ccnet-devel Mailing List for CruiseControl.NET

On 1/31/06, Ian Olsen <ia...@so...> wrote:
> That particular request did not make 3.1.7, however your problem may be s=
olved anyway.  Vault's command-line client "getversion" command does respec=
t cloaks, and the Vault sourcecontrol block uses that if you've got 3.1.7 i=
nstalled.

Ah you're right of course. The only reason I use getlabel is it's a
workaround.. I can already see a significant reduction in our nant
build scripts!

--
Perry Ismangil

Perry,

Thanks for the positive feedback.  Using CC.NET internally was indeed a =
great way for us to understand the annoyances, not to mention all the =
usual benefits of CI we now experience.

> Slight issue before I upgrade though, I need get by label to=20
> support cloaking, has that made it in 3.1.7?

That particular request did not make 3.1.7, however your problem may be =
solved anyway.  Vault's command-line client "getversion" command does =
respect cloaks, and the Vault sourcecontrol block uses that if you've =
got 3.1.7 installed.

> 4): Let me get this clear. The CCNET Vault source control=20
> block can now automatically retrieve a specific version, so=20
> there is no danger of a checkin making the build numbers inconsistent?

That's correct. :)

--
Ian=20

> -----Original Message-----
> From: ccn...@li...=20
> [mailto:ccn...@li...] On Behalf Of=20
> Perry Ismangil
> Sent: Tuesday, January 31, 2006 3:28 AM
> To: ccn...@li...
> Subject: Re: [Ccnet-devel] SourceGear Vault integration patch
>=20
> On 1/30/06, Ian Olsen <ia...@so...> wrote:
>=20
> Ian, excellent progress on resolving these issues! Being the=20
> person recommending Vault and then introducing CC.NET in our=20
> development process I feel vindicated. There is no better way=20
> to understand Vault issues with CC.NET than using it yourself, eh!
>=20
> > 1) Solved server contention issues with multiple=20
> simultaneous logins=20
> > when CC.NET starts
> > 2) Significantly faster change polling
> > 3) Significantly faster source retrieval, compared to the=20
> 1.0 release=20
> > with labeling enabled
>=20
> Slight issue before I upgrade though, I need get by label to=20
> support cloaking, has that made it in 3.1.7?
>=20
> > 4) The label, get source, remove label hack is no longer=20
> necessary.  We now retrieve and label a specific version of=20
> the source tree, making the entire integration significantly=20
> simpler and faster.
>=20
> 4): Let me get this clear. The CCNET Vault source control=20
> block can now automatically retrieve a specific version, so=20
> there is no danger of a checkin making the build numbers inconsistent?
>=20
> --
> Perry Ismangil
>=20
>=20
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep=20
> through log files for problems?  Stop!  Download the new AJAX=20
> search engine that makes searching your log files as easy as=20
> surfing the  web.  DOWNLOAD SPLUNK!
> http://sel.as-us.falkag.net/sel?cmd=3Dk&kid=103432&bid#0486&dat=121642
> _______________________________________________
> Ccnet-devel mailing list
> Ccn...@li...
> https://lists.sourceforge.net/lists/listinfo/ccnet-devel

On 1/30/06, Ian Olsen <ia...@so...> wrote:

Ian, excellent progress on resolving these issues! Being the person
recommending Vault and then introducing CC.NET in our development
process I feel vindicated. There is no better way to understand Vault
issues with CC.NET than using it yourself, eh!

> 1) Solved server contention issues with multiple simultaneous logins when=
 CC.NET starts
> 2) Significantly faster change polling
> 3) Significantly faster source retrieval, compared to the 1.0 release wit=
h labeling enabled

Slight issue before I upgrade though, I need get by label to support
cloaking, has that made it in 3.1.7?

> 4) The label, get source, remove label hack is no longer necessary.  We n=
ow retrieve and label a specific version of the source tree, making the ent=
ire integration significantly simpler and faster.

4): Let me get this clear. The CCNET Vault source control block can
now automatically retrieve a specific version, so there is no danger
of a checkin making the build numbers inconsistent?

--
Perry Ismangil

I never got any feedback on the last patch I submitted, have any of the =
maintainers had a chance to look at it?

