question about the patch #235

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

question about the patch #235

kazuro fukuhara
Dear everyone.

I have one question about this patch that I wrote.

http://sourceforge.net/p/checkstyle/patches/235/

Question : Is this feaure needed for Checkstyle?


If this is needed, I'll try to make pull request.
If not, I'll give up about this.
If this feature is needed but this have problem, Shall we have a discussion.
I will modify this patch.

Best Regards,
Kazuro Fukuhara



------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk
_______________________________________________
Checkstyle-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/checkstyle-devel
Reply | Threaded
Open this post in threaded view
|

Re: question about the patch #235

Roman Ivanov
Hello Kazuro Fukuhara.

Thanks a lot for time you spend with Checkstyle update. But would be better to discuss such update in advance in mail list.

From what I see in your description, you want feature that not all checks would possible support and do care about.
You asking not only detect a problem but also evaluate it significanse or weight of it.

But what is the real difference between (where limit is 2000)
Case 1. The number of line of java source code is 2100.
and
Case 1. The number of line of java source code is 5000.


For you 5000 lines length is critical , but for another project it is not a big deal. And even in your project 2100 and 5000 files size is same problem - just go and fix it,  If you do not want - application will not fail. Even more refactoring of 5000 to 2x2000 + 1000 could be easie then 2100 to 2000 + 100.
Checkstyle does not estimate problem for you, it only points you to problem, and you decide.

Just severity is more than enought:
<property name="severity" value="ignore"/>
Checks implementation does not care about severity at all, and do not see a serious benefit from your request.
Please extend your idea, and try to persuade other developers if you your position is strong.

Looks like you have different problem .... and just try to solve it from wrong side.
I advise to use Checkstyle just in time you write code in you IDE - so you will not have that problem.
If you want a solution please share your initial problem.
 
just an idea:
think about your custom Severity handler, you do trigger in configuration, checkstyle call you on required Checks and by parsing their message change severity as you want. But do not expect you custom Severity Handler be part of Checkstyle library.

Thanks,
Roman Ivanov

>
> Dear everyone.
>
> I have one question about this patch that I wrote.
>
> http://sourceforge.net/p/checkstyle/patches/235/
>
> Question : Is this feaure needed for Checkstyle?
>
>
> If this is needed, I'll try to make pull request.
> If not, I'll give up about this.
> If this feature is needed but this have problem, Shall we have a discussion.
> I will modify this patch.
>
> Best Regards,
> Kazuro Fukuhara
>
>
>

Роман Иванов
------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk
_______________________________________________
Checkstyle-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/checkstyle-devel
Reply | Threaded
Open this post in threaded view
|

Re: question about the patch #235

kazuro fukuhara
Hello Roman Ivanov,
I will thank you for your response.

> From what I see in your description, you want feature that not all checks would possible support and do care about.
> You asking not only detect a problem but also evaluate it significanse or weight of it.

Yes.

> But what is the real difference between (where limit is 2000)
> Case 1. The number of line of java source code is 2100.
> and
> Case 1. The number of line of java source code is 5000.
>
> For you 5000 lines length is critical , but for another project it is not a big deal. And even in your project 2100 and 5000 files size is same problem - just go and fix it,  If you do not want - application will not fail. Even more refactoring of 5000 to 2x2000 + 1000 could be easie then 2100 to 2000 + 100.
> Checkstyle does not estimate problem for you, it only points you to problem, and you decide.


Whether the size of excess is large or not , violation is violation. And it is needed to clear if it is the rule of the project.
And I know we can configure 'max' value to FileLength Check to support many project.

By the way, for both case, current checkstyle will just say 'this is violation' regardless of its
size of excess. So let's suppose a developer tried to reduce line number of this file.
This violation report will not be removed until he achieved to reduce to 2000.

But if checkstyle reported violation with its score, checkstyle will change
the score of violation message even if he could not reduce number of line to 2000 but only reduce size.
He will notice his effort is reported.

To make this kind of small notification available is that I want to do.


Please let me explain how I use checkstyle in my project.

In my project, The lead time is the most important value for customer. The customer
want to see that actual application runs and check its behavior and experience as early as possible after he showed what he want.
In this situation, we can hardly have code review meeting.

