RE: SVN server email notification function (SMTP) and How To configure it
Hi Ryan, Thanks a lot for your reply. so far I did search those tools such as commit_malier.py, svnmailer, SVN-Notify, all are required a lot of scripts writing, I do not find any tool with GUI that easily for administrator to do a setup and configure email notification. Any idea will be much appreciated! Regards, Zhao Wei -Original Message- From: Ryan Schmidt [mailto:subversion-2...@ryandesign.com] Sent: Wednesday, 8 November, 2017 3:05 AM To: Zhaowei Mao Cc: Subversion Users Subject: Re: SVN server email notification function (SMTP) and How To configure it On Nov 7, 2017, at 04:35, Zhaowei Mao wrote: > We use SVN server (CollabNetSubversion-server-1.5.5-1.win32) for years, I’d > like to configure email notification on server so that may trace any > changes in repositories. It is very likely that commit_email.pl is bundled > with Subversion, which allow administrator to configure it in hook (use > post commit hook) in repository. > > Other than email.pl, can anyone recommend any SMTP component or mailer > works for SVN but allow administrator to do configuration via GUI? As you can see at the top of the script: https://svn.apache.org/repos/asf/subversion/trunk/contrib/hook-scripts/commit-email.pl.in commit-email.pl is deprecated. The replacement is mailer.py: https://svn.apache.org/repos/asf/subversion/trunk/tools/hook-scripts/mailer/mailer.py
Minor Python bindings/API question
Hi,
Im trying to use the python-swig bindings to copy the log entries of our
repository into a database for indexing and cross referencing. There is one
minor problem: The callback seems to be called a few times with invalid log
entries:
(-1L, {}, None)
That is obviously easy to filter out, and there dont seem to be any gaps in
the revisions. Im just wondering if this points to repository problems, or
if these are just side effects of the database schema or API. Is it possible
that these are failed transactions?
* Debian 3.16.39-1+deb8u2
* svn, version 1.9.5 (r1770682)
Code:
start = svn.core.svn_opt_revision_t()
start.kind = svn.core.svn_opt_revision_number
start.value.number = startRev
end = svn.core.svn_opt_revision_t()
end.kind = svn.core.svn_opt_revision_head
revision_range = svn.core.svn_opt_revision_range_t()
revision_range.start = start
revision_range.end = end
limit = 0
discover_changed_paths = True
strict_node_history = True
include_merged_revisions = True
revprops = None
svn.client.log5(targets, end, (revision_range,), limit,
discover_changed_paths, strict_node_history, include_merged_revisions,
revprops, _receiver_callback, ctx)
Cheers,
Jens
Re: Apache subversion unable to open another ext4 partion svn repository
Nico: Please state references to these bugs. I have not heard anything
about them.
Setting it to "permissive" is fine if you are going to actually use
"audit2allow" to set a policy and then turn things back on. Otherwise it
is slower and accumulating data unnecessarily.
On Tue, Nov 7, 2017 at 6:55 PM, Nico Kadel-Garcia wrote:
>
>
> On Tue, Nov 7, 2017 at 8:16 AM, Doug Robinson
> wrote:
>
>> Kushal:
>>
>> Just as a slight possibility, you might check to see that the SELinux
>> stuff is turned off (use the "sestatus" command). If it's not disabled/off
>> then you're going to have to allow the http account to access that data by
>> setting the appropriate security context throughout all directories from
>> the root ('/') down to where the SVN data is stored in that 2nd EXT4
>> partition - and then the entire SVN data tree.
>>
>
> Do *not* disable SELinux while debugging it. Set it to "permissive".
>
> There have been a serious of nasty bugs when switching from "enabled" to
> "disabled", the likes of which get senior sysadmins paid good money to say
> that sort of things.
>
>
>
>> Cheers.
>>
>> Doug
>> --
>> *DOUGLAS B ROBINSON* SENIOR PRODUCT MANAGER
>>
>> T +1 925 396 1125
>> *E* doug.robin...@wandisco.com
>>
>> World Leader in Active Data Replication™
>> *Find out more wandisco.com
Re: Apache subversion unable to open another ext4 partion svn repository
On Wed, Nov 8, 2017 at 5:02 PM, Doug Robinson
wrote:
> Nico: Please state references to these bugs. I have not heard anything
> about them.
>
There have been multiple. https://bugzilla.redhat.com/show_bug.cgi?id=692573
shows up in a fast Google search: that was admittedly from 2011. I'll also
acknowledge that it's gotten better. A fast search used to show a lot more
reports of poor behavior with SELinux entirely disabled.
> Setting it to "permissive" is fine if you are going to actually use
> "audit2allow" to set a policy and then turn things back on. Otherwise it
> is slower and accumulating data unnecessarily.
>
In my experience, it's not noticeably slower than running SELinux at all.
And "permissive" can be a very useful tool to guide where and how, you
arrange Subversion server components for consistentcy with the already
activated web server SELinux policies. It's also helpful to point out where
your models of web configuration may be a little.. odd, and themselves
a source of error.
On Tue, Nov 7, 2017 at 6:55 PM, Nico Kadel-Garcia wrote:
>
>>
>>
>> On Tue, Nov 7, 2017 at 8:16 AM, Doug Robinson > > wrote:
>>
>>> Kushal:
>>>
>>> Just as a slight possibility, you might check to see that the SELinux
>>> stuff is turned off (use the "sestatus" command). If it's not disabled/off
>>> then you're going to have to allow the http account to access that data by
>>> setting the appropriate security context throughout all directories from
>>> the root ('/') down to where the SVN data is stored in that 2nd EXT4
>>> partition - and then the entire SVN data tree.
>>>
>>
>> Do *not* disable SELinux while debugging it. Set it to "permissive".
>>
>> There have been a serious of nasty bugs when switching from "enabled" to
>> "disabled", the likes of which get senior sysadmins paid good money to say
>> that sort of things.
>>
>>
>>
>>> Cheers.
>>>
>>> Doug
>>> --
>>> *DOUGLAS B ROBINSON* SENIOR PRODUCT MANAGER
>>>
>>> T +1 925 396 1125
>>> *E* doug.robin...@wandisco.com
>>>
>>> World Leader in Active Data Replication™
>>> *Find out more wandisco.com
