Re: [ALL] Commons Parent reports

2012-03-14 Thread Gilles Sadowski
On Tue, Mar 13, 2012 at 01:52:32PM -0400, Gary Gregory wrote:
 On Mar 13, 2012, at 12:40, Gilles Sadowski gil...@harfang.homelinux.org 
 wrote:
 
 
  [...]
 
  The tools are there, but you have to tell people that they _must_ use 
  them.
 
  Commons has already enough rules and process. As long as the releases
  are have clean code I wouldn't be too anal about the commits in
  between.
 
  I think that the main disagreement is here. Source code must be a clear 
  read
  for the _developers_. To put it bluntly, I don't care that the releases 
  have
  cleanly formatted code, as almost nobody is going to read those packaged
  sources!
 
 Au contraire, most users will /only/ look at the source jar because
 that is what you use in the debugger!

Here IMHO you are talking about _developers_.
By users I meant people who include the classes JAR in their classpath.
They will at most look at the API docs.


Gilles

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-14 Thread Gilles Sadowski
On Tue, Mar 13, 2012 at 01:59:25PM -0400, Gary Gregory wrote:
 On Mar 13, 2012, at 12:40, Gilles Sadowski gil...@harfang.homelinux.org 
 wrote:
 
  Hi.
 
 
  [...]
 
  The tools are there, but you have to tell people that they _must_ use 
  them.
 
  Commons has already enough rules and process. As long as the releases
  are have clean code I wouldn't be too anal about the commits in
  between.
 
  I think that the main disagreement is here. Source code must be a clear 
  read
  for the _developers_. To put it bluntly, I don't care that the releases 
  have
  cleanly formatted code, as almost nobody is going to read those packaged
  sources!
 
 And another thing: the formatting /is/ important in released sources
 because, again, this is what most users will see in their debuggers.
 Have you seen some of the JRE sources? Some files are a mess, others
 have blank lines in the middle of headers. Others look like they were
 entered by a prisoner blinded in the noon day sun after spending a
 month in the hole with bread and water ration and then given a stick
 of butter for lunch.

It sounds like I were arguing against well formatted sources. My point was
that sources must _always_ be well formatted, and not _only_ at release.


Gilles

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-14 Thread Gilles Sadowski
On Tue, Mar 13, 2012 at 06:37:21PM +0100, Emmanuel Bourg wrote:
 Le 13/03/2012 17:52, Gilles Sadowski a écrit :
 
 What about the Useful for the developer category?
 
 They are useful at release time only, then they become quickly
 outdated as the code evolves after the release.
 
 If I want to help Commons Math, I'll checkout the code and build the
 reports. Then I might inspect the Findbugs report and see if there
 is something to fix. But I'll never go to the website and browse
 months old reports.
 
 The point is, these reports are valuable if they are updated
 continuously, but that's not possible to do that, because if the
 documentation is updated, the site will expose information that is
 not applicable to the latest release, but to the next one.
 
 What I'd like to see is a more user oriented site, something clear
 and simple, and a developer oriented site with all the reports you
 want, automatically updated everyday.

+1
[All my comments were about the developer site (by which I mean the HTML
reports generated by mvn site) and not about the official web site.]


Gilles

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



RE: [ALL] Commons Parent reports

2012-03-14 Thread Honton, Charles
As a user of commons libraries, I look at several reports.  I'll consult 
Javadoc first.  In the case the javadoc is silent about behavior, I look at 
JXR.  When I am debugging my code's interaction with the commons library, I 
want to have the sources available to my IDE, so that I can step through the 
commons library.

chas

-Original Message-
From: Gilles Sadowski [mailto:gil...@harfang.homelinux.org] 
Sent: Wednesday, March 14, 2012 1:51 AM
To: dev@commons.apache.org
Subject: Re: [ALL] Commons Parent reports

On Tue, Mar 13, 2012 at 01:52:32PM -0400, Gary Gregory wrote:
 On Mar 13, 2012, at 12:40, Gilles Sadowski gil...@harfang.homelinux.org 
 wrote:
 
 
  [...]
 
  The tools are there, but you have to tell people that they _must_ use 
  them.
 
  Commons has already enough rules and process. As long as the releases
  are have clean code I wouldn't be too anal about the commits in
  between.
 
  I think that the main disagreement is here. Source code must be a clear 
  read
  for the _developers_. To put it bluntly, I don't care that the releases 
  have
  cleanly formatted code, as almost nobody is going to read those packaged
  sources!
 
 Au contraire, most users will /only/ look at the source jar because
 that is what you use in the debugger!

Here IMHO you are talking about _developers_.
By users I meant people who include the classes JAR in their classpath.
They will at most look at the API docs.


Gilles

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org


-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



RE: [ALL] Commons Parent reports

2012-03-14 Thread Honton, Charles
Why should the developer site be any different than the release site?

-Original Message-
From: Gilles Sadowski [mailto:gil...@harfang.homelinux.org] 
Sent: Wednesday, March 14, 2012 2:18 AM
To: dev@commons.apache.org
Subject: Re: [ALL] Commons Parent reports

On Tue, Mar 13, 2012 at 06:37:21PM +0100, Emmanuel Bourg wrote:
 Le 13/03/2012 17:52, Gilles Sadowski a écrit :
 
 What about the Useful for the developer category?
 
 They are useful at release time only, then they become quickly
 outdated as the code evolves after the release.
 
 If I want to help Commons Math, I'll checkout the code and build the
 reports. Then I might inspect the Findbugs report and see if there
 is something to fix. But I'll never go to the website and browse
 months old reports.
 
 The point is, these reports are valuable if they are updated
 continuously, but that's not possible to do that, because if the
 documentation is updated, the site will expose information that is
 not applicable to the latest release, but to the next one.
 
 What I'd like to see is a more user oriented site, something clear
 and simple, and a developer oriented site with all the reports you
 want, automatically updated everyday.