But to keep code beautiful and maintenanceability are also important to make our service goes.
That's why I execute checkstyle every midnight using cron.
(By the way, this is another story , but we also use the eclipse-cs Checkstyle plug-in in daily development)

The script is ant project. It retrieve (just like 'svn update') all code, execute checkstyle,
deliver an email report as explaind url below.
http://checkstyle.sourceforge.net/anttask.html

And more, the ant project have custom ant task that do below for each Java code

1. check who is the last committer of today.
2. calculate score for FileLength Check. it is how much lines exceeded threadshold of it.
3. calculate difference of compare today's score and yesterday's score.
4. save today's score to local file system to use the score tomorrow.

After that, the script will calculate total score-difference for each developer.
And include the score-difference to mail that will be send.

All developer will receive error report with its score-difference. Even if he could
reduce number of line but not reached 2000, he will be reported as his effort made sense to raise
maintenanceability. He will be praised as its effort , without to reach 2000.
On the contrary, if he incresead the number of line beyond 2000, he will accused to make violation.

What I want to do in the ant script is to praise a team menber as its effort/value and accuse him as its generating violation.

This is summarized behavior of daily code check ant script. I thought this score calculation have to be
done by Checkstyle itself. For now our ant custom task parses error message to pick up the size of excess. I guess parsing message should not to do.
And I thought describing severity in detail is natural extension of the feature of checkstyle that represent severity.
http://checkstyle.sourceforge.net/property_types.html#severity

At present , checkstyle 5.6 have only 4 level of severity. I want to introduce detailed estimation according as its size of excees.

I have a question.
What is the goal or intention of Severity? 
http://checkstyle.sourceforge.net/config.html#Severity
Is this just a small information and no need detailed information?

Generally speaking, apart from checkstyle, I think that keeping feature simple is also desired.
It will be depend on basic design.
If Severity feature should not be extended, I can give up this patch.

Best Regards,
Kazuro Fukuhara


On Sun, Sep 29, 2013 at 9:33 AM, Roman Ivanov <[hidden email]> wrote:
Hello Kazuro Fukuhara.

Thanks a lot for time you spend with Checkstyle update. But would be better to discuss such update in advance in mail list.

>From what I see in your description, you want feature that not all checks would possible support and do care about.
You asking not only detect a problem but also evaluate it significanse or weight of it.

But what is the real difference between (where limit is 2000)
Case 1. The number of line of java source code is 2100.
and
Case 1. The number of line of java source code is 5000.


For you 5000 lines length is critical , but for another project it is not a big deal. And even in your project 2100 and 5000 files size is same problem - just go and fix it,  If you do not want - application will not fail. Even more refactoring of 5000 to 2x2000 + 1000 could be easie then 2100 to 2000 + 100.
Checkstyle does not estimate problem for you, it only points you to problem, and you decide.

Just severity is more than enought:
<property name="severity" value="ignore"/>
Checks implementation does not care about severity at all, and do not see a serious benefit from your request.
Please extend your idea, and try to persuade other developers if you your position is strong.

Looks like you have different problem .... and just try to solve it from wrong side.
I advise to use Checkstyle just in time you write code in you IDE - so you will not have that problem.
If you want a solution please share your initial problem.

just an idea:
think about your custom Severity handler, you do trigger in configuration, checkstyle call you on required Checks and by parsing their message change severity as you want. But do not expect you custom Severity Handler be part of Checkstyle library.

Thanks,
Roman Ivanov
>
> Dear everyone.
>
> I have one question about this patch that I wrote.
>
> http://sourceforge.net/p/checkstyle/patches/235/
>
> Question : Is this feaure needed for Checkstyle?
>
>
> If this is needed, I'll try to make pull request.
> If not, I'll give up about this.
> If this feature is needed but this have problem, Shall we have a discussion.
> I will modify this patch.
>
> Best Regards,
> Kazuro Fukuhara
>
>
>

Роман Иванов
------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk
_______________________________________________
Checkstyle-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/checkstyle-devel


------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk
_______________________________________________
Checkstyle-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/checkstyle-devel
Reply | Threaded
Open this post in threaded view
|

Re: question about the patch #235

Oliver Burn
I have to wonder whether the complexity this feature adds is worth the benefit for most users.

Typically, when evaluating a new feature, I imagine how hard it will be to describe/document it. If it will be an effort to describe the new feature clearly, I tend to think it is not worth doing. Most people will not use the feature and may be confused by it (leading to more work later supporting them).

Just me $0.02 worth.

Regards,
Oliver


On Tue, Oct 1, 2013 at 3:10 AM, kazuro fukuhara <[hidden email]> wrote:
Hello Roman Ivanov,
I will thank you for your response.

> From what I see in your description, you want feature that not all checks would possible support and do care about.
> You asking not only detect a problem but also evaluate it significanse or weight of it.

