statcvs-users — StatCVS Users Mailing List

Re: [Statcvs-users] [PATCH] Make the author page display real names (if available)

[Richard C. <ri...@cy...>]

On 1 Dec 2005, at 01:31, Steffen Pingel wrote:

> Hi,
>
>>> AuthorProfileFactory? This sounds like the classic situation for a
>>> factory
>>> pattern.
>>
>> This implies that you have an AuthorProfile class with many
>> instances. Can we do without?
>
> Why would we want to? Creating these objects is a one time task,  
> just as
> parsing the cvs log file, and should be done on start up. IMHO a  
> simple
> Profile class with a couple of getters and setters will do the job.

OK, you win. Let's have an AuthorProfile type.

> But while
> we are at it, we may want to go with an interface that specifies  
> the get
> methods and provide a default implementation that has the setters  
> as well.

Nah. I don't see why an interface should be necessary at the current  
state. I don't want to generalize prematurely.

Would that cover your needs?

interface AuthorProfileFactory
    Profile getProfile(String loginName);

class AuthorProfile
    String getFullName();
    String getImageURL();
    String getWebPageURL();
    + setters

Richard



>>> Presentation layer classes would merely need to call
>>> Author.getProfile()
>>> and then render that profile any way they like.
>>
>> That's where I disagree. I'd prefer the log parsing and
>> representation to make no assumptions about the application that uses
>> the data. Having Author.getProfile() would add such an assumption to
>> the log representation.
>
> Yes, I certainly agree. The profile stuff should be kept separate  
> from the
> model code as there is no apparent connection to the cvs log.
>
>> I want to be able to re-use the low-level
>> classes (like Author) without having a dependency that forces me to
>> drag higher-level classes (like the Profile stuff) into my new  
>> project.
>
> Exactly. StatCvs-XML which shares the model code has a very  
> different approach
> to the profile . For maven we have a straight forward dom parser  
> that stores
> results in a map of maps.
>
>  SAXBuilder builder = new SAXBuilder();
>  Document suite = builder.build(file);
>  Element element = suite.getRootElement().getChild("developers");
>  if (element != null) {
>      for (Iterator it = element.getChildren().iterator(); it.hasNext 
> ();) {
>          Element developer = (Element)it.next();
>          String id = developer.getChildText("id");
>          if (id == null) {
>              continue;
>          }
>          if (developer.getChildText("name") != null) {
>              settings.setFullname(id, developer.getChildText("name"));
>          }
>          if (developer.getChildText("image") != null) {
>              settings.setAuthorPic(id, developer.getChildText 
> ("image"));
>          }
>          else if (developer.getChildText("url") != null) {
>              settings.setAuthorPic(id, developer.getChildText 
> ("url") + "/" +
> id + ".png");
>          }
>      }
>  }
>
> If anyone were to define a profile class I would probably go ahead and
> refactor the code though :).
>
> Steffen
>
>
> -------------------------------------------------------
> 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_id=7637&alloc_id=16865&op=click
> _______________________________________________
> Statcvs-users mailing list
> Sta...@li...
> https://lists.sourceforge.net/lists/listinfo/statcvs-users
>

Re: [Statcvs-users] statcvs output is just "null"

[Steffen P. <ste...@gm...>]

Hi,

> I created the log on the main branch.  Here is a small section of the log
> that produces the error.  I am not sure if I cut the log properly (I cut
> before and after the equal sign dividers).

It worked on my system using statcvs 0.2.2 downloaded from sf:

# java -jar statcvs-0.2.2.jar cvs.log . -output-dir output/ -verbose
StatCVS - CVS statistics generation

Parsing CVS log 'cvs.log'
Generating report for cvsroot into output/
Using default CSS file (statcvs.css)
could not get line count for 
'./tcdriver/db2/workflow/DB2_Product_Uninstall.wkf': 
java.io.FileNotFoundException: ./tcdriver/db2/workflow/DB2_Product_Uninstall.wkf 
(No such file or directory)
could not get line count for './tcdriver/db2/workflow/DB2_User_Add.wkf': 
java.io.FileNotFoundException: ./tcdriver/db2/workflow/DB2_User_Add.wkf (No 
such file or directory)
could not get line count for './tcdriver/db2/workflow/DB2_User_Delete.wkf': 
java.io.FileNotFoundException: ./tcdriver/db2/workflow/DB2_User_Delete.wkf 
(No such file or directory)
Creating CSS file at 'output/statcvs.css'
Creating page 'Authors'
Creating page 'Authors'
Creating page 'Development Statistics for  cvsroot'
Creating page 'Lines of Code'
Creating page 'File Sizes and File Counts'
Creating page 'Directory Sizes'
Creating page 'Module '
Creating page 'Module tcdriver/'
Creating page 'Module tcdriver/db2/'
Creating page 'Module tcdriver/db2/workflow/'
Creating page 'Commit Log'
Creating page 'User statistics for dconstan'
Creating page 'User statistics for epostea'
Creating page 'User statistics for jpacker'
Creating page 'User statistics for vwtsang'
runtime: 9.598 seconds
memory usage: 1204.0 kb


Could you please run this on your system and send the output (to me personally 
if you like, in case it contains private information)?

Steffen

-- 
Steffen Pingel - ste...@gm... - http://steffenpingel.de

Re: [Statcvs-users] [PATCH] Make the author page display real names (if available)

[Steffen P. <ste...@gm...>]

Hi,

> > AuthorProfileFactory? This sounds like the classic situation for a
> > factory
> > pattern.
>
> This implies that you have an AuthorProfile class with many
> instances. Can we do without?

Why would we want to? Creating these objects is a one time task, just as 
parsing the cvs log file, and should be done on start up. IMHO a simple 
Profile class with a couple of getters and setters will do the job. But while 
we are at it, we may want to go with an interface that specifies the get 
methods and provide a default implementation that has the setters as well. 