Even if not, the attached files include all of those changes (see my =
email from Jan 10 2005 below) plus vastly improved integration with =
Vault 3.1.7, which was released today:

1) Solved server contention issues with multiple simultaneous logins =
when CC.NET starts
2) Significantly faster change polling
3) Significantly faster source retrieval, compared to the 1.0 release =
with labeling enabled
4) The label, get source, remove label hack is no longer necessary.  We =
now retrieve and label a specific version of the source tree, making the =
entire integration significantly simpler and faster.

We detect Vault's version so that people running < 3.1.7 can continue to =
run as before.

Note that the Vault3.cs included here replaces the old Vault.cs, and =
Vault317.cs and VaultVersionChecker.cs have been added. =20

Unit tests for the new 3.1.7 integration are also included.  If there's =
anything I can do to help get these changes included, let me know, I'm =
more than happy to do so! :)  I'll definitely update the Vault docs on =
the wiki once there's a build that includes this stuff.

Thanks,

--
Ian Olsen
Developer
SourceGear

> -----Original Message-----
> From: Ian Olsen [mailto:ia...@so...]=20
> Sent: Tuesday, January 10, 2006 1:09 PM
> To: 'ccn...@li...'
> Subject: RE: [Ccnet-devel] SourceGear Vault integration patch
>=20
> Here (finally) is the Vault patch with the changes Owen=20
> requested.  (Note it's the full files, not an actual patch.)
>=20
> Since it's been so long, I'll reiterate what was done:
>=20
> CCNET-613: failure to label isn't handled correctly
> CCNET-614: useWorkingFolder with label isn't handled correctly
> CCNET-615: adds and deletes aren't correctly parsed when=20
> checking mods, so you can't filter them
> CCNET-616: logging is confusing when using multiple source=20
> control blocks
> CCNET-617: "checkin" for date/time on retrieved files is hard-coded
> CCNET-618: can't empty source folder before retrieval
>=20
> There are a number of documentation updates which I'll make=20
> once the patch makes it into a build:
>=20
> - How to use vault's rememberlogin command to "deposit"=20
> authentication information
> - Clarify what "useWorkingDirectory" means, particularly at=20
> it relates to the "workingDirectory" setting
> - If applyLabels is turned on, publishExceptions must also be=20
> turned on for the label to be removed in the event of a failed build.
> - cleanCopy property has been added to optionally empty out=20
> the source directory before retrieving new source
> - setFileTime property has been added for anybody who doesn't=20
> like the "checkin" default
>=20
> --
> Ian=20

Hi all

I found the problem : casing !
Nant.Core iso NAnt.Core  in the namespace

now it works like a charm.


sorry for the trouble, I need more sleep ;-)


regards

Hi all

Does anybody have experience with custom Nant trace listeners?
I want to write a fix for CCNET-625 : Provide more detailed feedback from
build during build process.

My initial idea was to write a custom Nant listener.
This listener provides all the info we need : see [1]
Writing the listener is no problem, (see quick VB code at [2])
Preliminary idea : write the starting of major events to a file,
which than could be read by CCnet and passed to dashboard/cctray.

According to Nant documentation, a custom listener can be used by
Nant -listener:<text>
but I can not get it to load the listener. :-(

I searched the Nant mailinglists, but found nothing use full.

Anybody any idea on getting this to work?


regards

[1]
http://nant.sourceforge.net/release/latest/help/fundamentals/listeners.html

[2]

Imports NAnt.Core

Namespace Nant.Core
    Public Class CCNetListener
        Implements IBuildListener

        Const traceFile As String =3D "c:\temp\tracernant.txt"
        Private targetfile As System.IO.StreamWriter

        Public Sub BuildFinished(ByVal sender As Object, ByVal e As
BuildEventArgs) Implements IBuildListener.BuildFinished

        End Sub

        Public Sub BuildStarted(ByVal sender As Object, ByVal e As
BuildEventArgs) Implements IBuildListener.BuildStarted
            targetfile =3D New IO.StreamWriter(traceFile, True)
            targetfile.WriteLine("Starting {0}", e.Project)
            targetfile.Close()
        End Sub

        Public Sub MessageLogged(ByVal sender As Object, ByVal e As
BuildEventArgs) Implements IBuildListener.MessageLogged

        End Sub

        Public Sub TargetFinished(ByVal sender As Object, ByVal e As
BuildEventArgs) Implements IBuildListener.TargetFinished

