🚀 go-pugleaf

"Amir Pashazadeh" <pasha@abdnet.com> writes:

> I decompiled a class file (using 3 different decompilers for
> accurancy), and every time I got the same source code, there is
> something amazing in it, it has "goto" instruction! how can it be?

Because VM instruction set is not the same as Java language keywords.

> does JVM (as executor) supports "goto" instruction, but no standard
> compiler supports it yet? so a custom compiler can use it?

The VM uses "goto" to jump around in the same sense x86 uses "jmp".
You need to have some sort of jumping instruction.

The "gotos" come from java-the-language loops (for, while) and sundry
others (break, continue).

On Sat, 23 Aug 2003 12:36:48 +0330, "Amir Pashazadeh"
<pasha@abdnet.com> wrote or quoted :

>I decompiled a class file (using 3 different decompilers for accurancy), and
>every time I got the same source code, there is something amazing in it, it
>has "goto" instruction! how can it be?
>does JVM (as executor) supports "goto" instruction, but no standard compiler
>supports it yet? so a custom compiler can use it?

Yes.  You can't  go crazy with it.  The byte code verifier checks your
loops and jumps to make sure they don't do goofy things with the stack
depth.

A parser might find it more convenient for example to generate byte
code directly where the GOTO is permissible.  GoTo is wicked only in
human programming where the spaghetti you can create confuses fellow
programmers.  Every IF and LOOP is built in byte code of conditional
jumps and gotos.

Fear of GoTo can reach absurd levels.   I crossed horns with a boss
who had a strict no goto rule.  I explained to him that using a goto
in an Eiffel-like macro could save an average of 6 nesting depths per
method in his style of C code, but he rejected it on the same grounds
a Christian rejects gay marriage.

--
Canadian Mind Products, Roedy Green.
Coaching, problem solving, economical contract programming. 
See http://mindprod.com/jgloss/jgloss.html for The Java Glossary.

> Fear of GoTo can reach absurd levels.   I crossed horns with a boss
> who had a strict no goto rule.  I explained to him that using a goto
> in an Eiffel-like macro could save an average of 6 nesting depths per
> method in his style of C code, but he rejected it on the same grounds
> a Christian rejects gay marriage.

Hehehe that last bit is hysterical! :) Goto is obviously evil because
people have told me so for the longest time, I think someone with a
lab coat might even have proved it. I dont fully understand it so
therefore it must be evil and morally wrong. </sarcasm>

     Lets all go fishing maybe

"Amir Pashazadeh" <pasha@abdnet.com> wrote in message
news:bi77c4$5ja9n$1@ID-170682.news.uni-berlin.de...
> Hi everybody,
> I decompiled a class file (using 3 different decompilers for accurancy),
and
> every time I got the same source code, there is something amazing in it,
it
> has "goto" instruction! how can it be?
> does JVM (as executor) supports "goto" instruction, but no standard
compiler
> supports it yet? so a custom compiler can use it?

Well, you haven't actually decompiled anything, you know.  Decompilers
produce source code, not opcodes.  What you have done is translate the
opcodes to their english names, and their arguments as well.

Like many others have said, goto is just a jumping instruction, critical in
both branching, looping, but also in method invocation.  Java does reserve
the goto keyword, although it is not implemented.  Thus, goto is a part of
java, albeit a nominal and useless one.
-- 
Elliott