Oh, I agree, the vast majority of them are failures to close (not as in your example, as that would be closed fine, but failing to close on exceptions and other dumb things).  But I've also seen some cases that made me think otherwise.  But, I don't know, I'm all for removing the hack and then waiting until someone whines. :)  Maybe it was just my imagination before, or maybe it has long since been fixed.

On 9/1/05, Gili <[EMAIL PROTECTED]> wrote:

        It's been around since JDK 1.3.

Gili

Martijn Dashorst wrote:
> How is the availability of the Task/Timer things on JDK 1.4?
>
> Martijn
>
> Johan Compagner wrote:
>
>> currently Task it is only used by the modificationwatcher
>> So if that one can be rewritten then we could remove that code yes
>> Because Thread pools/Schedulers and so on is not a think wicket has to
>> provide..
>>
>> johan
>>
>>
>> Gwyn Evans wrote:
>>
>>> It remains, however, true that the case for the existance of the bug
>>> isn't supported by the discussion at the link that's given as the
>>> reason for it's existance...  I'm not convinced it should be there.
>>>
>>> I think Gili's also got a good point with regards to the home-grown
>>> Task, as opposed to Timer/TimerTask...
>>>
>>> /Gwyn
>>>
>>> On 01/09/05, Johan Compagner <[EMAIL PROTECTED]> wrote:
>>>
>>>
>>>> this is not true.
>>>>
>>>> fileinputstream is closed that way.
>>>> If you only close the most outer in/out put stream then it doesn't
>>>> matter how far you nest
>>>> it will all be closed.
>>>>
>>>> Gili wrote:
>>>>
>>>>
>>>>>     Are you *absolutely* sure? All too often I see people wrapping
>>>>> open files in FileInputStream, BufferedInputStream, InputStreamReader,
>>>>> etc... and then forgetting to close the objects on every single level.
>>>>>
>>>>>     Often I see:
>>>>>
>>>>>     BufferedInputStream in = new BufferedInputStream(new
>>>>> FileInputStream("foo.xml"));
>>>>>     in.close();
>>>>>
>>>>>     but unless you close off the FileInputStream (which you never
>>>>> saved a handled to) you never really closed the file.
>>>>>
>>>>>     Until I see some sort of bug report in BugParade I strongly
>>>>> suspect this is more of a rumor or user-error than an actual bug in
>>>>> the JDK. Whenever I've ever run into this issue myself (and I have)
>>>>> it's always turned out to be user-error which I fixed by cleaning up
>>>>> my code.
>>>>>
>>>>> Gili
>>>>>
>>>>> Jesse Sightler wrote:
>>>>>
>>>>>
>>>>>> Actually, I have definitely seen weirdness with deleting files on
>>>>>> Windows on occasion when I was sure that streams were being closed.
>>>>>> I suspect it had to do with some unique circumstances (Java patch
>>>>>> level, windows level, possibly a network drive in the way).
>>>>>>
>>>>>> What worked in our case was to simply try the delete more than once.
>>>>>> It didn't even seem to matter if there was a delay between the two
>>>>>> retries.
>>>>>>
>>>>>> --
>>>>>> Jess
>>>>>>
>>>>>> On 8/31/05, *Gili* < [EMAIL PROTECTED]
>>>>>> <mailto:[EMAIL PROTECTED]>> wrote:
>>>>>>
>>>>>>
>>>>>>             You might want to revisit some of the classes in this
>>>>>>     package. I might
>>>>>>     understand Duration which is a immutable Date object, at least
>>>>>> you're
>>>>>>     gaining something tangible there. But what is actually gained by
>>>>>>     wicket.util.Task when Java already has Timer and TimerTask which
>>>>>> do the
>>>>>>     same thing? Also Files.remove(File file) looks like a hack at
>>>>>> best. I
>>>>>>     followed the URL mentioned in the method and it looks to me like
>>>>>> these
>>>>>>     people were on crack or something. There is no problem deleting
>>>>>> files
>>>>>>     under Windows. My application does this on a regular basis. If
>>>>>> you read
