Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.java.help > #571 > unrolled thread

Good practice or not to close the file before System.exit(1)?

Back to article view | Back to comp.lang.java.help

Contents

  Good practice or not to close the file before System.exit(1)? Merciadri Luca <Luca.Merciadri@student.ulg.ac.be> - 2011-04-22 00:13 +0200
    Re: Good practice or not to close the file before System.exit(1)? rossum <rossum48@coldmail.com> - 2011-04-22 00:11 +0100
      Re: Good practice or not to close the file before System.exit(1)? rossum <rossum48@coldmail.com> - 2011-04-22 12:39 +0100
        Re: Good practice or not to close the file before System.exit(1)? Patricia Shanahan <pats@acm.org> - 2011-04-22 05:09 -0700
          Re: Good practice or not to close the file before System.exit(1)? Lew <noone@lewscanon.com> - 2011-04-22 09:39 -0400
            Re: Good practice or not to close the file before System.exit(1)? Merciadri Luca <Luca.Merciadri@student.ulg.ac.be> - 2011-04-23 13:20 +0200
      Re: Good practice or not to close the file before System.exit(1)? Merciadri Luca <Luca.Merciadri@student.ulg.ac.be> - 2011-04-22 16:37 +0200
    Re: Good practice or not to close the file before System.exit(1)? Patricia Shanahan <pats@acm.org> - 2011-04-22 04:55 -0700
    Re: Good practice or not to close the file before System.exit(1)? Merciadri Luca <Luca.Merciadri@student.ulg.ac.be> - 2011-04-22 13:56 +0200
      Re: Good practice or not to close the file before System.exit(1)? Arved Sandstrom <asandstrom3minus1@eastlink.ca> - 2011-04-22 09:36 -0300
      Re: Good practice or not to close the file before System.exit(1)? markspace <-@.> - 2011-04-22 08:20 -0700
        Re: Good practice or not to close the file before System.exit(1)? Patricia Shanahan <pats@acm.org> - 2011-04-22 09:04 -0700
          Re: Good practice or not to close the file before System.exit(1)? markspace <-@.> - 2011-04-22 15:24 -0700
            Re: Good practice or not to close the file before System.exit(1)? markspace <-@.> - 2011-04-22 18:11 -0700
              Re: Good practice or not to close the file before System.exit(1)? markspace <-@.> - 2011-04-23 08:25 -0700
        Re: Good practice or not to close the file before System.exit(1)? Merciadri Luca <Luca.Merciadri@student.ulg.ac.be> - 2011-04-22 18:55 +0200
          Re: Good practice or not to close the file before System.exit(1)? Lew <noone@lewscanon.com> - 2011-04-22 13:40 -0400
          Re: Good practice or not to close the file before System.exit(1)? Patricia Shanahan <pats@acm.org> - 2011-04-22 10:55 -0700
            Re: Good practice or not to close the file before System.exit(1)? Merciadri Luca <Luca.Merciadri@student.ulg.ac.be> - 2011-04-22 22:58 +0200
              Re: Good practice or not to close the file before System.exit(1)? Merciadri Luca <Luca.Merciadri@student.ulg.ac.be> - 2011-04-22 23:09 +0200
                Re: Good practice or not to close the file before System.exit(1)? Patricia Shanahan <pats@acm.org> - 2011-04-22 14:30 -0700
                  Re: Good practice or not to close the file before System.exit(1)? markspace <-@.> - 2011-04-22 15:38 -0700
                    Re: Good practice or not to close the file before System.exit(1)? Patricia Shanahan <pats@acm.org> - 2011-04-22 17:00 -0700
                      Re: Good practice or not to close the file before System.exit(1)? Lew <noone@lewscanon.com> - 2011-04-22 22:10 -0400
                        Re: Good practice or not to close the file before System.exit(1)? Patricia Shanahan <pats@acm.org> - 2011-04-22 19:46 -0700
                          Re: Good practice or not to close the file before System.exit(1)? Lew <noone@lewscanon.com> - 2011-04-23 09:54 -0400
    Re: Good practice or not to close the file before System.exit(1)? Roedy Green <see_website@mindprod.com.invalid> - 2011-04-22 12:45 -0700
      Re: Good practice or not to close the file before System.exit(1)? Merciadri Luca <Luca.Merciadri@student.ulg.ac.be> - 2011-04-22 23:20 +0200

