On 05/23/2011 09:09 PM, Paul Burba wrote:
> On Mon, May 23, 2011 at 10:13 AM, Kamesh Jayachandran<kamesh_at_collab.net> wrote:
>> On 05/18/2011 02:04 AM, Paul Burba wrote:
>>>> Author: kameshj
>>>> Date: Thu Feb 25 13:40:22 2010
>>>> New Revision: 916286
>>>>
>>>> URL: http://svn.apache.org/viewvc?rev=916286&view=rev
>>>> Log:
>>>> With the below apache configuration(See the trailing slash at the end
>>>> of '/svn/').
>>>>
>>>> <Location /svn/>
>>>> DAV svn
>>>> SVNParentPath /repositories
>>>> #See the trailing slash on the master URI also can cause the confusion.
>>>> SVNMasterURI http://master/svn/
>>>> SVNAdvertiseV2Protocol Off
>>>> </Location>
>>>>
>>>> We get the following error on the client side.
>>>>
>>>> svn: Commit failed (details follow):
>>>> svn: MKACTIVITY of
>>>> '/svn/demujin/!svn/act/4b6d547c-018d-4e02-9d3f-2b283076cc06': Could
>>>> not
>>>> read status line: connection was closed by server (http://localhost)
>>> Hi Kamesh,
>>>
>> Sorry for the late response Paul.
>>> Are there any threads or IRC logs in which this is discussed
>>> (particularly a more detailed replication)?
>> No. Paul this error caught my eyes while testing something locally.
>>
>>> While reviewing r916286
>>> and r917512 I tried to replicate this problem by adding a trailing '/'
>>> to the location and SVNMasterURI. I *did* get an error, just a
>>> different one:
>>>
>>>
>>> C:\SVN\src-branch-1.6.x\Debug\subversion\tests\cmdline\svn-test-work\working_copies\slave>svn
>>> st
>>> M file.txt
>>>
>>>
>>> C:\SVN\src-branch-1.6.x\Debug\subversion\tests\cmdline\svn-test-work\working_copies\slave>svn
>>> ci -m "commit to slave"
>>> svn: Commit failed (details follow):
>>> svn: OPTIONS of 'http://localhost/svn-test-work/slave': 200 OK
>>> (http://localhost)
>>>
>>
>>> What's worse is that I get this error with a server at the latest
>>> trunk (r1104523), trunk right after you fixed this bug (r917512), and
>>> my own local attempt to backport this branch to 1.6.x (addressing the
>>> conflicts and the use of 1.7 APIs). They all fail the same way.
>> I could not see this error(I mean everything works as expected with
>> <Location /svn/> and SVNMasterURI witj trailing '/') with against
>> trunk_at_1126459
> Hi Kamesh,
>
> Hmmm, I get yet another error if the Location and SVNMasterURI have a
> trailing '/':
>
> C:\SVN\src-branch-1.6.x\Debug\subversion\tests\cmdline\svn-test-work\working_copies\slave-1.7-r1126459>svn
> st
> M file.txt
>
> slave-1.7-r1126459>svn ci -m "Commit to proxy with trailing / in
> Location and SVNMasterURI"
> ..\..\..\subversion\svn\commit-cmd.c:190: (apr_err=175002)
> ..\..\..\subversion\libsvn_client\commit.c:846: (apr_err=175002)
> ..\..\..\subversion\libsvn_client\commit.c:846: (apr_err=175002)
> svn: E175002: Commit failed (details follow):
> ..\..\..\subversion\libsvn_client\commit.c:644: (apr_err=175002)
> ..\..\..\subversion\libsvn_client\commit.c:644: (apr_err=175002)
> ..\..\..\subversion\libsvn_client\ra.c:356: (apr_err=175002)
> ..\..\..\subversion\libsvn_ra\ra_loader.c:496: (apr_err=175002)
> ..\..\..\subversion\libsvn_ra\ra_loader.c:496: (apr_err=175002)
> svn: E175002: Unable to connect to a repository at URL
> 'http://localhost/svn-test-work/slave'
> ..\..\..\subversion\libsvn_ra_neon\util.c:1556: (apr_err=175002)
> ..\..\..\subversion\libsvn_ra_neon\util.c:1152: (apr_err=175002)
> svn: E175002: The OPTIONS request returned invalid XML in the
> response: XML parse error at line 1: no element found
> (http://localhost/svn-test-work/slave)
>
> Without the trailing '/' this commit succeeds as expected.
>
>> I built trunk_at_916285(prior to my fix) and saw this error and with
>> trunk_at_916286 this error has gone away.
>>
>> May be something to do with win32 build.
> I have not used the write-through proxy functionality before so maybe
> I botched something with my configuration? I attached my httpd.conf
> in the event something problematic jumps out at you. Though, as I
> said, everything works fine without the trailing '/'.
>
>> Can you attach your patch against r916285, I will build it/test it and let
>> you know what is wrong?
> I don't have a patch against trunk_at_916285, I don't have any patch
> against trunk at all. The only patch I mentioned is for the backport
> of r916286 and r917512 against 1.6.x. Is this what you meant?
>
> Paul
Yes I could reproduce your original error even while doing a checkout
using my distro 1.6 client binaries against the following configuration.
This issue has nothing to do with write through proxy as it happens even
while checking out.
<Location /svn-test-work/slave/>
DAV svn
SVNPath /repositories/thanu
SVNMasterURI http://127.0.0.1:18080/svn-test-work/master/
AuthType Basic
AuthName Subversion
AuthBasicProvider csvn-file-users
AuthzSVNAccessFile "conf.d/svn_access_file"
SVNPathAuthz short_circuit
Require valid-user
</Location>
This error do not occur in the following configurations.
* when SVNParentPath is used instead of SVNPath
* When Location do not have trailing slash.
<snip of error>
[kamesh@kamesh tmp]$ svn co http://127.0.0.1/svn-test-work/slave
svn: OPTIONS of 'http://127.0.0.1/svn-test-work/slave': 200 OK
(http://127.0.0.1)
</snip>
With regards
Kamesh Jayachandran
Received on 2011-05-23 18:07:37 CEST