[Jython-bugs] [ jython-Bugs-1284344 ] __file__ compiled into .py.class file

[SourceForge.net <no...@so...>]

Bugs item #1284344, was opened at 2005-09-07 22:47
Message generated for change (Settings changed) made by fwierzbicki
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=112867&aid=1284344&group_id=12867

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
>Group: Deferred
Status: Open
Resolution: None
>Priority: 4
Submitted By: leouser (leouserz)
Assigned to: Nobody/Anonymous (nobody)
Summary: __file__ compiled into .py.class file

Initial Comment:
I ran across this in trying to figure out why my
jyleo.py startup script was failing.  What is happening
is that once a jython src file is compiled it has the
__file__ attribute permanently  written into the
compiled file.  This makes it impossible to derive
correctly the place where the file is being loaded
from.  Once compiled, you can move a set of jython code
to another machine or a different directory and code
that needs correct __file__ attributes will
mysteriously fail--> that is until you find out that
the __file__ attribute is returning a path that only
exists on another machine.

The requested fix: make the __file__ attribute
correctly reflect its location on the filesystem.

The severity of this depends upon if your
application/script depends upon this information to
function properly.  In my case it does, so from my
perspective it is a *bad* bug.  For others it may not
even be a bug.



----------------------------------------------------------------------

Comment By: James William Pye (jwpye)
Date: 2006-05-14 00:51

Message:
Logged In: YES 
user_id=1044177

I have experienced a __file__ related issue as well. It
seems that this is one of the subtle features that most
Python implementations seem to miss(ironpython seems to be
the only one that got it right, but they seem to be missing
the os module, so I dunno if they are actually "better")

jwp@lit:~/jythonRelease_2_2alpha1 % uname -a
FreeBSD lit 6.1-RC FreeBSD 6.1-RC #5: Mon Apr 24 21:21:02
MST 2006     root@lit:/usr/obj/usr/src/sys/void  amd64
jwp@lit:~/jythonRelease_2_2alpha1 % cat ../printfile.py
print '__file__ =', repr(__file__)
jwp@lit:~/jythonRelease_2_2alpha1 % ./jython ../printfile.py
__file__ =Traceback (innermost last):
  File "../printfile.py", line 1, in ?
NameError: __file__

Also, interestingly, it appears to have printed the first
element in the print statement(printing while iterating over
the list?)

Oh, and it does the same thing when I try to import the
'printfile' module.

If this is not deemed a bug, perhaps it should be mentioned
in the FAQ. Similarly, I can't find much of anything on
ironpython's lack of an os module for posix systems.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=112867&aid=1284344&group_id=12867