> > Presentation layer classes would merely need to call
> > Author.getProfile()
> > and then render that profile any way they like.
>
> That's where I disagree. I'd prefer the log parsing and
> representation to make no assumptions about the application that uses
> the data. Having Author.getProfile() would add such an assumption to
> the log representation. 

Yes, I certainly agree. The profile stuff should be kept separate from the 
model code as there is no apparent connection to the cvs log.

> I want to be able to re-use the low-level 
> classes (like Author) without having a dependency that forces me to
> drag higher-level classes (like the Profile stuff) into my new project.

Exactly. StatCvs-XML which shares the model code has a very different approach 
to the profile . For maven we have a straight forward dom parser that stores 
results in a map of maps.

 SAXBuilder builder = new SAXBuilder();
 Document suite = builder.build(file);
 Element element = suite.getRootElement().getChild("developers");
 if (element != null) {
     for (Iterator it = element.getChildren().iterator(); it.hasNext();) {
         Element developer = (Element)it.next();
         String id = developer.getChildText("id");
         if (id == null) {
             continue;
         }
         if (developer.getChildText("name") != null) {
             settings.setFullname(id, developer.getChildText("name"));
         }
         if (developer.getChildText("image") != null) {
             settings.setAuthorPic(id, developer.getChildText("image"));
         }
         else if (developer.getChildText("url") != null) {
             settings.setAuthorPic(id, developer.getChildText("url") + "/" + 
id + ".png");
         }
     }
 }

If anyone were to define a profile class I would probably go ahead and 
refactor the code though :).

Steffen

Re: [Statcvs-users] [PATCH] Make the author page display real names (if available)

[Richard C. <ri...@cy...>]

Hi James,

On 30 Nov 2005, at 23:43, James Ring wrote:

> Hi Richard,
>
>>
>> Yeah, the name sucks. I think the class should not represent a single
>> profile, but a method of getting author profiles. It's clear what the
>> implementation classes would be, things like PasswdFile,
>> MavenProjectFile, SourceforgeWebsiteScraper. Now find a good common
>> name for these.
>>
>> AuthorProfileSource?
>>
>
> AuthorProfileFactory? This sounds like the classic situation for a  
> factory
> pattern.

This implies that you have an AuthorProfile class with many  
instances. Can we do without?

> You could have the Factory as an abstract class with pretty much  
> the single
> method
>
> public abstract AuthorProfile getProfile( Author a );
>
> and then just subclass it and provide the correct concrete  
> implementation
> depending on the command line options.
>
> Alternatively, you could it as a Factory which calls other factories.
> For example, AuthorProfileFactory may call all its "sub-factories"
> (PasswdAuthorProfileFactory) until it finds a factory which can  
> give back
> a profile for the given author. This would allow StatCVS to source its
> author profiles from many different sources at the same time...  
> Whether
> that's useful or not is another question... :)

Take it easy here ... ;-)

>> Also, I see no need why the Author class should know about the
>> existence of the profiles. I want the Author class to remain as close
>> as possible to the concept of a CVS login. We're discussing a
>> presentation issue, and changing stuff on the presentation layer
>> shouldn't affect the core CVS classes.
>
> Actually, the AuthorProfile shouldn't have anything to do with the
> presentation.

You're right. Presentation was the wrong term.

> How you choose to represent the author's profile is totally
> up to the presentation layer. AuthorProfiles should be concerned  
> only with
> those things which make up an author's profile (like you said, an e- 
> mail
> address, URL, etc).
>
> Presentation layer classes would merely need to call  
> Author.getProfile()
> and then render that profile any way they like.

That's where I disagree. I'd prefer the log parsing and  
representation to make no assumptions about the application that uses  
the data. Having Author.getProfile() would add such an assumption to  
the log representation. I want to be able to re-use the low-level  
classes (like Author) without having a dependency that forces me to  
drag higher-level classes (like the Profile stuff) into my new project.

Hope that makes sense.

Richard



>
>> (Actually, the Author class should probably be called CVSLogin. My
>> suggestions may make more sense if seen from that point of view.)
>
> Ok, maybe I am thinking differently about the role of the Author  
> class.
>
>> Richard
>>
>
> Regards,
> James
>
> PS. I'm now using StatCVS at my work, and the bossman likes it a  
> lot ;)
> -- 
> James Ring
>
>
>
> -------------------------------------------------------
> 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_id=7637&alloc_id=16865&op=click
> _______________________________________________
> Statcvs-users mailing list
> Sta...@li...
> https://lists.sourceforge.net/lists/listinfo/statcvs-users
>

Re: [Statcvs-users] [PATCH] Make the author page display real names (if available)

[James R. <sj...@jd...>]

Hi Richard,

>
> Yeah, the name sucks. I think the class should not represent a single
> profile, but a method of getting author profiles. It's clear what the
> implementation classes would be, things like PasswdFile,
> MavenProjectFile, SourceforgeWebsiteScraper. Now find a good common
> name for these.
>
> AuthorProfileSource?
>

AuthorProfileFactory? This sounds like the classic situation for a factory
pattern.

You could have the Factory as an abstract class with pretty much the single
method

public abstract AuthorProfile getProfile( Author a );

and then just subclass it and provide the correct concrete implementation
depending on the command line options.

Alternatively, you could it as a Factory which calls other factories.
For example, AuthorProfileFactory may call all its "sub-factories"
(PasswdAuthorProfileFactory) until it finds a factory which can give back
a profile for the given author. This would allow StatCVS to source its
author profiles from many different sources at the same time... Whether
that's useful or not is another question... :)

>
> Imagine that profiles may provide other things except real names.
> Email address, web link etc. There's no need to duplicate all these
> methods on both the Author and Profiles class.

