jython-users

[Jython-users] Supporting + on Java objects

[Anthony E. <ae...@si...>]

Currently, in Jython, using any arbitrary Java object aside from a java.lang.String, the following fails:

  print "text" + obj

With the message: TypeError: __add__ nor __radd__ defined for these operands

But the following works:

  print "test", obj


Would the Jython implementers consider modifying the PyJavaInstance class with something similar to:

    public PyObject __add__(PyObject object) {
    	return __str__ + object.__str__();
    }

So that the + notation as used in Java works properly?  I am not sure if the code I provided above is correct as I am
not intimately familiar with the Jython source.  I am also not sure if this has been discussed before: the mailing list
archives provided at the Jython.org site do not have search facilities (AFAICT) and I was unable to find anything about
it in my mail application (back to 7/21/2001), so I apologize ahead of time if it has.

Sincerely,
Anthony Eden

Re: [Jython-users] Supporting + on Java objects

[Kevin B. <kb...@ca...>]

I'd be hesitant to do this.

A couple of reasons:

- In general, the Python string formatting idioms are preferred:
Instead of:
	print "text " + obj + ", " + obj2 + "/" + obj3
use:
	print "text %s, %s/%s" % (obj, obj2, obj3)
or:
	print "text %(obj)s, %(obj2)s/%(obj3)s" % vars()

- The proposed enhancement for Java instances would cause Python instances & java instances to behave differently:
	t = 5
	s = Object()
	print "text" + t + " " + s ## error

kb

Anthony Eden wrote:
> Currently, in Jython, using any arbitrary Java object aside from a java.lang.String, the following fails:
> 
>   print "text" + obj
> 
> With the message: TypeError: __add__ nor __radd__ defined for these operands
> 
> But the following works:
> 
>   print "test", obj
> 
> 
> Would the Jython implementers consider modifying the PyJavaInstance class with something similar to:
> 
>     public PyObject __add__(PyObject object) {
>     	return __str__ + object.__str__();
>     }
> 
> So that the + notation as used in Java works properly?  I am not sure if the code I provided above is correct as I am
> not intimately familiar with the Jython source.  I am also not sure if this has been discussed before: the mailing list
> archives provided at the Jython.org site do not have search facilities (AFAICT) and I was unable to find anything about
> it in my mail application (back to 7/21/2001), so I apologize ahead of time if it has.
> 
> Sincerely,
> Anthony Eden
> 
> 
> _______________________________________________
> Jython-users mailing list
> Jyt...@li...
> https://lists.sourceforge.net/lists/listinfo/jython-users

Re: [Jython-users] Supporting + on Java objects

[Samuele P. <pe...@in...>]

From: Kevin Butler <kb...@ca...>

> - The proposed enhancement for Java instances would cause Python instances &
java instances to behave differently:
> t = 5
> s = Object()
> print "text" + t + " " + s ## error
>

Exactly, that's the point, as long as Jython is an implementation
of Python <wink> , the behavior + and strings will not change, sorry.

regards.

RE: [Jython-users] Supporting + on Java objects

[Anthony E. <ae...@si...>]

No need to be sorry.  Thanks to all who have answered this to clear this up and for demonstrating the proper way to
handle String concatenation in Jython.  Perhaps this should be added to the FAQ. ;-)  For Java programmers getting into
Python via Jython, the behavior is not evident.

Sincerely,
Anthony Eden

> -----Original Message-----
> From: Samuele Pedroni [mailto:pe...@in...]
> Sent: Tuesday, March 26, 2002 2:13 PM
> To: kb...@ca...; Anthony Eden
> Cc: jyt...@li...
> Subject: Re: [Jython-users] Supporting + on Java objects
>
>
> From: Kevin Butler <kb...@ca...>
>
> > - The proposed enhancement for Java instances would cause Python instances &
> java instances to behave differently:
> > t = 5
> > s = Object()
> > print "text" + t + " " + s ## error
> >
>
> Exactly, that's the point, as long as Jython is an implementation
> of Python <wink> , the behavior + and strings will not change, sorry.
>
> regards.
>
>

RE: [Jython-users] Supporting + on Java objects

[Doug L. <lan...@go...>]

At 2:24 PM -0500 3/26/02, Anthony Eden wrote:
>No need to be sorry.  Thanks to all who have answered this to clear this up and for demonstrating the proper way to
>handle String concatenation in Jython.  Perhaps this should be added to the FAQ. ;-)  For Java programmers getting into
>Python via Jython, the behavior is not evident.