Yes.

> But what is the real difference between (where limit is 2000)
> Case 1. The number of line of java source code is 2100.
> and
> Case 1. The number of line of java source code is 5000.
>
> For you 5000 lines length is critical , but for another project it is not a big deal. And even in your project 2100 and 5000 files size is same problem - just go and fix it,  If you do not want - application will not fail. Even more refactoring of 5000 to 2x2000 + 1000 could be easie then 2100 to 2000 + 100.
> Checkstyle does not estimate problem for you, it only points you to problem, and you decide.


Whether the size of excess is large or not , violation is violation. And it is needed to clear if it is the rule of the project.
And I know we can configure 'max' value to FileLength Check to support many project.

By the way, for both case, current checkstyle will just say 'this is violation' regardless of its
size of excess. So let's suppose a developer tried to reduce line number of this file.
This violation report will not be removed until he achieved to reduce to 2000.

But if checkstyle reported violation with its score, checkstyle will change
the score of violation message even if he could not reduce number of line to 2000 but only reduce size.
He will notice his effort is reported.

To make this kind of small notification available is that I want to do.


Please let me explain how I use checkstyle in my project.

In my project, The lead time is the most important value for customer. The customer
want to see that actual application runs and check its behavior and experience as early as possible after he showed what he want.
In this situation, we can hardly have code review meeting.

But to keep code beautiful and maintenanceability are also important to make our service goes.
That's why I execute checkstyle every midnight using cron.
(By the way, this is another story , but we also use the eclipse-cs Checkstyle plug-in in daily development)

The script is ant project. It retrieve (just like 'svn update') all code, execute checkstyle,
deliver an email report as explaind url below.
http://checkstyle.sourceforge.net/anttask.html

And more, the ant project have custom ant task that do below for each Java code

1. check who is the last committer of today.
2. calculate score for FileLength Check. it is how much lines exceeded threadshold of it.
3. calculate difference of compare today's score and yesterday's score.
4. save today's score to local file system to use the score tomorrow.

After that, the script will calculate total score-difference for each developer.
And include the score-difference to mail that will be send.

All developer will receive error report with its score-difference. Even if he could
reduce number of line but not reached 2000, he will be reported as his effort made sense to raise
maintenanceability. He will be praised as its effort , without to reach 2000.
On the contrary, if he incresead the number of line beyond 2000, he will accused to make violation.

What I want to do in the ant script is to praise a team menber as its effort/value and accuse him as its generating violation.

This is summarized behavior of daily code check ant script. I thought this score calculation have to be
done by Checkstyle itself. For now our ant custom task parses error message to pick up the size of excess. I guess parsing message should not to do.
And I thought describing severity in detail is natural extension of the feature of checkstyle that represent severity.
http://checkstyle.sourceforge.net/property_types.html#severity

At present , checkstyle 5.6 have only 4 level of severity. I want to introduce detailed estimation according as its size of excees.

I have a question.
What is the goal or intention of Severity? 
http://checkstyle.sourceforge.net/config.html#Severity
Is this just a small information and no need detailed information?

Generally speaking, apart from checkstyle, I think that keeping feature simple is also desired.
It will be depend on basic design.
If Severity feature should not be extended, I can give up this patch.

Best Regards,
Kazuro Fukuhara


On Sun, Sep 29, 2013 at 9:33 AM, Roman Ivanov <[hidden email]> wrote:
Hello Kazuro Fukuhara.

Thanks a lot for time you spend with Checkstyle update. But would be better to discuss such update in advance in mail list.

>From what I see in your description, you want feature that not all checks would possible support and do care about.
You asking not only detect a problem but also evaluate it significanse or weight of it.

But what is the real difference between (where limit is 2000)
Case 1. The number of line of java source code is 2100.
and
Case 1. The number of line of java source code is 5000.


For you 5000 lines length is critical , but for another project it is not a big deal. And even in your project 2100 and 5000 files size is same problem - just go and fix it,  If you do not want - application will not fail. Even more refactoring of 5000 to 2x2000 + 1000 could be easie then 2100 to 2000 + 100.
Checkstyle does not estimate problem for you, it only points you to problem, and you decide.

Just severity is more than enought:
<property name="severity" value="ignore"/>
Checks implementation does not care about severity at all, and do not see a serious benefit from your request.
Please extend your idea, and try to persuade other developers if you your position is strong.