Agreed. It was a silly suggestion on my part.

> Also, I see no need why the Author class should know about the
> existence of the profiles. I want the Author class to remain as close
> as possible to the concept of a CVS login. We're discussing a
> presentation issue, and changing stuff on the presentation layer
> shouldn't affect the core CVS classes.

Actually, the AuthorProfile shouldn't have anything to do with the
presentation. How you choose to represent the author's profile is totally
up to the presentation layer. AuthorProfiles should be concerned only with
those things which make up an author's profile (like you said, an e-mail
address, URL, etc).

Presentation layer classes would merely need to call Author.getProfile()
and then render that profile any way they like.

> (Actually, the Author class should probably be called CVSLogin. My
> suggestions may make more sense if seen from that point of view.)

Ok, maybe I am thinking differently about the role of the Author class.

> Richard
>

Regards,
James

PS. I'm now using StatCVS at my work, and the bossman likes it a lot ;)
-- 
James Ring

Re: [Statcvs-users] [PATCH] Make the author page display real names (if available)

[Richard C. <ri...@cy...>]

On 30 Nov 2005, at 22:40, James Ring wrote:
> Ok, I think I agree in principle, but I don't like the name
> (AuthorProfilesIntegration). How about just a superclass  
> AuthorProfile, and
> subclasses like MavenAuthorProfile or PasswdAuthorProfile or
> LDAPAuthorProfile ... ?

Yeah, the name sucks. I think the class should not represent a single  
profile, but a method of getting author profiles. It's clear what the  
implementation classes would be, things like PasswdFile,  
MavenProjectFile, SourceforgeWebsiteScraper. Now find a good common  
name for these.

AuthorProfileSource?


> I still think the method getRealName() should remain on the Author  
> class.
> After all, Authors *do* have real names. Maybe it could just call  
> into the
> author's profile.

Imagine that profiles may provide other things except real names.  
Email address, web link etc. There's no need to duplicate all these  
methods on both the Author and Profiles class.

Also, I see no need why the Author class should know about the  
existence of the profiles. I want the Author class to remain as close  
as possible to the concept of a CVS login. We're discussing a  
presentation issue, and changing stuff on the presentation layer  
shouldn't affect the core CVS classes.

(Actually, the Author class should probably be called CVSLogin. My  
suggestions may make more sense if seen from that point of view.)

Richard



> I don't like the idea of having to say
>
> 	AuthorProfilesIntegration i = getProfileIntegration(); /* or  
> whatever */
> 	i.getRealName( author );
>
> I think that's backwards. ;)
>
> Better to say
> 	
> 	Author a = /* ... */
> 	a.getRealName();
> 	// or
> 	a.getProfile().getRealName();
>
> or something like that.
>
>> This is the same general approach as with File and
>> WebRepositoryIntegration. Please have a look at how these classes are
>> used for inspiration. Note that we have a couple of different
>> WebRepositoryIntegration implementations, and adding new ones is
>> easy. I'd like to have the same possiblity for external author data.
>
> That's fair enough, and I think it's possible with an  
> AuthorProfile, but I'll
> send patches and see what you think.
>
>> Steffen, any insight on this? What would be a good way to support
>> both Maven and passwd files as data sources?
>>
>> Cheers,
>> Richard
>>
>
> Thanks for your time.
>
> Regards,
> James
> -- 
> James Ring

Re: [Statcvs-users] [PATCH] Make the author page display real names (if available)

[James R. <sj...@jd...>]

Hi Richard,

On Thursday 01 December 2005 06:46, Richard Cyganiak wrote:
> Hi James,
>
> Thanks for the patches. I think that adding external author metadata
> (real names, but also email addresses and a link to a web page) is a
> useful addition. But I don't like the way you've architected it.
> Other users may want to get author data from other sources than
> passwd files, and your patch would make it a bit hard to add other
> sources.
>
> I'd suggest you change the patch to keep the passwd-specific logic
> out of the core classes. Some ideas for this:
>
> - Store the real names not in the Author instances, but in a new
> class, say PasswdFile, with a getRealName(Author) method.
> - make PasswdFile an implementation of a new interface, say
> AuthorProfilesIntegration, which has just that method for the moment.

Ok, I think I agree in principle, but I don't like the name=20
(AuthorProfilesIntegration). How about just a superclass AuthorProfile, and=
=20
subclasses like MavenAuthorProfile or PasswdAuthorProfile or=20
LDAPAuthorProfile ... ?

I still think the method getRealName() should remain on the Author class.=20
After all, Authors *do* have real names. Maybe it could just call into the=
=20
author's profile.

I don't like the idea of having to say

	AuthorProfilesIntegration i =3D getProfileIntegration(); /* or whatever */
	i.getRealName( author );

I think that's backwards. ;)

Better to say
=09
	Author a =3D /* ... */
	a.getRealName();
	// or
	a.getProfile().getRealName();

or something like that.

> This is the same general approach as with File and
> WebRepositoryIntegration. Please have a look at how these classes are
> used for inspiration. Note that we have a couple of different
> WebRepositoryIntegration implementations, and adding new ones is
> easy. I'd like to have the same possiblity for external author data.

That's fair enough, and I think it's possible with an AuthorProfile, but I'=
ll=20
send patches and see what you think.

> Steffen, any insight on this? What would be a good way to support
> both Maven and passwd files as data sources?
>
> Cheers,
> Richard
>

Thanks for your time.

Regards,
James
=2D-=20
James Ring

Re: [Statcvs-users] Jcvsweb integration

[Richard C. <ri...@cy...>]

Hi Eric,

Nice one. I didn't know about Jcvsweb. Looks like a nice project.