+1
[All my comments were about the developer site (by which I mean the HTML
reports generated by mvn site) and not about the official web site.]


Gilles

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org


-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-14 Thread sebb
On 14 March 2012 16:27, Honton, Charles charles_hon...@intuit.com wrote:
 As a user of commons libraries, I look at several reports.  I'll consult 
 Javadoc first.  In the case the javadoc is silent about behavior, I look at 
 JXR.

+1, JXR has been very useful to me.

 When I am debugging my code's interaction with the commons library, I want to 
 have the sources available to my IDE, so that I can step through the commons 
 library.
 chas

 -Original Message-
 From: Gilles Sadowski [mailto:gil...@harfang.homelinux.org]
 Sent: Wednesday, March 14, 2012 1:51 AM
 To: dev@commons.apache.org
 Subject: Re: [ALL] Commons Parent reports

 On Tue, Mar 13, 2012 at 01:52:32PM -0400, Gary Gregory wrote:
 On Mar 13, 2012, at 12:40, Gilles Sadowski gil...@harfang.homelinux.org 
 wrote:

 
  [...]
 
  The tools are there, but you have to tell people that they _must_ use 
  them.
 
  Commons has already enough rules and process. As long as the releases
  are have clean code I wouldn't be too anal about the commits in
  between.
 
  I think that the main disagreement is here. Source code must be a clear 
  read
  for the _developers_. To put it bluntly, I don't care that the releases 
  have
  cleanly formatted code, as almost nobody is going to read those packaged
  sources!

 Au contraire, most users will /only/ look at the source jar because
 that is what you use in the debugger!

 Here IMHO you are talking about _developers_.
 By users I meant people who include the classes JAR in their classpath.
 They will at most look at the API docs.


 Gilles

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org


 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org


-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-14 Thread sebb
On 13 March 2012 11:48, sebb seb...@gmail.com wrote:
 Commons Parent 24 includes the following reports:

 Javadoc
 Jxr
 Surefire
 RAT
 Cobertura
 Clirr
 JDepend

 I think the following should be added:

 Changes/JIRA

Done.

 The following could be added:

 Findbugs
 Checkstyle

 Any others?

There's obviously a lot of argument about CheckStyle (and probably PMD
too) because the report can be configured to check different things,
and not everyone agrees on what to check. Let's leave that for a
separate discussion (in another thread).

The other argument is about whether to include certain reports that
are perhaps more developer-oriented.

My view is that if a report is likely to be useful to a 3rd party
developer, we should include it.
Yes, it may make the report list a bit longer, but that is a minor
issue - and maybe one day Maven will give more control over the
layout.

Commons components tend to be fairly low-level, and are generally used
by developers in building larger applications, so it seems natural for
the developer-style reports to be available. As someone already
mentioned, reports such as Cobertura and Findbugs can help show that
we take testing and code quality seriously.

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Simone Tripodi
Hi!

 Commons Parent 24 includes the following reports:

 Javadoc
 Jxr
 Surefire
 RAT
 Cobertura
 Clirr
 JDepend

 I think the following should be added:

 Changes/JIRA

+1 the pattern is always the same


 The following could be added:

 Findbugs
 Checkstyle

+1


 Any others?

PMD/CPD would be fine as well, then I think it would be complete!
best,
-Simo

http://people.apache.org/~simonetripodi/
http://simonetripodi.livejournal.com/
http://twitter.com/simonetripodi
http://www.99soft.org/

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Torsten Curdt
 The following could be added:

 Findbugs
 Checkstyle

+1 for findbugs
-1 for checkstyle

cheers,
Torsten

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Simone Tripodi
Hi Torsten!

 -1 for checkstyle

With my +1 I meant that, as we discussed in another thread, the parent
could provide a default - but overridable - configuration; I think
that having at least one metric of code style measure in each
component would be nice to have, so unless other preferences, the
parent suggests a default config

I would like to understand better your PoV (that would influence
mine): which are your concerns about having the checkstyle?

many thanks in advance, all the best,
-Simo

http://people.apache.org/~simonetripodi/
http://simonetripodi.livejournal.com/
http://twitter.com/simonetripodi
http://www.99soft.org/

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Gary Gregory
On Tue, Mar 13, 2012 at 8:14 AM, Simone Tripodi simonetrip...@apache.orgwrote:

 Hi Torsten!

  -1 for checkstyle

 With my +1 I meant that, as we discussed in another thread, the parent
 could provide a default - but overridable - configuration; I think
 that having at least one metric of code style measure in each
 component would be nice to have, so unless other preferences, the
 parent suggests a default config


+1. We need a default, Commons components can continue to override or
decide to use the default. New components coming should use the default.

Gary


 I would like to understand better your PoV (that would influence
 mine): which are your concerns about having the checkstyle?

 many thanks in advance, all the best,
 -Simo

 http://people.apache.org/~simonetripodi/
 http://simonetripodi.livejournal.com/
 http://twitter.com/simonetripodi
 http://www.99soft.org/

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org




-- 
E-Mail: garydgreg...@gmail.com | ggreg...@apache.org
JUnit in Action, 2nd Ed: http://goog_1249600977http://bit.ly/ECvg0
Spring Batch in Action: http://s.apache.org/HOqhttp://bit.ly/bqpbCK
Blog: http://garygregory.wordpress.com
Home: http://garygregory.com/
Tweet! http://twitter.com/GaryGregory


Re: [ALL] Commons Parent reports

2012-03-13 Thread Torsten Curdt
I find checkstyle to be not very useful. It's more hassle than it's
worth. It's like pointing fingers instead of helping. If you want to
foster a certain code style provide eclipse and intellij formatter
settings instead - that's actually helping. Especially if you run them
before a release.

The basic code style is like logging - people spent just wait too much
time on this. Thinks we really should care about are in the findbugs
and PMD report. I don't see why we should make checkstyle part of the
projects by default.