All such programmers should read Tim Peters' "Zen of Python" at
least once a month until it sinks in...

  http://www.python.org/doc/Humor.html#zen

It's short enough to include here in its entirety (see below).

Brad Cox wrote:
> The token (+ vs , vs nothing) isn't important.  What *is* is
> that the result winds up as a string, not array, tuple, etc.

 ... which implies that there is a possible ambiguity.

So, for Jython to guess what you meant, it would have to
break *two* of the Zen suggestions:

   Explicit is better than implicit.
   In the face of ambiguity, refuse the temptation to guess.

I think one of these at a time is occasionally broken for a good
enough cause, but two?   Let's just say I don't think it'll happen
soon.  The workaround is painless enough -- and more explicit.

  -- Doug


PS, here's Tim's text:

The Zen of Python (by Tim Peters)

Beautiful is better than ugly.
Explicit is better than implicit.
Simple is better than complex.
Complex is better than complicated.
Flat is better than nested.
Sparse is better than dense.
Readability counts.
Special cases aren't special enough to break the rules.
Although practicality beats purity.
Errors should never pass silently.
Unless explicitly silenced.
In the face of ambiguity, refuse the temptation to guess.
There should be one-- and preferably only one --obvious way to do it.
Although that way may not be obvious at first unless you're Dutch.
Now is better than never.
Although never is often better than *right* now.
If the implementation is hard to explain, it's a bad idea.
If the implementation is easy to explain, it may be a good idea.
Namespaces are one honking great idea -- let's do more of those!

RE: [Jython-users] Supporting + on Java objects

[Anthony E. <ae...@si...>]

FWIW, the text below would make a good FAQ entry for Java programmers coming to the world of Python wondering about how
Strings work in Python.  Kevin's explanation was right on track and answered the question succinctly.  I am sure that
String concatenation is covered in the Python docs, but I still think it would be good to cover this in the Jython docs,
primarily because a lot of Java code transfers very easily to Python but some of it (such as for(x in y) and String
handling) do not.

Perhaps there are other stumbling blocks which Java developers coming to Jython will experience which could also be
documented better on the Jython site.  What do you think?

Sincerely,
Anthony Eden

> -----Original Message-----
> From: jyt...@li...
> [mailto:jyt...@li...]On Behalf Of Kevin
> Butler
> Sent: Tuesday, March 26, 2002 1:53 PM
> To: Anthony Eden
> Cc: jyt...@li...
> Subject: Re: [Jython-users] Supporting + on Java objects
>
>

<snip>

>
> - In general, the Python string formatting idioms are preferred:
> Instead of:
> 	print "text " + obj + ", " + obj2 + "/" + obj3
> use:
> 	print "text %s, %s/%s" % (obj, obj2, obj3)
> or:
> 	print "text %(obj)s, %(obj2)s/%(obj3)s" % vars()

>

Re: [Jython-users] Supporting + on Java objects

[Brad C. <bc...@vi...>]

As the one that asked Anythony about this (mis?)feature on the 
jpublish list, I'd like to add my support to his request. I'd have 
done it directly except that I'm so new to python that I'd thought 
this was my fault.

PS: The token (+ vs , vs nothing) isn't important. What *is* is that 
the result winds up as a string, not array, tuple, etc.

PPS: Probably harder, supporting expansion of variables within 
strings ala' perl would immediately make python a favorite. The 
${variable}s workaround that was offered here earlier is a bit 
obscure for heavily used stuff like expanding variable information 
into html template files.