        End Sub

        Public Sub TargetStarted(ByVal sender As Object, ByVal e As
BuildEventArgs) Implements IBuildListener.TargetStarted
            targetfile =3D New IO.StreamWriter(traceFile, True)
            targetfile.WriteLine("Starting target {0} of project {1}",
e.Target, e.Project)
            targetfile.Close()
        End Sub

        Public Sub TaskFinished(ByVal sender As Object, ByVal e As
BuildEventArgs) Implements IBuildListener.TaskFinished

        End Sub

        Public Sub TaskStarted(ByVal sender As Object, ByVal e As
BuildEventArgs) Implements IBuildListener.TaskStarted
            targetfile =3D New IO.StreamWriter(traceFile, True)
            targetfile.WriteLine("Starting task {0} of target {1} in projec=
t
{2}", e.Task, e.Target, e.Project)
            targetfile.Close()
        End Sub
    End Class
End Namespace

I ran in to the same problem earlier this week and do have a potential
fix, I just haven't had the time to post the diff yet. The change you
need to make is in the CvsHistoryParser.cs file to the regular
expression used to match the date line.

=20

Here is the patch:

=20

diff -u -r1.1 CvsHistoryParser.cs

--- CvsHistoryParser.cs   16 Dec 2005 18:05:59 -0000       1.1

+++ CvsHistoryParser.cs            27 Jan 2006 16:21:15 -0000

@@ -45,7 +45,7 @@

=20

                        private string currentLine;

                        private string workingFileName;

-                       private Regex dateLineRegex =3D new
Regex(@"date:\s+(?<date>\S+)\s+(?<time>\S+)\s*(?<timezone>\S*);\s+author
:\s+(?<author>.*);\s+state:\s+(?<state>.*);(\s+lines:\s+\+(?<line1>\d+)\
s+-(?<line2>\d+))?");

+                      private Regex dateLineRegex =3D new
Regex(@"date:\s+(?<date>\S+)\s+(?<time>\S+)\s*(?<timezone>\S*);\s+author
:\s+(?<author>.*);\s+state:\s+(?<state>\S*);(\s+lines:\s+\+(?<line1>\d+)
\s+-(?<line2>\d+))?");

=20

                        public Modification[] Parse(TextReader cvsLog,
DateTime from, DateTime to)

                        {

=20

________________________________

From: ccn...@li...
[mailto:ccn...@li...] On Behalf Of Robert
Angers
Sent: Friday, January 27, 2006 11:00 AM
To: ccn...@li...
Subject: [Ccnet-devel] Build result email

=20

In the email sent following a "Build Successful".  The file list, always
indicates the file as "added".  Eventhough I have deleted or modified=
 a
file ?!?!

In the email sent following a "Build Successful".  The file list, always =
indicates the file as "added".  Eventhough I have deleted or modified a =
file ?!?!

On 1/27/06, Owen Rogers <exo...@gm...> wrote:
> hmm... you'll notice that the first part of the url points to the xml
> file to download.  the file is not actually stored at this location;
> instead, the ccnet dashboard uses a custom httphandler to intercept
> the request and then route the call to the ccnet server to get it to
> stream back the contents of the ccnet log file.  i wonder if some part
> of the set up for the http handler could be missing.  i believe their
> is a step in the installer that creates an association between .xml
> extensions and the asp.net process.  maybe this is missing?

Yes that would be it.  How do I do this manually and or with a nAnt script?

P.S. is there a url to get the latest and greatest log file without
knowing the log files name?

--
Jay Flowers
----------------------------------------------------------------------
http://jayflowers.com
---------------------------------------------------------------------

On 26/01/06, Jay Flowers <jay...@gm...> wrote:
http://efbuild1/EF_COTS/default.aspx?_action_ViewBuildLog=3Dtrue&server=3DE=
F_COTS&project=3DEF_COTS&build=3Dlog20060126074946Lbuild.1.6.1.151.xml
>
> The url for the Click here to open... is:
> http://efbuild1/EF_COTS/log20060126074946Lbuild.1.6.1.151.xml?_action_Xml=
BuildLog=3Dtrue&server=3DEF_COTS&project=3DEF_COTS&build=3Dlog2006012607494=
6Lbuild.1.6.1.151.xml

so, this is the link that is failing?  comparing it to the link that i
get on ccnetlive, it looks fine.