Did you do the other changes that are necessary for proper  
integration? Like adding a command line switch and setting up the  
JcvswebIntegration instance in ConfigurationOptions?

If you did, could you please create a patch and add it to the  
SourceForge patch tracker [1]?

(Ideally, it would also contain a new section for the manual [2] --  
hint, hint!)

Cheers,
Richard

[1] http://sourceforge.net/tracker/?group_id=57558&atid=484571
[2] http://cvs.sourceforge.net/viewcvs.py/statcvs/htdocs/manual/ 
index.html?rev=HEAD&view=auto


On 21 Nov 2005, at 23:07, Eric Meaney wrote:

> Here is the source for the class I wrote to integrate with Jcvsweb
> (http://www.jcvs.org/jcvsweb/) if anyone is interested.
>
> /**
>  *
>  */
> package net.sf.statcvs.output;
>
> import java.util.HashSet;
> import java.util.Set;
>
> import net.sf.statcvs.model.CvsFile;
> import net.sf.statcvs.model.CvsRevision;
> import net.sf.statcvs.model.Directory;
>
> /**
>  * @author ERMEANEY
>  *
>  */
> public class JcvswebIntegration implements WebRepositoryIntegration {
>     private String baseURL;
>     private Set atticFileNames = new HashSet();
>     /**
>      *
>      */
>     public JcvswebIntegration(String baseURL) {
>         if (baseURL.endsWith("/")) {
>             this.baseURL = baseURL.substring(0, baseURL.length() - 1);
>         } else {
>             this.baseURL = baseURL;
>         }
>     }
>
>     /* (non-Javadoc)
>      * @see net.sf.statcvs.output.WebRepositoryIntegration#getName()
>      */
>     public String getName() {
>         return "JcvsWeb";
>     }
>
>     /* (non-Javadoc)
>      * @see  
> net.sf.statcvs.output.WebRepositoryIntegration#getDirectoryUrl 
> (net.sf.statcvs.model.Directory)
>      */
>     public String getDirectoryUrl(Directory directory) {
>         String path = baseURL + "/HEAD/list/" + directory.getPath();
>         return path;
>     }
>
>     /* (non-Javadoc)
>      * @see  
> net.sf.statcvs.output.WebRepositoryIntegration#getFileHistoryUrl 
> (net.sf.statcvs.model.CvsFile)
>      */
>     public String getFileHistoryUrl(CvsFile file) {
>         return baseURL + "/HEAD/vers/" + file.getFilenameWithPath();
>     }
>
>     /* (non-Javadoc)
>      * @see  
> net.sf.statcvs.output.WebRepositoryIntegration#getFileViewUrl 
> (net.sf.statcvs.model.CvsFile)
>      */
>     public String getFileViewUrl(CvsFile file) {
>         return baseURL + "/HEAD/view/" + file.getFilenameWithPath() +
> "/" + file.getLatestRevision().getRevisionNumber();
>     }
>
>     /* (non-Javadoc)
>      * @see  
> net.sf.statcvs.output.WebRepositoryIntegration#getFileViewUrl 
> (net.sf.statcvs.model.CvsRevision)
>      */
>     public String getFileViewUrl(CvsRevision revision) {
>         return baseURL + "/HEAD/view/" +
> revision.getFile().getFilenameWithPath() + "/" +
> revision.getRevisionNumber();
>     }
>
>     /* (non-Javadoc)
>      * @see  
> net.sf.statcvs.output.WebRepositoryIntegration#getDiffUrl 
> (net.sf.statcvs.model.CvsRevision,
> net.sf.statcvs.model.CvsRevision)
>      */
>     public String getDiffUrl(CvsRevision oldRevision, CvsRevision  
> newRevision) {
>         return baseURL + "/HEAD/pdiff/" +
> newRevision.getFile().getFilenameWithPath() + "/" +
> oldRevision.getRevisionNumber() + "/" +
> newRevision.getRevisionNumber();
>     }
>
>     /* (non-Javadoc)
>      * @see  
> net.sf.statcvs.output.WebRepositoryIntegration#setAtticFileNames 
> (java.util.Set)
>      */
>     public void setAtticFileNames(Set atticFileNames) {
>         this.atticFileNames = atticFileNames;
>     }
>
> }
>
> --
> Eric Meaney
> erm...@gm...
> <JcvswebIntegration.java>

Re: [Statcvs-users] [PATCH] Make the author page display real names (if available)

[Richard C. <ri...@cy...>]

Hi James,

Thanks for the patches. I think that adding external author metadata  
(real names, but also email addresses and a link to a web page) is a  
useful addition. But I don't like the way you've architected it.  
Other users may want to get author data from other sources than  
passwd files, and your patch would make it a bit hard to add other  
sources.

I'd suggest you change the patch to keep the passwd-specific logic  
out of the core classes. Some ideas for this:

- Store the real names not in the Author instances, but in a new  
class, say PasswdFile, with a getRealName(Author) method.
- make PasswdFile an implementation of a new interface, say  
AuthorProfilesIntegration, which has just that method for the moment.

This is the same general approach as with File and  
WebRepositoryIntegration. Please have a look at how these classes are  
used for inspiration. Note that we have a couple of different  
WebRepositoryIntegration implementations, and adding new ones is  
easy. I'd like to have the same possiblity for external author data.

Steffen, any insight on this? What would be a good way to support  
both Maven and passwd files as data sources?