My 2 cents,
Torsten

On Tue, Mar 13, 2012 at 1:14 PM, Simone Tripodi
simonetrip...@apache.org wrote:
 Hi Torsten!

 -1 for checkstyle

 With my +1 I meant that, as we discussed in another thread, the parent
 could provide a default - but overridable - configuration; I think
 that having at least one metric of code style measure in each
 component would be nice to have, so unless other preferences, the
 parent suggests a default config

 I would like to understand better your PoV (that would influence
 mine): which are your concerns about having the checkstyle?

 many thanks in advance, all the best,
 -Simo

 http://people.apache.org/~simonetripodi/
 http://simonetripodi.livejournal.com/
 http://twitter.com/simonetripodi
 http://www.99soft.org/

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org


-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Gilles Sadowski
On Tue, Mar 13, 2012 at 01:39:40PM +0100, Torsten Curdt wrote:
 I find checkstyle to be not very useful. It's more hassle than it's
 worth. It's like pointing fingers instead of helping. If you want to
 foster a certain code style provide eclipse and intellij formatter
 settings instead - that's actually helping. Especially if you run them
 before a release.

CheckStyle reports should be checked regularly. Only doing so just before a
release indeed leads to a lot of tedious work, because coders did not
respect the basic, agreed on, style.

I don't use an IDE, so for me, CheckStyle helps but formatter IDE plugins
would not. Our mileage do vary but the end-product (clean code) should not.  

As said in another post, you can always disable reports that you find
unhelpful.

 The basic code style is like logging - people spent just wait too much
 time on this.

The real problem is that some coders do not do their part of the job when
they commit badly formatted code.
Those whose spend too much time are the ones who try to clean up the mess
afterwards.
CheckStyle indeed points a finger to the right person, which IMHO helps by
making this person aware that he should fix it.

 Thinks we really should care about are in the findbugs
 and PMD report. I don't see why we should make checkstyle part of the
 projects by default.

They are all useful in different ways.

Best regards,
Gilles

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Gary Gregory
On Tue, Mar 13, 2012 at 8:39 AM, Torsten Curdt tcu...@vafer.org wrote:

 I find checkstyle to be not very useful. It's more hassle than it's
 worth. It's like pointing fingers instead of helping. If you want to
 foster a certain code style provide eclipse and intellij formatter
 settings instead - that's actually helping. Especially if you run them
 before a release.


I'd /love/ to have IDE settings for formatting saved in a project. This is
the 21st century, all IDEs support this, if you do not use an IDE (hi
Gilles), then, well, you probably also like driving a stick for control
:) The only tricky part is how organize such a folder to account for
different IDEs and versions. For example ide/eclipse/3.7.1,
ide/intellij/10.5.3, and so on. Then you can move the IDE files to where
each IDE wants it.

Gary


 The basic code style is like logging - people spent just wait too much
 time on this. Thinks we really should care about are in the findbugs
 and PMD report. I don't see why we should make checkstyle part of the
 projects by default.


 My 2 cents,
 Torsten

 On Tue, Mar 13, 2012 at 1:14 PM, Simone Tripodi
 simonetrip...@apache.org wrote:
  Hi Torsten!
 
  -1 for checkstyle
 
  With my +1 I meant that, as we discussed in another thread, the parent
  could provide a default - but overridable - configuration; I think
  that having at least one metric of code style measure in each
  component would be nice to have, so unless other preferences, the
  parent suggests a default config
 
  I would like to understand better your PoV (that would influence
  mine): which are your concerns about having the checkstyle?
 
  many thanks in advance, all the best,
  -Simo
 
  http://people.apache.org/~simonetripodi/
  http://simonetripodi.livejournal.com/
  http://twitter.com/simonetripodi
  http://www.99soft.org/
 
  -
  To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
  For additional commands, e-mail: dev-h...@commons.apache.org
 

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org




-- 
E-Mail: garydgreg...@gmail.com | ggreg...@apache.org
JUnit in Action, 2nd Ed: http://goog_1249600977http://bit.ly/ECvg0
Spring Batch in Action: http://s.apache.org/HOqhttp://bit.ly/bqpbCK
Blog: http://garygregory.wordpress.com
Home: http://garygregory.com/
Tweet! http://twitter.com/GaryGregory


Re: [ALL] Commons Parent reports

2012-03-13 Thread Torsten Curdt
 CheckStyle reports should be checked regularly. Only doing so just before a
 release indeed leads to a lot of tedious work, because coders did not
 respect the basic, agreed on, style.

I guess we are disagreeing here.

 I don't use an IDE, so for me, CheckStyle helps but formatter IDE plugins
 would not. Our mileage do vary but the end-product (clean code) should not.

Then you are probably a vocal minority here.
As long as there is someone that can run a code formatter before a
release that does not matter though.

 As said in another post, you can always disable reports that you find
 unhelpful.

Fair enough. But projects that find it useful could also just add the report :)
We are discussing about what should be the default here.
That said I rather just disable it in the POM that continue with the discussion.

 The basic code style is like logging - people spent just wait too much
 time on this.

...because I dont' want to contribute to that time any more.

 The real problem is that some coders do not do their part of the job when
 they commit badly formatted code.
 Those whose spend too much time are the ones who try to clean up the mess
 afterwards.

If you prefer to not use code formatter - that is. But that's your decision.

 Menu  Format Source Code  Done.

...plus I am bet there are ways to set up code formatting for vim and
friends - if one wanted to.

 CheckStyle indeed points a finger to the right person, which IMHO helps by
 making this person aware that he should fix it.

And I say - better let's give people the tools and not just point at them.

cheers,
Torsten

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Benedikt Ritter
Am 13. März 2012 14:15 schrieb Gary Gregory garydgreg...@gmail.com:
 On Tue, Mar 13, 2012 at 8:39 AM, Torsten Curdt tcu...@vafer.org wrote:

 I find checkstyle to be not very useful. It's more hassle than it's
 worth. It's like pointing fingers instead of helping. If you want to
 foster a certain code style provide eclipse and intellij formatter
 settings instead - that's actually helping. Especially if you run them
 before a release.


