
> # ldd /usr/local/bin/svnadmin| grep sqlite
>     libsqlite3.so.0 => /usr/lib64/libsqlite3.so.0 (0x003f7a40)
> # echo $LD_LIBRARY_PATH
> /usr/local/lib64
>
>
> for some unkown reason LD_LIBRARY_PATH work-a-round is not working
>
> I could not remove system provided sqlite 3.3.6 because there are
> dependencies to RPM and so on.
> I do not want to break other tools.
>
> - Marcus
>
>
>

Re: confirm unsubscribe from users@subversion.apache.org

[Kay Kay]

On 03/11/2010 11:19 PM, users-h...@subversion.apache.org wrote:

Hi! This is the ezmlm program. I'm managing the
users@subversion.apache.org mailing list.

To confirm that you would like

kaykay.uni...@gmail.com

removed from the users mailing list, please send a short reply
to this address:


users-uc.1268378387.pndepcbkaaialjfpnamn-kaykay.unique=gmail@subversion.apache.org

Usually, this happens when you just hit the "reply" button.
If this does not work, simply copy the address and paste it into
the "To:" field of a new message.

I haven't checked whether your address is currently on the mailing list.
To see what address you used to subscribe, look at the messages you are
receiving from the mailing list. Each message has your address hidden
inside its return path; for example, m...@xdd.ff.com receives messages
with return path:-mary=xdd.ff@subversion.apache.org.

Some mail programs are broken and cannot handle long addresses. If you
cannot reply to this request, instead send a message to
  and put the entire address listed above
into the "Subject:" line.


--- Administrative commands for the users list ---

I can handle administrative requests automatically. Please
do not send them to the list address! Instead, send
your message to the correct command address:

To subscribe to the list, send a message to:


To remove your address from the list, send a message to:


Send mail to the following for info and FAQ for this list:



Similar addresses exist for the digest list:



To get messages 123 through 145 (a maximum of 100 per request), mail:


To get an index with subject and author for messages 123-456 , mail:


They are always returned as sets of 100, max 2000 per request,
so you'll actually get 100-499.

To receive all messages with the same subject as message 12345,
send a short message to:


The messages should contain one line or word of text to avoid being
treated as s...@m, but I will ignore their content.
Only the ADDRESS you send to is important.

You can start a subscription for an alternate address,
for example "j...@host.domain", just add a hyphen and your
address (with '=' instead of '@') after the command word:


To stop subscription for this address, mail:


In both cases, I'll send a confirmation message to that address. When
you receive it, simply reply to it to complete your subscription.

If despite following these instructions, you do not get the
desired results, please contact my owner at
users-ow...@subversion.apache.org. Please be patient, my owner is a
lot slower than I am ;-)

--- Enclosed is a copy of the request I received.

Return-Path:
Received: (qmail 87501 invoked by uid 99); 12 Mar 2010 07:19:47 -
Received: from nike.apache.org (HELO nike.apache.org) (192.87.106.230)
 by apache.org (qpsmtpd/0.29) with ESMTP; Fri, 12 Mar 2010 07:19:47 +
X-ASF-Spam-Status: No, hits=-2.9 required=10.0

tests=ASF_EMPTY_LIST_OPS,ASF_LIST_OPS,EMPTY_MESSAGE,FREEMAIL_FROM,MISSING_SUBJECT,RCVD_IN_DNSWL_NONE,SPF_PASS,T_TO_NO_BRKTS_FREEMAIL
X-Spam-Check-By: apache.org
Received-SPF: pass (nike.apache.org: domain of kaykay.uni...@gmail.com 
designates 209.85.222.186 as permitted sender)
Received: from [209.85.222.186] (HELO mail-pz0-f186.google.com) (209.85.222.186)
 by apache.org (qpsmtpd/0.29) with ESMTP; Fri, 12 Mar 2010 07:19:38 +
Received: by pzk16 with SMTP id 16so537189pzk.22
 for; Thu, 11 Mar 2010 
23:19:17 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed;
 d=gmail.com; s=gamma;
 h=domainkey-signature:received:received:message-id:date:from
  :user-agent:mime-version:to:content-type:content-transfer-encoding;
 bh=47DEQpj8HBSa+/TImW+5JCeuQeRkm5NMpJWZG3hSuFU=;
 b=OePh2rvnDppP+Z6epniA3kW4pOXvFvI2LOKEKx2lrqgUyI5hzaqAC9zdUQ1XrYivnj
  FAkaCDIDZkmaoFJVz3lcwWp32ot2kdayrre8hrfwZcOlfcycJkNgtHjbjJeGvJC28yTZ
  fAXrcYo4d9a8sTE6Go2KoL4ImxgoZ8dWLKwFM=
DomainKey-Signature: a=rsa-sha1; c=nofws;
 d=gmail.com; s=gamma;
 h=message-id:date:from:user-agent:mime-version:to:content-type
  :content-transfer-encoding;
 b=ObaVudr09dBZm16NT9Pnn6lwOFljwzU0MBitAMF2Ge308+xs3jxYfEG6G1UU5ujo4e
  o8NagowVw1YiZvAXJ4i87QoBmCUSZZhciqMlaGYF42y3ugClH4wT3dO93eFSSY3n1Wry
  Eg+FxaTxtpkyygY62DHfsbx4KjpswY3Au6HXU=
Received: by 10.142.202.7 with SMTP id z7mr2162627wff.101.1268378356933;
 Thu, 11 Mar 2010 23:19:16 -0800 (PST)
Received: from emerald.haysearch.com (c-67-188-164-101.hsd1.ca.comcast.net 
[67.188.164.101])
 by mx.google.com with ESMTPS id 36sm293416yxh.49.2010.03.11.23.19.15
 (version=SSLv3 cipher=RC4-MD5);
 Thu, 11 Mar 2010 23:19:16 -0800 (PST)
Message-ID:<4b99eaf2.7030...@gmail.com>
Date: Thu, 11 Mar 2010 23:19:14 -0800
From: Kay Kay
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.8) Gecko/20100227 
Thunderbird/3.0.3
MIME-Version: 1.0
To: users-unsubscr...@subversion.apache.o