Page 2 of 2 — ← Prev page 1 [2]

#593

On 4/22/2011 2:09 PM, Merciadri Luca wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Another question comes to my mind:
>
> 3) Let us consider that your main deals with I/O. Then, it is common
> to use exceptions, e.g. IOException (for dealing with files for
> example) or NumberFormatException (for parseInt for example). However,
> it is disadviced to use exceptions in the main. As a result, I can
> reasonably conclude that dealing with I/O should go into (a) seperate
> classe[s], but is it really a good idea when you don't need to deal
> with that much input from the user?

There is no problem using exceptions inside main, as long as you catch
them, so that they don't get thrown from main.

Patricia

[toc] | [prev] | [next] | [standalone]

#595

On 4/22/2011 2:30 PM, Patricia Shanahan wrote:

>
> There is no problem using exceptions inside main, as long as you catch
> them, so that they don't get thrown from main.

During development, throwing exceptions from main is fine because it 
gives the developer the most about of useful information.

If you release a program into the wild, then yes, end-users might not 
want to see stack traces.  Although, some might.  It all depends on your 
user requirements.  Some technically sophisticated users with access to 
your source code might prefer a stack trace.

So I don't think there's any hard and fast rule that says always do or 
always don't.  It depends on what you want the program to do.

[toc] | [prev] | [next] | [standalone]

#596

On 4/22/2011 3:38 PM, markspace wrote:
> On 4/22/2011 2:30 PM, Patricia Shanahan wrote:
>
>>
>> There is no problem using exceptions inside main, as long as you catch
>> them, so that they don't get thrown from main.
>
> During development, throwing exceptions from main is fine because it
> gives the developer the most about of useful information.

Yes, I was over-simplifying a bit.

Patricia

[toc] | [prev] | [next] | [standalone]

#600

Patricia Shanahan wrote:
> markspace wrote:
>> Patricia Shanahan wrote:
>>
>>>
>>> There is no problem using exceptions inside main, as long as you catch
>>> them, so that they don't get thrown from main.
>>
>> During development, throwing exceptions from main is fine because it
>> gives the developer the most about of useful information.
>
> Yes, I was over-simplifying a bit.

I disagree.  Throwing exceptions from main() provides no benefit that logging 
doesn't provide better.

I'm using "main()" as a catch-all term for the top business layer of an 
application, be it a Java application, applet, web app, service or whatever. 
The observable behavior must conform to the problem domain being solved by the 
application.  Emitting implementation-level messages at the business-domain 
level is inappropriate unless the business domain *is* the implementation. 
Even there, there is a conceptual difference between the business domain and 
the implementation domain.

Design your business logic to be business logic.  DO NOT make the mistake of 
"temporarily" inserting inappropriate code thinking, "We can pull this out 
later for production."  Later doesn't come around enough to count on that.

Design it right from day one and do not stray from being good.

-- 
Lew
Honi soit qui mal y pense.
http://upload.wikimedia.org/wikipedia/commons/c/cf/Friz.jpg

[toc] | [prev] | [next] | [standalone]

#601

On 4/22/2011 7:10 PM, Lew wrote:
> Patricia Shanahan wrote:
>> markspace wrote:
>>> Patricia Shanahan wrote:
>>>
>>>>
>>>> There is no problem using exceptions inside main, as long as you catch
>>>> them, so that they don't get thrown from main.
>>>
>>> During development, throwing exceptions from main is fine because it
>>> gives the developer the most about of useful information.
>>
>> Yes, I was over-simplifying a bit.
>
> I disagree. Throwing exceptions from main() provides no benefit that
> logging doesn't provide better.

The benefit is simplicity.

I often write short test programs, for example when looking into a
newsgroup question, for which I'm the only user and that will only be
run once. Just sticking "throws Exception" on main is simpler than
catching exceptions and logging them.