Looks like you have different problem .... and just try to solve it from wrong side.
I advise to use Checkstyle just in time you write code in you IDE - so you will not have that problem.
If you want a solution please share your initial problem.

just an idea:
think about your custom Severity handler, you do trigger in configuration, checkstyle call you on required Checks and by parsing their message change severity as you want. But do not expect you custom Severity Handler be part of Checkstyle library.

Thanks,
Roman Ivanov
>
> Dear everyone.
>
> I have one question about this patch that I wrote.
>
> http://sourceforge.net/p/checkstyle/patches/235/
>
> Question : Is this feaure needed for Checkstyle?
>
>
> If this is needed, I'll try to make pull request.
> If not, I'll give up about this.
> If this feature is needed but this have problem, Shall we have a discussion.
> I will modify this patch.
>
> Best Regards,
> Kazuro Fukuhara
>
>
>

Роман Иванов
------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk
_______________________________________________
Checkstyle-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/checkstyle-devel


------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk
_______________________________________________
Checkstyle-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/checkstyle-devel



------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60134791&iu=/4140/ostg.clktrk
_______________________________________________
Checkstyle-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/checkstyle-devel
Reply | Threaded
Open this post in threaded view
|

Re: question about the patch #235

kazuro fukuhara
Dear Oliver Burn,

Ok, I understand.
I'll give up the patch.

Best Regards,
Fukuhara




On Thu, Oct 3, 2013 at 10:05 PM, Oliver Burn <[hidden email]> wrote:
I have to wonder whether the complexity this feature adds is worth the benefit for most users.

Typically, when evaluating a new feature, I imagine how hard it will be to describe/document it. If it will be an effort to describe the new feature clearly, I tend to think it is not worth doing. Most people will not use the feature and may be confused by it (leading to more work later supporting them).

Just me $0.02 worth.

Regards,
Oliver


On Tue, Oct 1, 2013 at 3:10 AM, kazuro fukuhara <[hidden email]> wrote:
Hello Roman Ivanov,
I will thank you for your response.

> From what I see in your description, you want feature that not all checks would possible support and do care about.
> You asking not only detect a problem but also evaluate it significanse or weight of it.

Yes.

> But what is the real difference between (where limit is 2000)
> Case 1. The number of line of java source code is 2100.
> and
> Case 1. The number of line of java source code is 5000.
>
> For you 5000 lines length is critical , but for another project it is not a big deal. And even in your project 2100 and 5000 files size is same problem - just go and fix it,  If you do not want - application will not fail. Even more refactoring of 5000 to 2x2000 + 1000 could be easie then 2100 to 2000 + 100.
> Checkstyle does not estimate problem for you, it only points you to problem, and you decide.


Whether the size of excess is large or not , violation is violation. And it is needed to clear if it is the rule of the project.
And I know we can configure 'max' value to FileLength Check to support many project.

By the way, for both case, current checkstyle will just say 'this is violation' regardless of its
size of excess. So let's suppose a developer tried to reduce line number of this file.
This violation report will not be removed until he achieved to reduce to 2000.

But if checkstyle reported violation with its score, checkstyle will change
the score of violation message even if he could not reduce number of line to 2000 but only reduce size.
He will notice his effort is reported.

To make this kind of small notification available is that I want to do.


Please let me explain how I use checkstyle in my project.

In my project, The lead time is the most important value for customer. The customer
want to see that actual application runs and check its behavior and experience as early as possible after he showed what he want.
In this situation, we can hardly have code review meeting.

But to keep code beautiful and maintenanceability are also important to make our service goes.
That's why I execute checkstyle every midnight using cron.
(By the way, this is another story , but we also use the eclipse-cs Checkstyle plug-in in daily development)

The script is ant project. It retrieve (just like 'svn update') all code, execute checkstyle,
deliver an email report as explaind url below.
http://checkstyle.sourceforge.net/anttask.html

And more, the ant project have custom ant task that do below for each Java code

1. check who is the last committer of today.
2. calculate score for FileLength Check. it is how much lines exceeded threadshold of it.
3. calculate difference of compare today's score and yesterday's score.
4. save today's score to local file system to use the score tomorrow.

After that, the script will calculate total score-difference for each developer.
And include the score-difference to mail that will be send.

All developer will receive error report with its score-difference. Even if he could
reduce number of line but not reached 2000, he will be reported as his effort made sense to raise
maintenanceability. He will be praised as its effort , without to reach 2000.
On the contrary, if he incresead the number of line beyond 2000, he will accused to make violation.

What I want to do in the ant script is to praise a team menber as its effort/value and accuse him as its generating violation.