> This is for CCNET version 1.0 Final.

so is ccnetlive and it works fine for it.  i wonder if this could be
an iis configuration issue.  we saw similar problems when trying to
get cctray to download when the directory was set up to have execute
permissions.

hmm... you'll notice that the first part of the url points to the xml
file to download.  the file is not actually stored at this location;
instead, the ccnet dashboard uses a custom httphandler to intercept
the request and then route the call to the ccnet server to get it to
stream back the contents of the ccnet log file.  i wonder if some part
of the set up for the http handler could be missing.  i believe their
is a step in the installer that creates an association between .xml
extensions and the asp.net process.  maybe this is missing?
cheers,
owen.

--
Owen Rogers | http://dotnetjunkies.com/weblog/exortech |
CruiseControl.NET - http://ccnet.thoughtworks.com

The returned page is immediate.  It is a "The page can not be found." error=
.

The url for the View Build Log Project Plugin is:
http://efbuild1/EF_COTS/default.aspx?_action_ViewBuildLog=3Dtrue&server=3DE=
F_COTS&project=3DEF_COTS&build=3Dlog20060126074946Lbuild.1.6.1.151.xml

The url for the Click here to open... is:
http://efbuild1/EF_COTS/log20060126074946Lbuild.1.6.1.151.xml?_action_XmlBu=
ildLog=3Dtrue&server=3DEF_COTS&project=3DEF_COTS&build=3Dlog20060126074946L=
build.1.6.1.151.xml

This is for CCNET version 1.0 Final.

--
Jay Flowers
----------------------------------------------------------------------
http://jayflowers.com
---------------------------------------------------------------------

Hi Owen

let me know what you think of it ;-)


regards

On 1/26/06, Owen Rogers <exo...@gm...> wrote:
>
> hi ruben,
>
> On 25/01/06, Ruben Willems <rub...@gm...> wrote:
> >  This patch adds a name to the trigger, so we can identify the trigger
> > causing a build.
> >  If no name is specified, the default name is the trigger type :
> > 'IntervalTrigger', for an interval trigger.
> >  The name of the trigger causing the will be passed to Nant as the
> property:
> > 'ccnettriggerid'
>
> thanks for the patch!! i've only had a preliminary look at your patch,
> but it looks great -- exactly what i had in mind for implementing this
> functionality.  i like the fact that you've wrapped the BuildCondition
> in a class (BuildInformation).  it totally makes sense and it is
> in-line with my design goal of getting enums out of our interfaces.
>
> i hope to have more time to look at it tomorrow night.
> cheers,
> owen.
> --
> Owen Rogers | http://dotnetjunkies.com/weblog/exortech |
> CruiseControl.NET - http://ccnet.thoughtworks.com
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log
> files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> http://sel.as-us.falkag.net/sel?cmdlnk&kid=103432&bid#0486&dat=121642
> _______________________________________________
> Ccnet-devel mailing list
> Ccn...@li...
> https://lists.sourceforge.net/lists/listinfo/ccnet-devel
>

hi ruben,

On 25/01/06, Ruben Willems <rub...@gm...> wrote:
>  This patch adds a name to the trigger, so we can identify the trigger
> causing a build.
>  If no name is specified, the default name is the trigger type :
> 'IntervalTrigger', for an interval trigger.
>  The name of the trigger causing the will be passed to Nant as the proper=
ty:
> 'ccnettriggerid'

thanks for the patch!! i've only had a preliminary look at your patch,
but it looks great -- exactly what i had in mind for implementing this
functionality.  i like the fact that you've wrapped the BuildCondition
in a class (BuildInformation).  it totally makes sense and it is
in-line with my design goal of getting enums out of our interfaces.

i hope to have more time to look at it tomorrow night.
cheers,
owen.
--
Owen Rogers | http://dotnetjunkies.com/weblog/exortech |
CruiseControl.NET - http://ccnet.thoughtworks.com

hi jay,

On 25/01/06, Jay Flowers <jay...@gm...> wrote:
> The "Click here to open log in its own page" does not for for me.
> Anyone else get this to work?