Cheers,
Richard


On 27 Nov 2005, at 02:35, James Ring wrote:

> Hi all,
>
> This patch makes the author page display real names (if available).  
> You can
> see an example of the output generated by this patch at
>
> http://xtuml.jdns.org/statcvs/user_u3957053.html
>
> Please note that only the page title is changed. No other output is  
> affected.
>
> I'll also post this on the SourceForge patch tracker.
>
> Regards,
> James
> -- 
> James Ring
> <author-page-real-name.patch>

Re: [Statcvs-users] statcvs output is just "null"

[Steffen P. <ste...@gm...>]

Hi,

On Friday 25 November 2005 15:58, Jeff McRae wrote:
> I tried to run cvsstat on my cvs log and I get
>
> [tioadmin@jmcrae3 statcvs-0.2.2]$ java -jar statcvs.jar
> /tmp/statcvs-0.2.2/cvs-log.txt /home/tioadmin/cvs/tcdriver
> StatCVS - CVS statistics generation
>
> null
> [tioadmin@jmcrae3 statcvs-0.2.2]$

Make sure the log was created on the MAIN branch (do a fresh checkout or run 
'cvs up -A').

Steffen

-- 
Steffen Pingel - ste...@gm... - http://steffenpingel.de

Re: [Statcvs-users] statcvs output is just "null"

[Richard C. <ri...@cy...>]

Hi Jeff,

I can't think of anything causing this problem right now. Not having  
access to the log, there's not really much we can do, unless you are  
willing to do all the work ...

I suggest you have a look at the CVS log. You'll see that it contains  
one section for each file in the repository. Please copy only the  
first section into a new file and try to run StatCVS on that. Does it  
work?

If it does, then try to find the file that caused the problem (using  
the debug output from your initial file). The files are processed in  
the order they appear in the file, so it shouldn't be hard. Can you  
spot anything suspicious in that file's log? Try to run StatCVS on  
only that part.

Hope that helps,
Richard



On 29 Nov 2005, at 20:23, Jeff McRae wrote:

>
> sta...@li... wrote on 2005-11-25  
> 05:42:36 PM:
>
> > Hi Jeff,
> >
> > On Saturday 26 November 2005 01:58, Jeff McRae wrote:
> > > I tried to run cvsstat on my cvs log and I get
> > >
> > > [tioadmin@jmcrae3 statcvs-0.2.2]$ java -jar statcvs.jar
> > > /tmp/statcvs-0.2.2/cvs-log.txt /home/tioadmin/cvs/tcdriver
> > > StatCVS - CVS statistics generation
> > >
> > > null
> > > [tioadmin@jmcrae3 statcvs-0.2.2]$
> >
> > I don't know if it matters or not, but which version of Java are  
> you using?
> > (java -version)
>
> Java is
>
> java version "1.4.2"
> Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.2)
> Classic VM (build 1.4.2, J2RE 1.4.2 IBM build cxia321420-20040626  
> (JIT enabled: jitc))
>
> >
> > It might also be worthwhile attaching the log file that causes the
> > problem, if
> > that's not a drama. :)
>
> Unfortunately I cannot post the entire log file or I'm sure to  
> cause heart attacks in my company.  Maybe someone else could attach  
> a sample log file here and then I will try it out to see if the  
> problem is with my log file or my setup.
>
> >
> > > When I run with debug mode on, here are the last few lines  
> before it
> > > stops.
> > >
> > > <lots of stuff cut out>
> > > FINE statcvs.input.FileBuilder <init>(): logging
> > > db2/workflow/DB2_Product_Check_Status.wkf
> > > FINE statcvs.input.FileBuilder <init>(): logging
> > > db2/workflow/DB2_Product_Install.wkf
> > > FINE statcvs.input.FileBuilder <init>(): logging
> > > db2/workflow/DB2_Product_Uninstall.wkf
> > > null
> > > [tioadmin@jmcrae3 statcvs-0.2.2]$
> > >
> > > Any ideas what is wrong?
> > > My log is successfully parsed by cvstat (another tool) but I  
> want to use
> > > cvsstat because it shows more information.
> > > My machine is Redhat EL 3.
> > >
> > > Thanks
> > >
> > > Jeff
> >
> > Regards,
> > James
> >
> > --
> > James Ring
> > [attachment "att9369r.dat" deleted by Jeff McRae/Toronto/IBM]
>
> Jeff

Re: [Statcvs-users] statcvs output is just "null"

[Jeff M. <jm...@ca...>]

sta...@li... wrote on 2005-11-25 05:42:36 PM:

> Hi Jeff,
> 
> On Saturday 26 November 2005 01:58, Jeff McRae wrote:
> > I tried to run cvsstat on my cvs log and I get
> >
> > [tioadmin@jmcrae3 statcvs-0.2.2]$ java -jar statcvs.jar
> > /tmp/statcvs-0.2.2/cvs-log.txt /home/tioadmin/cvs/tcdriver
> > StatCVS - CVS statistics generation
> >
> > null
> > [tioadmin@jmcrae3 statcvs-0.2.2]$
> 
> I don't know if it matters or not, but which version of Java are you 
using? 
> (java -version)

Java is

java version "1.4.2"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.2)
Classic VM (build 1.4.2, J2RE 1.4.2 IBM build cxia321420-20040626 (JIT 
enabled: jitc))

> 
> It might also be worthwhile attaching the log file that causes the 
> problem, if 
> that's not a drama. :)

Unfortunately I cannot post the entire log file or I'm sure to cause heart 
attacks in my company.  Maybe someone else could attach a sample log file 
here and then I will try it out to see if the problem is with my log file 
or my setup.