This is summarized behavior of daily code check ant script. I thought this score calculation have to be
done by Checkstyle itself. For now our ant custom task parses error message to pick up the size of excess. I guess parsing message should not to do.
And I thought describing severity in detail is natural extension of the feature of checkstyle that represent severity.
http://checkstyle.sourceforge.net/property_types.html#severity

At present , checkstyle 5.6 have only 4 level of severity. I want to introduce detailed estimation according as its size of excees.

I have a question.
What is the goal or intention of Severity? 
http://checkstyle.sourceforge.net/config.html#Severity
Is this just a small information and no need detailed information?

Generally speaking, apart from checkstyle, I think that keeping feature simple is also desired.
It will be depend on basic design.
If Severity feature should not be extended, I can give up this patch.

Best Regards,
Kazuro Fukuhara


On Sun, Sep 29, 2013 at 9:33 AM, Roman Ivanov <[hidden email]> wrote:
Hello Kazuro Fukuhara.

Thanks a lot for time you spend with Checkstyle update. But would be better to discuss such update in advance in mail list.

>From what I see in your description, you want feature that not all checks would possible support and do care about.
You asking not only detect a problem but also evaluate it significanse or weight of it.

But what is the real difference between (where limit is 2000)
Case 1. The number of line of java source code is 2100.
and
Case 1. The number of line of java source code is 5000.


For you 5000 lines length is critical , but for another project it is not a big deal. And even in your project 2100 and 5000 files size is same problem - just go and fix it,  If you do not want - application will not fail. Even more refactoring of 5000 to 2x2000 + 1000 could be easie then 2100 to 2000 + 100.
Checkstyle does not estimate problem for you, it only points you to problem, and you decide.

Just severity is more than enought:
<property name="severity" value="ignore"/>
Checks implementation does not care about severity at all, and do not see a serious benefit from your request.
Please extend your idea, and try to persuade other developers if you your position is strong.

Looks like you have different problem .... and just try to solve it from wrong side.
I advise to use Checkstyle just in time you write code in you IDE - so you will not have that problem.
If you want a solution please share your initial problem.

just an idea:
think about your custom Severity handler, you do trigger in configuration, checkstyle call you on required Checks and by parsing their message change severity as you want. But do not expect you custom Severity Handler be part of Checkstyle library.

Thanks,
Roman Ivanov
>
> Dear everyone.
>
> I have one question about this patch that I wrote.
>
> http://sourceforge.net/p/checkstyle/patches/235/
>
> Question : Is this feaure needed for Checkstyle?
>
>
> If this is needed, I'll try to make pull request.
> If not, I'll give up about this.
> If this feature is needed but this have problem, Shall we have a discussion.
> I will modify this patch.
>
> Best Regards,
> Kazuro Fukuhara
>
>
>

Роман Иванов
------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk
_______________________________________________
Checkstyle-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/checkstyle-devel


------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk
_______________________________________________
Checkstyle-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/checkstyle-devel



------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60134791&iu=/4140/ostg.clktrk
_______________________________________________
Checkstyle-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/checkstyle-devel



------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60134791&iu=/4140/ostg.clktrk
_______________________________________________
Checkstyle-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/checkstyle-devel
Reply | Threaded
Open this post in threaded view
|

Re: question about the patch #235

Roman Ivanov
In reply to this post by kazuro fukuhara
Hi Kazuro

>> That's why I execute checkstyle every midnight using cron.
In my experience, if developer was not alarmed after pressing Save in file editor all other validation of its code does not make huge sense - code unlikely to be changed (there will be enormous amount of reasons and other problems so you will unlikely comeback).

>> What is the goal or intention of Severity?  
>> Is this just a small information and no need detailed information?
Severity is feature of Checkstyle configuration from perspective of user.  Individual Checks classes have no idea with what severity this alarm will be shown to user.
Message and locaton is that all that  Checks could tell about problem.
You want to play with severity to your needs so you need to parse messages and make a decision, as I proposed before CustomSeverityHandler that receive all messages and raise Severity according to your needs. Checkstyle right now do "changes" for severity :) by suppressing errors copletely according to configuration of special comments in code.
Checkstyle is not ready right now to have customization on that level but I do believe that my be useful in some cases, and will not cost Checkstyle a lot.
If you provide patch for Checkstyle we could consider it (I do not promiss to apply) .... our team will be focused on other more acute problems to fix/implement.


Thanks,
Roman Ivanov
------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60134791&iu=/4140/ostg.clktrk
_______________________________________________
Checkstyle-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/checkstyle-devel