 I'd /love/ to have IDE settings for formatting saved in a project. This is
 the 21st century, all IDEs support this, if you do not use an IDE (hi
 Gilles), then, well, you probably also like driving a stick for control
 :) The only tricky part is how organize such a folder to account for
 different IDEs and versions. For example ide/eclipse/3.7.1,
 ide/intellij/10.5.3, and so on. Then you can move the IDE files to where
 each IDE wants it.


Very big +1 Gary! It would make contributing to the various components
so much easier (at least for people how use an IDE ;). I guess we
should discuss this on a new thread.

 Gary


 The basic code style is like logging - people spent just wait too much
 time on this. Thinks we really should care about are in the findbugs
 and PMD report. I don't see why we should make checkstyle part of the
 projects by default.


 My 2 cents,
 Torsten

 On Tue, Mar 13, 2012 at 1:14 PM, Simone Tripodi
 simonetrip...@apache.org wrote:
  Hi Torsten!
 
  -1 for checkstyle
 
  With my +1 I meant that, as we discussed in another thread, the parent
  could provide a default - but overridable - configuration; I think
  that having at least one metric of code style measure in each
  component would be nice to have, so unless other preferences, the
  parent suggests a default config
 
  I would like to understand better your PoV (that would influence
  mine): which are your concerns about having the checkstyle?
 
  many thanks in advance, all the best,
  -Simo
 
  http://people.apache.org/~simonetripodi/
  http://simonetripodi.livejournal.com/
  http://twitter.com/simonetripodi
  http://www.99soft.org/
 
  -
  To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
  For additional commands, e-mail: dev-h...@commons.apache.org
 

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org




 --
 E-Mail: garydgreg...@gmail.com | ggreg...@apache.org
 JUnit in Action, 2nd Ed: http://goog_1249600977http://bit.ly/ECvg0
 Spring Batch in Action: http://s.apache.org/HOqhttp://bit.ly/bqpbCK
 Blog: http://garygregory.wordpress.com
 Home: http://garygregory.com/
 Tweet! http://twitter.com/GaryGregory

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Gilles Sadowski
On Tue, Mar 13, 2012 at 02:27:32PM +0100, Torsten Curdt wrote:
  CheckStyle reports should be checked regularly. Only doing so just before a
  release indeed leads to a lot of tedious work, because coders did not
  respect the basic, agreed on, style.
 
 I guess we are disagreeing here.

I didn't think we disagreed on refusing to spend a lot of time of correcting
formatting mess.

 
  I don't use an IDE, so for me, CheckStyle helps but formatter IDE plugins
  would not. Our mileage do vary but the end-product (clean code) should not.
 
 Then you are probably a vocal minority here.

Because I don't use an IDE or because I like clean code?

 As long as there is someone that can run a code formatter before a
 release that does not matter though.

But would you be against running a formatter before committing the code.

Modern editors would do this as you type (like e.g. in emacs or the IDE
plugins referred to in this thread).
It's tiresome to have to read badly formatted code and it leads to spurios
commits just to straighten up the formatting.

  As said in another post, you can always disable reports that you find
  unhelpful.
 
 Fair enough. But projects that find it useful could also just add the report 
 :)
 We are discussing about what should be the default here.
 That said I rather just disable it in the POM that continue with the 
 discussion.
 
  The basic code style is like logging - people spent just wait too much
  time on this.
 
 ...because I dont' want to contribute to that time any more.
 
  The real problem is that some coders do not do their part of the job when
  they commit badly formatted code.
  Those whose spend too much time are the ones who try to clean up the mess
  afterwards.
 
 If you prefer to not use code formatter - that is. But that's your decision.

Did I say so?
I just said the opposite: I don't like that people commit badly formatted
code and...

  Menu  Format Source Code  Done.

... if they don't perform the above, CheckStyle is there to remind them they
forgot to do it.

 ...plus I am bet there are ways to set up code formatting for vim and
 friends - if one wanted to.

I hope that misunderstanding has been cleared now.

  CheckStyle indeed points a finger to the right person, which IMHO helps by
  making this person aware that he should fix it.
 
 And I say - better let's give people the tools and not just point at them.

The tools are there, but you have to tell people that they _must_ use them.

Regards,
Gilles

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Simone Tripodi
Hi again Torsten!

I thought it would be useful having a checkstyle also becasue it is
not just a matter of code formatting rules, there are also useful
basic checks http://checkstyle.sourceforge.net/checks.html that help
- me, at least - on taking care of some rules I can occasionally
forget to apply.

 And I say - better let's give people the tools and not just point at them.

I agree. IIUC the proposal of putting checkstyle in the parent came
out because it is plugged in every plugin, so the purpose is more
generalize rather then point to it.

All the best!
-Simo

http://people.apache.org/~simonetripodi/
http://simonetripodi.livejournal.com/
http://twitter.com/simonetripodi
http://www.99soft.org/



On Tue, Mar 13, 2012 at 2:27 PM, Torsten Curdt tcu...@vafer.org wrote:
 CheckStyle reports should be checked regularly. Only doing so just before a
 release indeed leads to a lot of tedious work, because coders did not
 respect the basic, agreed on, style.

 I guess we are disagreeing here.

 I don't use an IDE, so for me, CheckStyle helps but formatter IDE plugins
 would not. Our mileage do vary but the end-product (clean code) should not.

 Then you are probably a vocal minority here.
 As long as there is someone that can run a code formatter before a
 release that does not matter though.

 As said in another post, you can always disable reports that you find
 unhelpful.

 Fair enough. But projects that find it useful could also just add the report 
 :)
 We are discussing about what should be the default here.
 That said I rather just disable it in the POM that continue with the 
 discussion.