> 
> > When I run with debug mode on, here are the last few lines before it
> > stops.
> >
> > <lots of stuff cut out>
> > FINE statcvs.input.FileBuilder <init>(): logging
> > db2/workflow/DB2_Product_Check_Status.wkf
> > FINE statcvs.input.FileBuilder <init>(): logging
> > db2/workflow/DB2_Product_Install.wkf
> > FINE statcvs.input.FileBuilder <init>(): logging
> > db2/workflow/DB2_Product_Uninstall.wkf
> > null
> > [tioadmin@jmcrae3 statcvs-0.2.2]$
> >
> > Any ideas what is wrong?
> > My log is successfully parsed by cvstat (another tool) but I want to 
use
> > cvsstat because it shows more information.
> > My machine is Redhat EL 3.
> >
> > Thanks
> >
> > Jeff
> 
> Regards,
> James
> 
> -- 
> James Ring
> [attachment "att9369r.dat" deleted by Jeff McRae/Toronto/IBM] 

Jeff

[Statcvs-users] [PATCH] Make the author page display real names (if available)

[James R. <sj...@jd...>]

Hi all,

This patch makes the author page display real names (if available). You can 
see an example of the output generated by this patch at

http://xtuml.jdns.org/statcvs/user_u3957053.html

Please note that only the page title is changed. No other output is affected.

I'll also post this on the SourceForge patch tracker.

Regards,
James
-- 
James Ring

[Statcvs-users] [PATCH] Use new author(String,String) constructor

[James R. <sj...@jd...>]

Hi all,

This patch makes the input/Builder class use the mappings obtained from the 
passwd file specified by the -passwd option. Authors are now created with the 
two-arg (String, String) constructor.

It is possible for the second parameter to be null in the case where no 
mapping exists for the given user or no -passwd option is specified.

I'll also put this into the patch tracker on Sourceforge.

James
-- 
James Ring

Re: [PATCH] Re: [Statcvs-users] Add -passwd command line option

[James R. <sj...@jd...>]

Hi Steffen,

On Saturday 26 November 2005 12:00, Steffen Pingel wrote:
> Hi,
>
> please post the patch to the SourceForge tracker:
>
> http://sourceforge.net/tracker/?group_id=3D57558&atid=3D484571

Done, thanks for the link. The patch is request ID 1367210.

> Psersonally I often run StatCvs on a different machine than the one that
> hosts the repository and has the passwd file with all login names.
> Nevertheless I think the patch might be useful to a few people.
>
> StatCvs-XML has a similiar switch that is used to parse real names from t=
he
> maven project.xml file, therefore I think we may apply this patch as well.
>
> Richard, what is your opinon on the patch?
>
> Steffen

Regards,
James

=2D-=20
James Ring

Re: [PATCH] Re: [Statcvs-users] Add -passwd command line option

[Steffen P. <ste...@gm...>]

Hi,

please post the patch to the SourceForge tracker:

http://sourceforge.net/tracker/?group_id=57558&atid=484571

Psersonally I often run StatCvs on a different machine than the one that hosts 
the repository and has the passwd file with all login names. Nevertheless I 
think the patch might be useful to a few people. 

StatCvs-XML has a similiar switch that is used to parse real names from the 
maven project.xml file, therefore I think we may apply this patch as well.

Richard, what is your opinon on the patch? 

Steffen

-- 
Steffen Pingel - ste...@gm... - http://steffenpingel.de

[PATCH] Re: [Statcvs-users] Add -passwd command line option

[James R. <sj...@jd...>]

Hi all,

This patch actually makes the command line parser recognise the -passwd 
option. Sorry, I left this out of the previous patch.

Please apply. :)

Regards,
James
-- 
James Ring

[Statcvs-users] [PATCH] Add -passwd command line option

[James R. <sj...@jd...>]

Hi all,

I used StatCVS on a repository with 12 active authors, and I found it useful 
to be able to see their names alongside their usernames.

The attached patch adds a -passwd <path> option to StatCVS, which allows you 
to specify the location of a Unix-style passwd file containing username->real 
name mappings. Users of cvs2cl will already be familiar with a similar 
feature in that utility.

So, if you have a file that contains

        sjr:x:blah:blah:James Ring:blah
        jmd:x:blah:blah:Jon M. Doe:blah
        abc:x:blah:blah:Anne B. Collins:blah

the -passwd option with this file will map the username sjr to "James Ring". 
This patch doesn't actually change the HTML or graph output of StatCVS. I'll 
send that patch in a follow-up e-mail.

This should apply cleanly to the latest CVS (which I think is still the same 
as 0.2.2). Here's the diffstat:

Main.java                        |   13 ++++++
input/Builder.java               |   71 +++++++++++++++++++++++++++++++++++++
model/Author.java                |   20 ++++++++++
output/ConfigurationOptions.java |   21 ++++++++++
4 files changed, 123 insertions(+), 2 deletions(-)

Please try it out!

Regards,
James
-- 
James Ring

Re: [Statcvs-users] statcvs output is just "null"

[James R. <sj...@jd...>]

Hi Jeff,

On Saturday 26 November 2005 01:58, Jeff McRae wrote:
> I tried to run cvsstat on my cvs log and I get
>
> [tioadmin@jmcrae3 statcvs-0.2.2]$ java -jar statcvs.jar
> /tmp/statcvs-0.2.2/cvs-log.txt /home/tioadmin/cvs/tcdriver
> StatCVS - CVS statistics generation
>
> null
> [tioadmin@jmcrae3 statcvs-0.2.2]$

I don't know if it matters or not, but which version of Java are you using?=
=20
(java -version)

It might also be worthwhile attaching the log file that causes the problem,=
 if=20
that's not a drama. :)

> When I run with debug mode on, here are the last few lines before it
> stops.
>
> <lots of stuff cut out>
> FINE statcvs.input.FileBuilder <init>(): logging
> db2/workflow/DB2_Product_Check_Status.wkf
> FINE statcvs.input.FileBuilder <init>(): logging
> db2/workflow/DB2_Product_Install.wkf
> FINE statcvs.input.FileBuilder <init>(): logging
> db2/workflow/DB2_Product_Uninstall.wkf
> null
> [tioadmin@jmcrae3 statcvs-0.2.2]$
>
> Any ideas what is wrong?
> My log is successfully parsed by cvstat (another tool) but I want to use
> cvsstat because it shows more information.
> My machine is Redhat EL 3.
>
> Thanks
>
> Jeff

Regards,
James

=2D-=20
James Ring

[Statcvs-users] statcvs output is just "null"

[Jeff M. <jm...@ca...>]

I tried to run cvsstat on my cvs log and I get

[tioadmin@jmcrae3 statcvs-0.2.2]$ java -jar statcvs.jar 
/tmp/statcvs-0.2.2/cvs-log.txt /home/tioadmin/cvs/tcdriver
StatCVS - CVS statistics generation
 
null
[tioadmin@jmcrae3 statcvs-0.2.2]$

When I run with debug mode on, here are the last few lines before it 
stops.