Patricia

[toc] | [prev] | [next] | [standalone]

#606

Patricia Shanahan wrote:
> Lew wrote:
>> Patricia Shanahan wrote:
>>> markspace wrote:
>>>> Patricia Shanahan wrote:
>>>>
>>>>>
>>>>> There is no problem using exceptions inside main, as long as you catch
>>>>> them, so that they don't get thrown from main.
>>>>
>>>> During development, throwing exceptions from main is fine because it
>>>> gives the developer the most about of useful information.
>>>
>>> Yes, I was over-simplifying a bit.
>>
>> I disagree. Throwing exceptions from main() provides no benefit that
>> logging doesn't provide better.

> The benefit is simplicity.
>
> I often write short test programs, for example when looking into a
> newsgroup question, for which I'm the only user and that will only be
> run once. Just sticking "throws Exception" on main is simpler than
> catching exceptions and logging them.

I repeat the part that you omitted, which changes the sense of the comment you 
cited,
> I'm using "main()" as a catch-all term for the top business layer of an application,
> be it a Java application, applet, web app, service or whatever.
> The observable behavior must conform to the problem domain being solved by the application.
> Emitting implementation-level messages at the business-domain level is inappropriate
> unless the business domain *is* the implementation.

You speak of the case where the implementation is the domain, which I 
explicitly called out as a case where I agree that exceptions can be useful in 
the "main".  Thanks for elucidating that point.

-- 
Lew
Honi soit qui mal y pense.
http://upload.wikimedia.org/wikipedia/commons/c/cf/Friz.jpg

[toc] | [prev] | [next] | [standalone]

#589

On Fri, 22 Apr 2011 00:13:48 +0200, Merciadri Luca
<Luca.Merciadri@student.ulg.ac.be> wrote, quoted or indirectly quoted
someone who said :

>In some cases, I need to make a System.exit(1) in my program. However,
>a file has been opened some time before, and it has not been yet
>closed. Do I need to close it for good practice? How is it handled by
>Java if I don't close the file?

This is not FORTRAN.  Java does not close files for you if you don't.
If you were writing, the final buffers full would not get written. The
OS of course would close the files at that level, but it knows nothing
of the buffers.
-- 
Roedy Green Canadian Mind Products
http://mindprod.com
Politicians complain that Kindles and iBooks are killing jobs by 
destroying the paper book industry.  I see it that they have create a way 
to produce books for less than a third the cost without destroying forests 
and emitting greenhouse gases in the process.  They have created wealth.  
They are encouraging literacy and cutting the costs of education.  

[toc] | [prev] | [next] | [standalone]

#592

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Roedy Green <see_website@mindprod.com.invalid> writes:

> On Fri, 22 Apr 2011 00:13:48 +0200, Merciadri Luca
> <Luca.Merciadri@student.ulg.ac.be> wrote, quoted or indirectly quoted
> someone who said :
>
>>In some cases, I need to make a System.exit(1) in my program. However,
>>a file has been opened some time before, and it has not been yet
>>closed. Do I need to close it for good practice? How is it handled by
>>Java if I don't close the file?
>
> This is not FORTRAN.  Java does not close files for you if you don't.
> If you were writing, the final buffers full would not get written. The
> OS of course would close the files at that level, but it knows nothing
> of the buffers.

4) I knew this, and I've now changed everything that had to be changed
regarding the files' closing, be it for input or output. However, I
was wondering... if you don't close a file opened for reading, what does it
actually change? The way UNIX deals with files fortunately never locks
a[n] (currently) opened file. I simply ask what kind of problems the
non-closing of an opened file for reading would cause.

(Consider here that I only speak about the UNIX case; I know that Java
is evidently platform independent; as a result, one might want to
depend only on the VM and not on the OS. I just want to be informed
about the kind of troubles it might cause, and how the non-closing is dealt.)
- -- 
Merciadri Luca
See http://www.student.montefiore.ulg.ac.be/~merciadri/
- -- 

First deserve, then desire.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/>

iEYEARECAAYFAk2x8TcACgkQM0LLzLt8MhydiACfTU7tMHXSXY6lM+mIMj1CYn5O
AWgAoKrE6+kmAYDuahVkEbPiOiQ0LToS
=FEVU
-----END PGP SIGNATURE-----

[toc] | [prev] | [standalone]

Page 2 of 2 — ← Prev page 1 [2]

Back to top | Article view | comp.lang.java.help

csiph-web