 The basic code style is like logging - people spent just wait too much
 time on this.

 ...because I dont' want to contribute to that time any more.

 The real problem is that some coders do not do their part of the job when
 they commit badly formatted code.
 Those whose spend too much time are the ones who try to clean up the mess
 afterwards.

 If you prefer to not use code formatter - that is. But that's your decision.

  Menu  Format Source Code  Done.

 ...plus I am bet there are ways to set up code formatting for vim and
 friends - if one wanted to.

 CheckStyle indeed points a finger to the right person, which IMHO helps by
 making this person aware that he should fix it.

 And I say - better let's give people the tools and not just point at them.

 cheers,
 Torsten

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org


-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Christian Grobmeier
+1 on the mentioned plugins, except:
-1 on checkstyle.

I dont see a benefit in making checkstyle default. If somebody wants
to use that tool, he can.

What checkstyle-style would be the default? Sun conventions or maven
style? I am for sun, Simone is for maven (i guess). Probably there is
another component with more specific desires. I really don't want to
discuss this issue...

We are all grown up here and a small party. And we have commit
notifications. And if a component wishes, it can use checkstyle. Just
don't make a specific style default... it probably makes a little bit
more sense if components would all use the same style. But we don't
even manage to agree on a uni-build system, I really see no way to
make it uni-styled.


On Tue, Mar 13, 2012 at 2:27 PM, Torsten Curdt tcu...@vafer.org wrote:
 CheckStyle reports should be checked regularly. Only doing so just before a
 release indeed leads to a lot of tedious work, because coders did not
 respect the basic, agreed on, style.

 I guess we are disagreeing here.

 I don't use an IDE, so for me, CheckStyle helps but formatter IDE plugins
 would not. Our mileage do vary but the end-product (clean code) should not.

 Then you are probably a vocal minority here.
 As long as there is someone that can run a code formatter before a
 release that does not matter though.

 As said in another post, you can always disable reports that you find
 unhelpful.

 Fair enough. But projects that find it useful could also just add the report 
 :)
 We are discussing about what should be the default here.
 That said I rather just disable it in the POM that continue with the 
 discussion.

 The basic code style is like logging - people spent just wait too much
 time on this.

 ...because I dont' want to contribute to that time any more.

 The real problem is that some coders do not do their part of the job when
 they commit badly formatted code.
 Those whose spend too much time are the ones who try to clean up the mess
 afterwards.

 If you prefer to not use code formatter - that is. But that's your decision.

  Menu  Format Source Code  Done.

 ...plus I am bet there are ways to set up code formatting for vim and
 friends - if one wanted to.

 CheckStyle indeed points a finger to the right person, which IMHO helps by
 making this person aware that he should fix it.

 And I say - better let's give people the tools and not just point at them.

 cheers,
 Torsten

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org




-- 
http://www.grobmeier.de
https://www.timeandbill.de

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Torsten Curdt
 I didn't think we disagreed on refusing to spend a lot of time of correcting 
 formatting mess.

Not on that one :)

I just argue that in the days of code formatters lot of time has
become is quite significant.

 Then you are probably a vocal minority here.

 Because I don't use an IDE or because I like clean code?

I hope that's a rhetorical question. I would think most java people use an IDE.

 As long as there is someone that can run a code formatter before a
 release that does not matter though.

 But would you be against running a formatter before committing the code.

It would be nice - but I would not want to require it.

  Menu  Format Source Code  Done.

 ... if they don't perform the above, CheckStyle is there to remind them they
 forgot to do it.

Why not just run the formatter in the very moment you are annoyed?

 And I say - better let's give people the tools and not just point at them.

 The tools are there, but you have to tell people that they _must_ use them.

Commons has already enough rules and process. As long as the releases
are have clean code I wouldn't be too anal about the commits in
between.

Summary: if you guys insist - add it ...as long as components are OK
to disable it.
I personally just don't find it useful. But I can live with the fact
that you disagree.

Now that was already 4 cents :)

cheers,
Torsten

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Gilles Sadowski
On Tue, Mar 13, 2012 at 03:08:01PM +0100, Christian Grobmeier wrote:
 +1 on the mentioned plugins, except:
 -1 on checkstyle.
 
 I dont see a benefit in making checkstyle default. If somebody wants
 to use that tool, he can.
 
 What checkstyle-style would be the default? Sun conventions or maven
 style? I am for sun, Simone is for maven (i guess). Probably there is
 another component with more specific desires. I really don't want to
 discuss this issue...
 
 We are all grown up here and a small party. And we have commit
 notifications. And if a component wishes, it can use checkstyle. Just
 don't make a specific style default... it probably makes a little bit
 more sense if components would all use the same style. But we don't
 even manage to agree on a uni-build system, I really see no way to
 make it uni-styled.

Unless I'm missing something:

  (Using CheckStyle) != (Using the same formatting style in all projects)


Gilles

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Ralph Goers

On Mar 13, 2012, at 6:44 AM, Benedikt Ritter wrote:

 Am 13. März 2012 14:15 schrieb Gary Gregory garydgreg...@gmail.com:
 On Tue, Mar 13, 2012 at 8:39 AM, Torsten Curdt tcu...@vafer.org wrote:
 
 I find checkstyle to be not very useful. It's more hassle than it's
 worth. It's like pointing fingers instead of helping. If you want to
 foster a certain code style provide eclipse and intellij formatter
 settings instead - that's actually helping. Especially if you run them
 before a release.
 
 
 I'd /love/ to have IDE settings for formatting saved in a project. This is
 the 21st century, all IDEs support this, if you do not use an IDE (hi
 Gilles), then, well, you probably also like driving a stick for control
 :) The only tricky part is how organize such a folder to account for
 different IDEs and versions. For example ide/eclipse/3.7.1,
 ide/intellij/10.5.3, and so on. Then you can move the IDE files to where
 each IDE wants it.
 