See http://virtualschool.edu/jwaa under Mls (multi line 
preprocessor), velocity, or JSP/ASP for workarounds to the same 
problem in Java. This really "should" be supported by the language, 
not outside as in mls, velocity, or jsp. Like this:

title = "my title"
print '''
<html>
	<head>
		<title>$title</title>
	</head>
</html>
'''

Ideally, this should work for any expression (variables, subroutine 
calls, arrays, tuples, etc). As I recall, perl didn't do all of these 
right. Don't recall exactly; I haven't used perl in some years.

At 1:23 PM -0500 3/26/02, Anthony Eden wrote:
>Currently, in Jython, using any arbitrary Java object aside from a 
>java.lang.String, the following fails:
>
>   print "text" + obj
>
>With the message: TypeError: __add__ nor __radd__ defined for these operands
>
>But the following works:
>
>   print "test", obj
>
>Would the Jython implementers consider modifying the PyJavaInstance 
>class with something similar to:
>
>     public PyObject __add__(PyObject object) {
>    	return __str__ + object.__str__();
>     }
>
>So that the + notation as used in Java works properly?  I am not 
>sure if the code I provided above is correct as I am
>not intimately familiar with the Jython source.  I am also not sure 
>if this has been discussed before: the mailing list
>archives provided at the Jython.org site do not have search 
>facilities (AFAICT) and I was unable to find anything about
>it in my mail application (back to 7/21/2001), so I apologize ahead 
>of time if it has.
>
>Sincerely,
>Anthony Eden
>
>
>_______________________________________________
>Jython-users mailing list
>Jyt...@li...
>https://lists.sourceforge.net/lists/listinfo/jython-users


-- 
Brad Cox, PhD; bc...@vi... 703 361 4751
o For industrial age goods there were checks and credit cards.
    For everything else there is http://virtualschool.edu/mybank
o Java Interactive Learning Environment http://virtualschool.edu/jile
o Java Web Application Architecture: http://virtualschool.edu/jwaa

Re: [Jython-users] Supporting + on Java objects

[Samuele P. <pe...@in...>]

From: Brad Cox <bc...@vi...>
> 
> PPS: Probably harder, supporting expansion of variables within 
> strings ala' perl would immediately make python a favorite. The 
> ${variable}s workaround that was offered here earlier is a bit 
> obscure for heavily used stuff like expanding variable information 
> into html template files.
> 

Incredible as it may seem <wink>, there have been serious 
discussions about this on python-dev, and there's even a PEP

http://python.sourceforge.net/peps/pep-0215.html

But for the moment the discsussions
 have not converged anywhere,
the main problem being how to disentangle
the confusion wrt to the old idioms (deprecate them
(hard because of legacy),
or organzize things in a hierarchy of increasing 
functionality but how) and the syntax.

Don't hold your breath (see 
http://www.python.org/doc/essays/pepparade.html
), anyway we cannot
support this indipendently from Python.

regards.

Re: [Jython-users] Supporting + on Java objects

[Kevin B. <kb...@ca...>]

Brad Cox wrote:

 > PS: The token (+ vs , vs nothing) isn't important. What *is* is that
 > the result winds up as a string, not array, tuple, etc.

The string formatting always yields a string.

> PPS: Probably harder, supporting expansion of variables within strings 
> ala' perl would immediately make python a favorite. The ${variable}s 
> workaround that was offered here earlier is a bit obscure for heavily 
> used stuff like expanding variable information into html template files.

I'm not sure how string formatting is "obscure" - it is described
in the python tutorial:
http://www.python.org/doc/current/tut/node9.html#SECTION009100000000000000000

and it is heavily used in Python code (cd jython/Lib; grep \% *py if you'd like
a lot of examples...).

Also, it is likely to be faster than '+' would be...

BTW, watch your characters: percent and parens, not dollar and braces

> Ideally, this should work for any expression (variables, subroutine 
> calls, arrays, tuples, etc). 