one thing about clicking on this link is that i caused the entire xml
log file to be loaded into your browser.  most browsers (at least ie
and firefox) load this into their own dom (so you can expand and
collapse nodes).  from what i've seen this can consume huge amounts of
memory even for a moderate size build file (say 10 megs).  so, i)
depending on your server, the connection might have timed out while
the log was being downloaded to your browser or ii) your browser might
have been crippled from trying to load this big xml file.  those are
just some possibilities off the top of my head.

could you provide some more details in terms of what you are experiencing?
cheers,
owen.
--
Owen Rogers | http://dotnetjunkies.com/weblog/exortech |
CruiseControl.NET - http://ccnet.thoughtworks.com

Works for me (I'm still on RC2 though) 
 


-----Original Message-----
From: ccn...@li...
[mailto:ccn...@li...] On Behalf Of Jay Flowers
Sent: Thursday, 26 January 2006 8:50 a.m.
To: ccn...@li...
Subject: [Ccnet-devel] Click here to open log in its own page


The "Click here to open log in its own page" does not for for me. 
Anyone else get this to work?

The "Click here to open log in its own page" does not for for me.=20
Anyone else get this to work?

--
Jay Flowers
----------------------------------------------------------------------
http://jayflowers.com
---------------------------------------------------------------------

Something I wrote a while back, I'm wondering if you think it's relevant to
add this to the wiki (under using CCNet with NAnt maybe), what do you think
?


original post :
http://www.dotnetguru2.org/tbarrere/index.php?p=3D229&more=3D1&c=3D1&tb=3D1=
&pb=3D1

<updateversion major=3D"-1" minor=3D"-1" build=3D"${build.number}"
revision=3D"-1" verbose=3D"true">
       <fileset basedir=3D".">
               <include name=3D"**/AssemblyInfo.cs"/>
               <include name=3D"**/CommonAssemblyInfo.cs"/>
               <include name=3D"**/AssemblyInfo.cpp"/>
       </fileset>
</updateversion>



Thibaut Barr=E8re
--
[blog] http://www.dotnetguru2.org/tbarrere

hi philip,

On 21/01/06, Philip Mateescu <phi...@da...> wrote:
> Would you guys be willing to accept two patches to CCTray:
> - stores the path to the settings file in the Registry so that it can be =
stored in a different location
> - allows CCTray to load a remote settings file (HTTP)
>
> The reason for both is that all our developers monitor the same project s=
o it would be nice if we could either store the file in a different place -=
 so that it gets updated when source-code gets updated - or access it remot=
ely - so that it's always up-to-date.

both patches seem reasonable.  one option that had been discussed was
to have cctray pick up the projects to monitor from the ccnet
dashboard.  the dashboard would provide a central location to
aggregate the results across multiple ccnet servers.  you would still
need to specify the location of the dashboard; however, the same holds
true even in the event of a cctray configuration service.

with the new 'dowload cctray' in the webdashboard, it would be nice to
be able to have cctray be able to point back to the dashboard that it
was downloaded from.  i don't have a great solution for this.

btw, one quick and dirty approach that we've been using is to upload a
custom zip file to the dashboard in the cctray folder.  this zip file
contains the cctray binaries plus the configuration file that points
to the right servers.  this at least makes it easy for users to get up
and going quickly.  if i remember correctly, when clicking on the
cctray dashboard link it will return the first file it finds in the
cctray directory with a name of cctray*.*.

anyone done anything better?

as to the initial question, please feel free to post patches to this
list.  or better yet, open a new issue on the jira server and upload
your patches there: http://jira.public.thoughtworks.org/browse/CCNET

(incidentally, i believe that i've fixed the permissions so that files
can now be uploaded to jira by logged in users).

cheers,
owen.

--
Owen Rogers | http://dotnetjunkies.com/weblog/exortech |
CruiseControl.NET - http://ccnet.thoughtworks.com

Hi,

Would you guys be willing to accept two patches to CCTray:
- stores the path to the settings file in the Registry so that it can be =
stored in a different location
- allows CCTray to load a remote settings file (HTTP)=20

The reason for both is that all our developers monitor the same project =
so it would be nice if we could either store the file in a different =
place - so that it gets updated when source-code gets updated - or =
access it remotely - so that it's always up-to-date.

Thanks.
philip

Hi Matthew

for my current customer (must have around 90 ccnet projects today) what I d=
o
is generate ccnet.config from a much smaller xml file (basically : the list
of all projects) using a XSLT transform.