 
 Very big +1 Gary! It would make contributing to the various components
 so much easier (at least for people how use an IDE ;). I guess we
 should discuss this on a new thread.

Although I would also like to have the checkstyle configuration for IntelliJ as 
part of the project, I also like having the checkstyle report to get the 
overview of the whole project.

Ralph
-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Simone Tripodi
 Unless I'm missing something:

  (Using CheckStyle) != (Using the same formatting style in all projects)

+1

http://people.apache.org/~simonetripodi/
http://simonetripodi.livejournal.com/
http://twitter.com/simonetripodi
http://www.99soft.org/

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Gilles Sadowski
Hi.

  [...]
 
  The tools are there, but you have to tell people that they _must_ use them.
 
 Commons has already enough rules and process. As long as the releases
 are have clean code I wouldn't be too anal about the commits in
 between.

I think that the main disagreement is here. Source code must be a clear read
for the _developers_. To put it bluntly, I don't care that the releases have
cleanly formatted code, as almost nobody is going to read those packaged
sources!

 [...]

Gilles

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Christian Grobmeier
On Tue, Mar 13, 2012 at 5:00 PM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
 Hi.

  [...]
 
  The tools are there, but you have to tell people that they _must_ use them.

 Commons has already enough rules and process. As long as the releases
 are have clean code I wouldn't be too anal about the commits in
 between.

 I think that the main disagreement is here. Source code must be a clear read
 for the _developers_. To put it bluntly, I don't care that the releases have
 cleanly formatted code, as almost nobody is going to read those packaged
 sources!

Nobody objects using Checkstyle. Personally I object a default
Checkstyle config, which everybody must override. Nearly every
components has specifics, so everybody MUST override. What if you
don't want to use Checkstyle? Can you disable it?
What, if you use Sun conventions and Maven conventions are the
default? Much work! Please leave the checkstyle question to where it
belongs, and this is not parent pom, but the individual component.

And thats what I meant with: as long as we don't have a common
codestyle, i does not make much sense to have a common checkstyle
configuration.

Secondly, I have not had the feeling in the past years that checkstyle
helped me so much (including non open source projects). And so far, my
code was readable.




 [...]

 Gilles

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org




-- 
http://www.grobmeier.de
https://www.timeandbill.de

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Emmanuel Bourg
I'd like to see less reports by default. Most of them are only useful 
for checking a release candidate by the commons devs, otherwise it just 
clutters the sites with useless information for a general usage.


The most important reports to publish for the public are:
- Javadoc
- JXR
- Clirr
- Changes (I prefer the changes plugin, it's much more explicit than a 
list of issues in JIRA)



The reports only visible when voting on a release candidate:
- Cobertura
- Surefire
- RAT
- Findbugs
- Checkstyle

The useless ones:
- JDepend


Emmanuel Bourg


Le 13/03/2012 12:48, sebb a écrit :

Commons Parent 24 includes the following reports:

Javadoc
Jxr
Surefire
RAT
Cobertura
Clirr
JDepend

I think the following should be added:

Changes/JIRA

The following could be added:

Findbugs
Checkstyle

Any others?

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org






smime.p7s
Description: S/MIME Cryptographic Signature


Re: [ALL] Commons Parent reports

2012-03-13 Thread Gilles Sadowski
Hi.

 
   [...]
  
   The tools are there, but you have to tell people that they _must_ use 
   them.
 
  Commons has already enough rules and process. As long as the releases
  are have clean code I wouldn't be too anal about the commits in
  between.
 
  I think that the main disagreement is here. Source code must be a clear read
  for the _developers_. To put it bluntly, I don't care that the releases have
  cleanly formatted code, as almost nobody is going to read those packaged
  sources!
 
 Nobody objects using Checkstyle. Personally I object a default
 Checkstyle config, which everybody must override. Nearly every
 components has specifics, so everybody MUST override. What if you
 don't want to use Checkstyle? Can you disable it?
 What, if you use Sun conventions and Maven conventions are the
 default? Much work! Please leave the checkstyle question to where it
 belongs, and this is not parent pom, but the individual component.
 
 And thats what I meant with: as long as we don't have a common
 codestyle, i does not make much sense to have a common checkstyle
 configuration.

I thought that the question was whether to generate a CheckStyle report, not
whether the configuration should be the same...

 Secondly, I have not had the feeling in the past years that checkstyle
 helped me so much (including non open source projects). And so far, my
 code was readable.

My code is also readable...

I forgot to mention earlier in this thread that a code formatter will not
detect missing comments; I've also seen that some people using IDE let the
software generate totally useless place-holder Javadoc comments. Hence
no tool can afterwards detect that documentation is missing.


Regards,
Gilles

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Gilles Sadowski
On Tue, Mar 13, 2012 at 05:23:06PM +0100, Emmanuel Bourg wrote:
 I'd like to see less reports by default. Most of them are only
 useful for checking a release candidate by the commons devs,
 otherwise it just clutters the sites with useless information for a
 general usage.
 
 The most important reports to publish for the public are:
 - Javadoc
 - JXR
 - Clirr
 - Changes (I prefer the changes plugin, it's much more explicit than
 a list of issues in JIRA)
 
 
 The reports only visible when voting on a release candidate:
 - Cobertura
 - Surefire
 - RAT
 - Findbugs
 - Checkstyle
 
 The useless ones:
 - JDepend
 

What about the Useful for the developer category? 

I think that we originally mentioned a way to enable/disable the running of
the report generators. E.g. one could call from the command-line:

 mvn site -Dcheckstyle=off -Dfindbugs=on

The issue is to provide a top-level working config, then everyone can easily
run whatever they need.
 

Gilles

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Christian Grobmeier
On Tue, Mar 13, 2012 at 5:40 PM, Gilles Sadowski
gil...@harfang.homelinux.org wrote:
 And thats what I meant with: as long as we don't have a common
 codestyle, i does not make much sense to have a common checkstyle
 configuration.

