Le 01/08/2010 19:06, Gilles Sadowski a écrit :
>> /** {...@inheritdoc} */
>> - public void setConvergenceChecker(RealConvergenceChecker checker) {
>> - this.checker = checker;
>> + public void setConvergenceChecker(RealConvergenceChecker
>> convergenceChecker) {
>> + this.checker = convergenceChecker;
>> }
>
> Why does "checkstyle" complain about the name "checker" for the parameter?
It hides a field.
>
>> Modified:
>> commons/proper/math/trunk/src/main/java/org/apache/commons/math/util/MathUtils.java
>> URL:
>> http://svn.apache.org/viewvc/commons/proper/math/trunk/src/main/java/org/apache/commons/math/util/MathUtils.java?rev=981238&r1=981237&r2=981238&view=diff
>> ==============================================================================
>> ---
>> commons/proper/math/trunk/src/main/java/org/apache/commons/math/util/MathUtils.java
>> (original)
>> +++
>> commons/proper/math/trunk/src/main/java/org/apache/commons/math/util/MathUtils.java
>> Sun Aug 1 15:01:30 2010
>> @@ -1868,10 +1868,18 @@ public final class MathUtils {
>> }
>>
>> public static class Order {
>> +
>> + /** Enumerate type for increasing/decreasing directions. */
>> public static enum Direction {
>> - INCREASING,
>> - DECREASING
>> +
>> + /** Constant for increasing direction. */
>> + INCREASING,
>> +
>> + /** Constant for decreasing direction. */
>> + DECREASING
>
> Indeed, I don't know why I had created a two-level enum...
> Although, maybe it was because I was thinking that we could put other
> "characteristics" there, such as "strictness" (which now is specified
> through a boolean).
The perhaps we could revert to one level only ?
>
>> - * @throws {...@link NotStrictlyPositiveException} if one of the sizes
>> is
>> + * @throws NotStrictlyPositiveException if one of the sizes is
>
> Does Javadoc automatically create an HTML link for the argument of the
> "@throws" tag?
>
>> - public int getCount(int ... c) {
>> + public int getCount(int ... c) throws OutOfRangeException {
>
> Does "checkstyle" complain about unchecked exceptions not appearing in the
> throws clause?
It depends on the configuration. The current setting we have in [math]
is to allow missing RuntimeException. This specific change is an error
from me. I tried different solutions because checkstyle had a problem
with this file and stopped with an error. Feel free to revert it.
> On the one hand I would agree to put unchecked exceptions if it prevents
> tools from complaining about unused "import". On the other hand, it is
> considered to be bad programming style since it cannot mean that only the
> unchecked exceptions referred to there can actually be thrown.
I'm on the fence here. Personnally I would like to have all exceptions
declared and checkstyle could help us here if we adjust its setting. But
we did set it up this way and I don't remember why.
Luc
>
>
> Gilles
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