I automated this generation by creating a custom CCNet project which
triggers a NAnt script relying on the NAnt 'style' task.

If I had to do it again right now I'd invoke a CodeSmith template (easier t=
o
maintain, but probably just a simple in terms of integration), or a Ruby ER=
B
or xml builder generated file.

For very simple cases I'd use xml entities (as advertised by someone
recently in the user list if my memory serves me).

I definitely see no good point to manually maintain this file either :)

cheers

Thibaut
--
[blog] http://www.dotnetguru2.org/tbarrere

On 05/01/06, Matthew Bass <Mat...@sa...> wrote:
>
> Hello fellow CCers.
>
> I'm curious if anyone is currently working on an admin tool of some sort
> for CruiseControl.NET's main XML config file? I hate having to update the
> file by hand. I would find it quite convenient if the config file could b=
e
> updated from the dashboard. Even a desktop tool would be better than
> nothing.
>
> I'd be interested in writing and contributing one myself if nobody is
> currently working on one.
>
> Thanks,
> Matt
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log
> files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> http://ads.osdn.com/?ad_idv37&alloc_id=16865&opclick
> _______________________________________________
> Ccnet-devel mailing list
> Ccn...@li...
> https://lists.sourceforge.net/lists/listinfo/ccnet-devel
>

Hi

I'd like to add some content to the wiki: the source code of a NAnt task to
force the build of a given CCNet project (to be used inside a NAnt script).

I was thinking of adding it somewhere in
http://confluence.public.thoughtworks.org/display/CCNET/Using+CruiseControl=
.NET+with+NAnt,
but maybe another place would be more appropriate (or maybe in the contrib
area ?)

What do you think?

regards

Thibaut
--
[blog] http://www.dotnetguru2.org/tbarrere

On 19/01/06, ste...@th... <ste...@th...> wrote:
> After getting CC.Net working with msbuild (which was pretty easy) I
> started to look at the MSTest reports.  I ended up downloading a copy of
> the stylesheet from
> http://confluence.public.thoughtworks.org/display/CCNETCOMM/XSL+Transfor
> ms which I used as the basis for the detailed report.  I would recommend
> that this be included in the release to suppliment the summary report
> that already exists.

i agree totally.  the main thing that was pending was verifying the
quality of the stylesheet.  i'm content to go with your recommendation
as a second opinion.
i'll commit your updates to the stylesheets as well as try to process
some of the other pending submissions tonight.
thanks again for your help with this steve!
cheers,
owen.

--
Owen Rogers | http://dotnetjunkies.com/weblog/exortech |
CruiseControl.NET - http://ccnet.thoughtworks.com

On 19/01/06, ste...@th... <ste...@th...> wrote:
> I've attached a diff file to the Simian summary report so that it only
> shows it's section in the email / dashboard if there is data in the log
> file for Simian.  This appears to be more consistent with the other
> summary reports (NUnit, NCover, FxCop etc).

awesome.  i've been meaning to do this for ages.
thanks for the patch.
cheers,
owen.
--
Owen Rogers | http://dotnetjunkies.com/weblog/exortech |
CruiseControl.NET - http://ccnet.thoughtworks.com

The final email for xsl changes (for now ;).

One thing that appeared to be missing for .NET 2.0 / VS 2005 was a
report for the code coverage functionality from Team Suite.  After a bit
of digging I managed to find the code coverage data that is produced
when running the unit tests (for reference a file named data.coverage in
a folder below where the test output is written).  The problem is that
this file is in a binary format.  I found an article online
(http://blogs.msdn.com/ms_joc/articles/406608.aspx) which gave a basic
example of how to extract the data from this file (be warned, for a 6000
line c# project the resulting xml was around 3MB which when merged into
each build can produce some large logs over time).  I did have to make
some changes from the example to get it working correctly and I can
provide some sample code to do this (I will eventually put this in a
NAnt or MSBuild task when I get a chance).

Once you have got the xml output, you can then just merge this into the
log file and process the xml file as you do for any other log in CC.Net.

I've attached a working version of a summary report (usable on the
dashboard / email report) and a full report (not much testing outside of
the 1 project I've been using as a guinea pig at the moment I'm afraid).

We might need to do a more detailed setup example for the confluence
site similar to the msbuild example for this.

Steve