 I thought that the question was whether to generate a CheckStyle report, not
 whether the configuration should be the same...

if you add a checkstyle report without a custom config, it has the
default configuration, right? In other terms, everybody would create a
report based on this. Or am I wrong?

 Secondly, I have not had the feeling in the past years that checkstyle
 helped me so much (including non open source projects). And so far, my
 code was readable.

 My code is also readable...

Sure, I didn't want to say otherwise :-)

 I forgot to mention earlier in this thread that a code formatter will not
 detect missing comments; I've also seen that some people using IDE let the
 software generate totally useless place-holder Javadoc comments. Hence
 no tool can afterwards detect that documentation is missing.

Right. I really like Torsten old blog post on JavaDoc comments:
http://vafer.org/blog/20050323095453/

If you have no good javadoc, leave it out. Sometimes you simply do not
have good javadocs checkdoc should not complain about my
decisions.

Cheers



 Regards,
 Gilles

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org




-- 
http://www.grobmeier.de
https://www.timeandbill.de

-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Emmanuel Bourg

Le 13/03/2012 17:52, Gilles Sadowski a écrit :


What about the Useful for the developer category?


They are useful at release time only, then they become quickly outdated 
as the code evolves after the release.


If I want to help Commons Math, I'll checkout the code and build the 
reports. Then I might inspect the Findbugs report and see if there is 
something to fix. But I'll never go to the website and browse months old 
reports.


The point is, these reports are valuable if they are updated 
continuously, but that's not possible to do that, because if the 
documentation is updated, the site will expose information that is not 
applicable to the latest release, but to the next one.


What I'd like to see is a more user oriented site, something clear and 
simple, and a developer oriented site with all the reports you want, 
automatically updated everyday.


Emmanuel Bourg



smime.p7s
Description: S/MIME Cryptographic Signature


Re: [ALL] Commons Parent reports

2012-03-13 Thread Gary Gregory
On Mar 13, 2012, at 12:23, Emmanuel Bourg ebo...@apache.org wrote:

 I'd like to see less reports by default. Most of them are only useful for 
 checking a release candidate by the commons devs, otherwise it just clutters 
 the sites with useless information for a general usage.

 The most important reports to publish for the public are:
 - Javadoc
 - JXR
 - Clirr
 - Changes (I prefer the changes plugin, it's much more explicit than a list 
 of issues in JIRA)


 The reports only visible when voting on a release candidate:
 - Cobertura
 - Surefire

These two are important IMO. It shows that we are not releasing
garbage, at least if the tests pass 100% with decent code coverage.
This is part of the open in open source IMO. Here is our component,
warts and all, use this information as a guide to help you decide if
this component meets your quality standard. Could you imagine
Microsoft releasing Office with this information? Never! That would
help competitors too much. Hey look at this over there, they don't
even test it. Or, looky here, they released knowing this test fails!

Gary

 - RAT
 - Findbugs
 - Checkstyle

 The useless ones:
 - JDepend


 Emmanuel Bourg


 Le 13/03/2012 12:48, sebb a écrit :
 Commons Parent 24 includes the following reports:

 Javadoc
 Jxr
 Surefire
 RAT
 Cobertura
 Clirr
 JDepend

 I think the following should be added:

 Changes/JIRA

 The following could be added:

 Findbugs
 Checkstyle

 Any others?

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org




-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Gary Gregory
On Mar 13, 2012, at 12:40, Gilles Sadowski gil...@harfang.homelinux.org wrote:


 [...]

 The tools are there, but you have to tell people that they _must_ use 
 them.

 Commons has already enough rules and process. As long as the releases
 are have clean code I wouldn't be too anal about the commits in
 between.

 I think that the main disagreement is here. Source code must be a clear read
 for the _developers_. To put it bluntly, I don't care that the releases have
 cleanly formatted code, as almost nobody is going to read those packaged
 sources!

Au contraire, most users will /only/ look at the source jar because
that is what you use in the debugger!

Gary


 Nobody objects using Checkstyle. Personally I object a default
 Checkstyle config, which everybody must override. Nearly every
 components has specifics, so everybody MUST override. What if you
 don't want to use Checkstyle? Can you disable it?
 What, if you use Sun conventions and Maven conventions are the
 default? Much work! Please leave the checkstyle question to where it
 belongs, and this is not parent pom, but the individual component.

 And thats what I meant with: as long as we don't have a common
 codestyle, i does not make much sense to have a common checkstyle
 configuration.

 I thought that the question was whether to generate a CheckStyle report, not
 whether the configuration should be the same...

 Secondly, I have not had the feeling in the past years that checkstyle
 helped me so much (including non open source projects). And so far, my
 code was readable.

 My code is also readable...

 I forgot to mention earlier in this thread that a code formatter will not
 detect missing comments; I've also seen that some people using IDE let the
 software generate totally useless place-holder Javadoc comments. Hence
 no tool can afterwards detect that documentation is missing.


 Regards,
 Gilles

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org


-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Gary Gregory
On Mar 13, 2012, at 12:40, Gilles Sadowski gil...@harfang.homelinux.org wrote:

 Hi.


 [...]

 The tools are there, but you have to tell people that they _must_ use 
 them.

 Commons has already enough rules and process. As long as the releases
 are have clean code I wouldn't be too anal about the commits in
 between.

 I think that the main disagreement is here. Source code must be a clear read
 for the _developers_. To put it bluntly, I don't care that the releases have
 cleanly formatted code, as almost nobody is going to read those packaged
 sources!

And another thing: the formatting /is/ important in released sources
because, again, this is what most users will see in their debuggers.
Have you seen some of the JRE sources? Some files are a mess, others
have blank lines in the middle of headers. Others look like they were
entered by a prisoner blinded in the noon day sun after spending a
month in the hole with bread and water ration and then given a stick
of butter for lunch.