>>>>>>     the 2nd to last comment on the first page (by tommyxman1) you
>>>>>> will
>>>>>>     notice that he was forgetting to close a FileInputStream, and I
>>>>>> suspect
>>>>>>     similar problems are responsible for the rest of the posters'
>>>>>> problems.
>>>>>>
>>>>>>             Our hack seems to encourage a finalizer to run but the
>>>>>>     correct behavior
>>>>>>     is to properly close the file in the first place. Anyway, you
>>>>>> might want
>>>>>>     to drop some of these classes and begin using the built-in Java
>>>>>>     equivilents.
>>>>>>
>>>>>>             Some of the classes in wicket.util have a good usage
>>>>>> (they
>>>>>>     provide
>>>>>>     helper functions around preexisting Java classes) but others
>>>>>> such as
>>>>>>     Task simply reinvent the wheel with no obvious benefit.
>>>>>>
>>>>>>     Gili
>>>>>>     --
>>>>>>     http://www.desktopbeautifier.com/
>>>>>> <http://www.desktopbeautifier.com/ >
>>>>>>
>>>>>>
>>>>>>     -------------------------------------------------------
>>>>>>     SF.Net email is Sponsored by the Better Software Conference &
>>>>>> EXPO
>>>>>>     September 19-22, 2005 * San Francisco, CA * Development Lifecycle
>>>>>>     Practices
>>>>>>     Agile & Plan-Driven Development * Managing Projects & Teams *
>>>>>>     Testing & QA
>>>>>>     Security * Process Improvement & Measurement *
>>>>>>     http://www.sqe.com/bsce5sf
>>>>>>     _______________________________________________
>>>>>>     Wicket-develop mailing list
>>>>>>     [email protected]
>>>>>>     <mailto:[email protected]>
>>>>>>     https://lists.sourceforge.net/lists/listinfo/wicket-develop
>>>>>>     <https://lists.sourceforge.net/lists/listinfo/wicket-develop >
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>> -------------------------------------------------------
>>>> SF.Net email is Sponsored by the Better Software Conference & EXPO
>>>> September 19-22, 2005 * San Francisco, CA * Development Lifecycle
>>>> Practices
>>>> Agile & Plan-Driven Development * Managing Projects & Teams *
>>>> Testing & QA
>>>> Security * Process Improvement & Measurement *
>>>> http://www.sqe.com/bsce5sf
>>>> _______________________________________________
>>>> Wicket-develop mailing list
>>>> [email protected]
>>>> https://lists.sourceforge.net/lists/listinfo/wicket-develop
>>>>
>>>>
>>>
>>>
>>>
>>>
>>> -------------------------------------------------------
>>> SF.Net email is Sponsored by the Better Software Conference & EXPO
>>> September 19-22, 2005 * San Francisco, CA * Development Lifecycle
>>> Practices
>>> Agile & Plan-Driven Development * Managing Projects & Teams * Testing
>>> & QA
>>> Security * Process Improvement & Measurement *
>>> http://www.sqe.com/bsce5sf
>>> _______________________________________________
>>> Wicket-develop mailing list
>>> [email protected]
>>> https://lists.sourceforge.net/lists/listinfo/wicket-develop
>>>
>>>
>>
>>
>>
>>
>> -------------------------------------------------------
>> SF.Net email is Sponsored by the Better Software Conference & EXPO
>> September 19-22, 2005 * San Francisco, CA * Development Lifecycle
>> Practices
>> Agile & Plan-Driven Development * Managing Projects & Teams * Testing
>> & QA
>> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
>> _______________________________________________
>> Wicket-develop mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/wicket-develop
>
>
>
>
>
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> Wicket-develop mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/wicket-develop
>

--
http://www.desktopbeautifier.com/


-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
Wicket-develop mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/wicket-develop

Reply via email to