<lots of stuff cut out>
FINE statcvs.input.FileBuilder <init>(): logging 
db2/workflow/DB2_Product_Check_Status.wkf
FINE statcvs.input.FileBuilder <init>(): logging 
db2/workflow/DB2_Product_Install.wkf
FINE statcvs.input.FileBuilder <init>(): logging 
db2/workflow/DB2_Product_Uninstall.wkf
null
[tioadmin@jmcrae3 statcvs-0.2.2]$

Any ideas what is wrong?
My log is successfully parsed by cvstat (another tool) but I want to use 
cvsstat because it shows more information. 
My machine is Redhat EL 3.

Thanks

Jeff

[Statcvs-users] Jcvsweb integration

[Eric M. <erm...@gm...>]

Here is the source for the class I wrote to integrate with Jcvsweb
(http://www.jcvs.org/jcvsweb/) if anyone is interested.

/**
 *
 */
package net.sf.statcvs.output;

import java.util.HashSet;
import java.util.Set;

import net.sf.statcvs.model.CvsFile;
import net.sf.statcvs.model.CvsRevision;
import net.sf.statcvs.model.Directory;

/**
 * @author ERMEANEY
 *
 */
public class JcvswebIntegration implements WebRepositoryIntegration {
    private String baseURL;
    private Set atticFileNames =3D new HashSet();
    /**
     *
     */
    public JcvswebIntegration(String baseURL) {
        if (baseURL.endsWith("/")) {
            this.baseURL =3D baseURL.substring(0, baseURL.length() - 1);
        } else {
            this.baseURL =3D baseURL;
        }
    }

    /* (non-Javadoc)
     * @see net.sf.statcvs.output.WebRepositoryIntegration#getName()
     */
    public String getName() {
        return "JcvsWeb";
    }

    /* (non-Javadoc)
     * @see net.sf.statcvs.output.WebRepositoryIntegration#getDirectoryUrl(=
net.sf.statcvs.model.Directory)
     */
    public String getDirectoryUrl(Directory directory) {
        String path =3D baseURL + "/HEAD/list/" + directory.getPath();
        return path;
    }

    /* (non-Javadoc)
     * @see net.sf.statcvs.output.WebRepositoryIntegration#getFileHistoryUr=
l(net.sf.statcvs.model.CvsFile)
     */
    public String getFileHistoryUrl(CvsFile file) {
        return baseURL + "/HEAD/vers/" + file.getFilenameWithPath();
    }

    /* (non-Javadoc)
     * @see net.sf.statcvs.output.WebRepositoryIntegration#getFileViewUrl(n=
et.sf.statcvs.model.CvsFile)
     */
    public String getFileViewUrl(CvsFile file) {
        return baseURL + "/HEAD/view/" + file.getFilenameWithPath() +
"/" + file.getLatestRevision().getRevisionNumber();
    }

    /* (non-Javadoc)
     * @see net.sf.statcvs.output.WebRepositoryIntegration#getFileViewUrl(n=
et.sf.statcvs.model.CvsRevision)
     */
    public String getFileViewUrl(CvsRevision revision) {
        return baseURL + "/HEAD/view/" +
revision.getFile().getFilenameWithPath() + "/" +
revision.getRevisionNumber();
    }

    /* (non-Javadoc)
     * @see net.sf.statcvs.output.WebRepositoryIntegration#getDiffUrl(net.s=
f.statcvs.model.CvsRevision,
net.sf.statcvs.model.CvsRevision)
     */
    public String getDiffUrl(CvsRevision oldRevision, CvsRevision newRevisi=
on) {
        return baseURL + "/HEAD/pdiff/" +
newRevision.getFile().getFilenameWithPath() + "/" +
oldRevision.getRevisionNumber() + "/" +
newRevision.getRevisionNumber();
    }

    /* (non-Javadoc)
     * @see net.sf.statcvs.output.WebRepositoryIntegration#setAtticFileName=
s(java.util.Set)
     */
    public void setAtticFileNames(Set atticFileNames) {
        this.atticFileNames =3D atticFileNames;
    }

}

--
Eric Meaney
erm...@gm...

Re: [Statcvs-users] Paper on performance indicators in CVS repositories

[Richard C. <ri...@cy...>]

James,

On 18 Oct 2005, at 07:07, James Ring wrote:
> In the case of my final year university group project, the  
> lecturers are
> explicitly basing some of our assessment on the number of lines  
> contributed
> to CVS. I think that this is a sorry situation, given how poorly a  
> simple
> line count can represent somebody's input to a software engineering  
> project.