Ahh, you want a formatter that evaluates expressions?

Not hard to make one.

Since named-value % formatting is based on dictionaries, you can replace the dictionary
with your own class with a "__getitem__" function. This lets you completely replace
the % evaluation with whatever you'd like:

--- formatter.py ---
import operator

class Formatter:
     def __init__( self, locals=None ):
         self.locals = locals or globals()

     def __getitem__( self, name ):
         return eval( name, self.locals )

asdf=10

template = """asdf=%(asdf)s
simple expression=%(20+30)s
complex expression=%(reduce(operator.add, range( 10 )))s
hairier than you want: %(Template( {"one":1, "two":4, "four":16, "4":16} ).format( "Square of four is %(four)s, but the digit 4 is %(4)s" ))s
"""

print template % Formatter()

----
output is:

 >>> ## working on region in file d:/TEMP/python-3181zD...
asdf=10
simple expression=50
complex expression=45
hairier than you want: Square of four is 16, but the digit 4 is 4

 >>>

Gosh, but I love Python.  :-)

Bonus points if you figure out why the outputs of four & 4 are different.  :-)

kb

Re: [Jython-users] Supporting + on Java objects

[Samuele P. <pe...@in...>]

From: Kevin Butler <kb...@ca...>
> 
> print template % Formatter()
> 

You can use Formatter also in this way:

import sys

def i(str):
  uplocals = sys._getframe().f_back.f_locals
  return str % Formatter(locals = uplocals)

print i(template)

As taste goes YMMV. 

regards.

Re: [Jython-users] Supporting + on Java objects

[Kevin B. <kb...@ca...>]

I wrote:

 > hairier than you want: %(Template( {"one":1, "two":4, "four":16, "4":16}
 > ).format( "Square of four is %(four)s, but the digit 4 is %(4)s" ))s
 > """

Woops!  Just realized I included code using my original Template class name & format method.
Gotta remember to run all examples in a new interpreter before posting.  :-)

This example will work better:

import operator

class Formatter:
     def __init__( self, locals=None ):
         self.locals = locals or globals()

     def __getitem__( self, name ):
         return eval( name, self.locals )

asdf=10

template = """asdf=%(asdf)s
simple expression=%(20+30)s
complex expression=%(reduce(operator.add, range( 10 )))s
hairier than you want: %("Square of four is %(four)s, but the digit 4 is %(4)s" % Formatter( {"one":1, "two":4, "four":16, "4":16} ))s
"""

print template % Formatter()

Output:

 >>> ## working on region in file d:/TEMP/python-318cSW...
asdf=10
simple expression=50
complex expression=45
hairier than you want: Square of four is 16, but the digit 4 is 4

 >>>

(I should also point out that eval may not be blindingly fast, and there
are security concerns in evaluating any user-provided expressions...)

kb

Re: [Jython-users] Supporting + on Java objects

[Brad C. <bc...@vi...>]

Thanks to all who replied. Will see if Kevin's suggestion floats my 
boat. If not, I'll wheel out the MLS preprocessor again.

PS: MLS addresses the syntax issues in Java by using {{diggraphs}} as 
syntactic sugar for "string". So...

	print {{
<html><title>{{anyExpressionWhatSoEver}}</title></html>
}};

expands into precisely this
	print "\n"+
<html><title>"+anyExpressionWhatSoEver+"</title></html>\n"+
"";

Even-level nestings (including zeroth) expand into strings while 
odd-level nestings expand as the original text. A simple solution to 
a common problem (FAR simpler than JSP or Velocity) that doesn't 
involve language modifications at all.
-- 
Brad Cox, PhD; bc...@vi... 703 361 4751
o For industrial age goods there were checks and credit cards.
    For everything else there is http://virtualschool.edu/mybank
o Java Interactive Learning Environment http://virtualschool.edu/jile
o Java Web Application Architecture: http://virtualschool.edu/jwaa