Gary


 Nobody objects using Checkstyle. Personally I object a default
 Checkstyle config, which everybody must override. Nearly every
 components has specifics, so everybody MUST override. What if you
 don't want to use Checkstyle? Can you disable it?
 What, if you use Sun conventions and Maven conventions are the
 default? Much work! Please leave the checkstyle question to where it
 belongs, and this is not parent pom, but the individual component.

 And thats what I meant with: as long as we don't have a common
 codestyle, i does not make much sense to have a common checkstyle
 configuration.

 I thought that the question was whether to generate a CheckStyle report, not
 whether the configuration should be the same...

 Secondly, I have not had the feeling in the past years that checkstyle
 helped me so much (including non open source projects). And so far, my
 code was readable.

 My code is also readable...

 I forgot to mention earlier in this thread that a code formatter will not
 detect missing comments; I've also seen that some people using IDE let the
 software generate totally useless place-holder Javadoc comments. Hence
 no tool can afterwards detect that documentation is missing.


 Regards,
 Gilles

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org


-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Honton, Charles
When I'm researching competing open-source implementations, I use reports
as part of my evaluation.  The ones that are important to me are:

Javadoc
Findbugs
Surefire
Failsafe
Cobertura

If I'm evaluating an upgrade, I'll look at:

Changes
Clirr


Chas Honton


Le 13/03/2012 17:52, Gilles Sadowski a écrit :

 What about the Useful for the developer category?


-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread sebb
On 13 March 2012 17:59, Gary Gregory garydgreg...@gmail.com wrote:
 On Mar 13, 2012, at 12:40, Gilles Sadowski gil...@harfang.homelinux.org 
 wrote:

 Hi.


 [...]

 The tools are there, but you have to tell people that they _must_ use 
 them.

 Commons has already enough rules and process. As long as the releases
 are have clean code I wouldn't be too anal about the commits in
 between.

 I think that the main disagreement is here. Source code must be a clear 
 read
 for the _developers_. To put it bluntly, I don't care that the releases 
 have
 cleanly formatted code, as almost nobody is going to read those packaged
 sources!

 And another thing: the formatting /is/ important in released sources
 because, again, this is what most users will see in their debuggers.
 Have you seen some of the JRE sources? Some files are a mess, others
 have blank lines in the middle of headers. Others look like they were
 entered by a prisoner blinded in the noon day sun after spending a
 month in the hole with bread and water ration and then given a stick
 of butter for lunch.

No, that was a 'tab' of butter (which then sometimes got stuck into the source).

 Gary


 Nobody objects using Checkstyle. Personally I object a default
 Checkstyle config, which everybody must override. Nearly every
 components has specifics, so everybody MUST override. What if you
 don't want to use Checkstyle? Can you disable it?
 What, if you use Sun conventions and Maven conventions are the
 default? Much work! Please leave the checkstyle question to where it
 belongs, and this is not parent pom, but the individual component.

 And thats what I meant with: as long as we don't have a common
 codestyle, i does not make much sense to have a common checkstyle
 configuration.

 I thought that the question was whether to generate a CheckStyle report, not
 whether the configuration should be the same...

 Secondly, I have not had the feeling in the past years that checkstyle
 helped me so much (including non open source projects). And so far, my
 code was readable.

 My code is also readable...

 I forgot to mention earlier in this thread that a code formatter will not
 detect missing comments; I've also seen that some people using IDE let the
 software generate totally useless place-holder Javadoc comments. Hence
 no tool can afterwards detect that documentation is missing.


 Regards,
 Gilles

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org


 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org


-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org



Re: [ALL] Commons Parent reports

2012-03-13 Thread Gary Gregory
On Mar 13, 2012, at 14:05, sebb seb...@gmail.com wrote:

 On 13 March 2012 17:59, Gary Gregory garydgreg...@gmail.com wrote:
 On Mar 13, 2012, at 12:40, Gilles Sadowski gil...@harfang.homelinux.org 
 wrote:

 Hi.


 [...]

 The tools are there, but you have to tell people that they _must_ use 
 them.

 Commons has already enough rules and process. As long as the releases
 are have clean code I wouldn't be too anal about the commits in
 between.

 I think that the main disagreement is here. Source code must be a clear 
 read
 for the _developers_. To put it bluntly, I don't care that the releases 
 have
 cleanly formatted code, as almost nobody is going to read those packaged
 sources!

 And another thing: the formatting /is/ important in released sources
 because, again, this is what most users will see in their debuggers.
 Have you seen some of the JRE sources? Some files are a mess, others
 have blank lines in the middle of headers. Others look like they were
 entered by a prisoner blinded in the noon day sun after spending a
 month in the hole with bread and water ration and then given a stick
 of butter for lunch.

 No, that was a 'tab' of butter (which then sometimes got stuck into the 
 source).

Darn, I should have checkstyled my message!

Gary


 Gary


 Nobody objects using Checkstyle. Personally I object a default
 Checkstyle config, which everybody must override. Nearly every
 components has specifics, so everybody MUST override. What if you
 don't want to use Checkstyle? Can you disable it?
 What, if you use Sun conventions and Maven conventions are the
 default? Much work! Please leave the checkstyle question to where it
 belongs, and this is not parent pom, but the individual component.

 And thats what I meant with: as long as we don't have a common
 codestyle, i does not make much sense to have a common checkstyle
 configuration.

 I thought that the question was whether to generate a CheckStyle report, not
 whether the configuration should be the same...

 Secondly, I have not had the feeling in the past years that checkstyle
 helped me so much (including non open source projects). And so far, my
 code was readable.

 My code is also readable...

 I forgot to mention earlier in this thread that a code formatter will not
 detect missing comments; I've also seen that some people using IDE let the
 software generate totally useless place-holder Javadoc comments. Hence
 no tool can afterwards detect that documentation is missing.


 Regards,
 Gilles

 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org


 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org


 -
 To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
 For additional commands, e-mail: dev-h...@commons.apache.org


-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org