True. In the words of Bill Gates: Measuring programming progress by  
lines of code is like measuring aircraft building progress by weight.  
(I love this quote. You teachers would probably hate it ;-)


>> Link and some commentary on my weblog:
>> http://dowhatimean.net/2005/10/cvs-and-performance
>
> You mention that the lack of correlation between code metrics and  
> access
> patterns does not bode well for StatCVS. I don't think that it's
> necessarily a bad thing... Finding correlations between student grades
> and CVS activity is only one possible application
> for a good tool such as StatCVS.
>
> For example, in an industry project, the project manager may use a  
> whole
> bunch of statistics from various sources (including, but not  
> limited to,
> StatCVS) to evaluate the health and progress of their project.
>
> I guess what I'm trying to say is that by not providing more  
> metrics, we'll
> never see if any interesting correlations come up in unexpected  
> places.
>
> I'd certainly like to see some sort of Java code metrics as an "add- 
> on"
> for StatCVS, like the ability to measure (for example) method fan-in
> and fan-out in a Java project over time.

That would certainly be interesting, but there are two issues: 1)  
It's programming language dependent, which limits the target  
audience. 2) I can't see how to present this in an end user  
application. If we have a graph of method fan-in over time, I can see  
everybody going: "OK, so method fan-in has risen 30% over the last  
half year. Now is this good or bad?"

Richard


>
> The more stats, the merrier!
>
>
>> Best,
>> Richard
>>
>>
>
> Thanks,
> James
> -- 
> James Ring
>
>
>

Re: [Statcvs-users] Paper on performance indicators in CVS repositories

[James R. <sj...@jd...>]

Hi Richard,

> Hi all,
>
> I came across an interesting paper on using information from CVS
> repositories to predict student performance. Summary: they found no
> good performance indicators. The best one was plain ol' lines of code.

In the case of my final year university group project, the lecturers are
explicitly basing some of our assessment on the number of lines contributed
to CVS. I think that this is a sorry situation, given how poorly a simple
line count can represent somebody's input to a software engineering project.

> Link and some commentary on my weblog:
> http://dowhatimean.net/2005/10/cvs-and-performance

You mention that the lack of correlation between code metrics and access
patterns does not bode well for StatCVS. I don't think that it's
necessarily a bad thing... Finding correlations between student grades
and CVS activity is only one possible application
for a good tool such as StatCVS.

For example, in an industry project, the project manager may use a whole
bunch of statistics from various sources (including, but not limited to,
StatCVS) to evaluate the health and progress of their project.

I guess what I'm trying to say is that by not providing more metrics, we'll
never see if any interesting correlations come up in unexpected places.

I'd certainly like to see some sort of Java code metrics as an "add-on"
for StatCVS, like the ability to measure (for example) method fan-in
and fan-out in a Java project over time.

The more stats, the merrier!

> Best,
> Richard
>

Thanks,
James
-- 
James Ring

[Statcvs-users] Paper on performance indicators in CVS repositories

[Richard C. <ri...@cy...>]

Hi all,

I came across an interesting paper on using information from CVS  
repositories to predict student performance. Summary: they found no  
good performance indicators. The best one was plain ol' lines of code.

Link and some commentary on my weblog:
http://dowhatimean.net/2005/10/cvs-and-performance

Best,
Richard

[Statcvs-users] $BG'>Z$5$l$^$7$?$N$G!D(B

[<in...@nh...>]

$B"#<+8J>R2p"#(B
$B!!Ev%5%$%H$OIaCJ$N;d@83h$dIW$KITK~$rJz$($F$k1|MM$dFH?H=w@-$N(B
$B=8$^$j$G$9!#=w@-2q0w$N;q6b$r%a%$%s$G1?1D$7$F$*$j$^$9!#(B
$B!!$3$NEY!"5.J}$N%"%+%&%s%H(B($BG[?.$5$l$?%"%I%l%9(B)$B$OEv2q0w@)%/%i(B
$B%V$N!LFCJL=`2q0w!M$HG'Dj$5$l$^$7$?$N$G!"%a!<%k$G$*CN$i$;$5$;(B
$B$FD:$-$^$7$?!#(B
$B!!!!!!!!""4JC1$JL5NAEPO?$HF1;~$KFCJL2q0w$X0\9T""(B
$B!!!!!!!!!!(Bhttp://www.flameblow.net?num=200
$B!LFCJL2q0w!M$NJ}$X2<5-FCE5$rMQ0U$7$F$*$j$^$9$N$G!"@'Hs$4MxMQ(B
$B$7$F$_$F2<$5$$!#(B

$B!z(B1,000$B1_J,L5NA%]%$%s%H(B
$BG'>ZITMW$G!Z(B1,000$B1_![J,$NL5NA%]%$%s%H$r?JDhCW$7$^$9!#HV9fG'>Z8e$5$i$K(B12000$BJ,L5NA%]%$%s%H$4MxMQ$$$?$@$1$^$9(B!$BL5NA%]%$%s%H$G%a!<%k$N$d$j$H$j!"%W%m%U%#!<%k$N1\Mw!"<L??$N3NG'$J$I(B
$BA4$F$N5!G=$,;H$($^$9$N$G!"$4<+M3$K$4MxMQ2<$5$$!#(B

$B!z!L"(!MD>%a!<%kAw!"<u?.2DG=(B
$BCO0h=w@-$N4uK>$h$j!":G?7EPO?>pJs$r<L%aE:IU$GG[?.$N>e!"D>%a$N(B
$B>R2p$r9T$$$^$9!#4uK>0[@-$+$iD>@\5.J}$N08@h$K%a!<%k$,F~$C$F$/(B
$B$k2DG=@-$b9b$$$N$G!"9%$_$N=w@->pJs$r?o;~%A%'%C%/$7$^$7$g(B
$B$&!*!*!*(B

$BFCJL2q0w$H$7$F$N$4F~>l$r$*BT$A$7$F$*$j$^$9!#(B

$B!JFCJL2q0w0\9T(Bhttp://www.flameblow.net?num=200)










$BG[?.ITMW$NJ}$O:!J}(B
$B!Z(BReception refusal address$B![(B
no...@he...