OK, now it's clear.
But why you did not specify earlier, that the JAR/not JAR issue affects
performance only in COMPILE time.
But the original question was (if you remember):

"Does having imports such as "import java.util.*" as opposed to "import
java.util.Vector" (when only the Vector class is used in the program) affect
the performance of a stand alone program?"
                               ^^^^^^^^^^^^^^^^^^^^^^^

Radim

>
> At 08:38 AM 6/28/2002 +0200, Radim Tlusty wrote:
>  >> At 03:20 PM 6/27/2002 +0200, Radim Tlusty wrote:
>  >>  >But the following description is misleading:
>
>  >>  >There is no relation between imports and storage of class files.
>  >>
>  >>   ??? I'm not sure how your inference engine works.
>  >>   Who said there was a relationship?
>  >
>  >YOU.
>
>       ?????  I said "it used to be overhead in the old days..."
>       Compilation took longer when your class path only has
>       class files because there was a directory lookup involved.
>       Jar files eliminated that issue a while ago.
>
>  >You said, that there was overhead in "old days" before JAR
> files and today
>  >with JAR files with manifest there should not be difference
> between "import
>  >java.util.*" and "import java.util.Vector".
>
>  >This for me means, that you see
>  >relation between storage of classes and "import style" in
> source code. But
>  >this is misleading.
>
>       "see relation"???   "import style"???  I have no idea
>       what you are saying.  Please explain.
>
>  >There is of course difference in speed with classes stored in
> JAR files or
>  >standalone, but in loading time of class and not because of
> imports, as you
>  >said (cite from your statement: "imports used to cause
> directory lookups,
>  >which were slow").
>
>       There *WAS* a difference (between import foo.* and import
> foo.MyClass)
>       in compiling Java classes in the pre-jar days.  It *USED* to take
>       longer to compile Java classes when we only had *class* files and
>       did not have *jar* files.  This difference was due to the directory
>       lookup and was particularly noticeable with large numbers
> of class files.
>
>       If you want further clarification, send me direct email.
>
>       Frank G.
> +======================================================================+
> | Crossroads Technologies Inc, 55 Broad Street, 28th Fl, NYC, NY 10004 |
> |  Enterprise Java Engineering                                         |
> | Email: [EMAIL PROTECTED]         Web: www.CrossroadsTech.com |
> | Voice: 212-482-5280 x229                 Fax: 212-482-5281           |
> +======================================================================+
>
>
> _______________________________________________
> Advanced-swing mailing list
> [EMAIL PROTECTED]
> http://eos.dk/mailman/listinfo/advanced-swing
>

_______________________________________________
Advanced-swing mailing list
[EMAIL PROTECTED]
http://eos.dk/mailman/listinfo/advanced-swing

Reply via email to