[IronPython] Python Command Line Compiler issues and questions

[Michael Foord]

Dino Viehland wrote:
> I am surprised one is behaving different.  It works for me w/ pyc if I keep the package directory on the path, but if I don't then it breaks.
> 
> I'm curious did you include all of your individual package files, or just __init__.py?  And were the packages on your path?
> 

It was a colleague of mine who tried this, but another user (in this 
thread) did report the same problem.

Michael
http://www.voidspace.org.uk/ironpython/index.shtml

> -----Original Message-----
> From: users-bounces at lists.ironpython.com [mailto:users-bounces at lists.ironpython.com] On Behalf Of Michael Foord
> Sent: Monday, January 01, 2007 8:20 AM
> To: Discussion of IronPython
> Subject: Re: [IronPython] Python Command Line Compiler issues and questions
> 
> Mark Rees wrote:
>> I have a simple IronPython script that will compile an IronPython
>> executable
>> with multiple source files
>>
>> http://hex-dump.googlecode.com/svn/trunk/ironpython-gdata-tutorial/tut03/makeexe.py
> 
>> and to handle the "if __name__ == '__main__':" idom you just need to change
>> it to:
>>
>> if __name__ == '__main__' or __name__ == sys.executable:
>>
>> It's discussed in this post:
>> http://hex-dump.blogspot.com/2006/08/deploying-gdata-reader-as-executable.html
>>
> 
> Ahh... cool.
> 
> I've had a play with makeexe.py, it's very nice.
> 
> You can compile multiple Python files (including imports) into a single
> executable.
> 
> The difficulty we had was with maintaining our package structure. We
> want import semantics like 'from Main.Something import Something', which
>   didn't seem to work with PYC.
> 
> It *does* seem to work (initial experiments) with makeexe.py, which is
> cool. (I included the empty 'package/__init__.py' file in the arguments
> to makeexe.py.)
> 
> We were intending to ship source code with our application (which has
> various advantages for us - our target market includes programmers in
> business environments), but we have an issue with startup time - mainly
> caused by the parsing of the Python files. Precompilation could solve that.
> 
> Hmmm....
> 
> Michael Foord
> http://www.voidspace.org.uk/python/index.shtml
> 
>> Mark
>>
>>
>>
>> On 1/1/07, Michael Foord <fuzzyman at voidspace.org.uk> wrote:
>>> Patrick O'Brien wrote:
>>>> If you compile a python module using PYC and run the resulting
>>> executable,
>>>> the module name is no longer '__main__' and so you cannot make use of
>>> the
>>>> popular idiom:
>>>>
>>>> if __name__ == '__main__':
>>>>
>>>> If you don't use that idiom it makes it impossible to unit test any
>>>> functions in your main module because running the unit tests imports
>>> the
>>>> module and runs the main code.  The alternative is to have the main
>>> module
>>>> be as small as possible and put the bulk of the functionality into
>>> other
>>>> modules that get imported by the main module.  Which would be fine if
>>> there
>>>> was a way to include those modules in the resulting exe created by PYC,
>>> but
>>>> that doesn't seem to happen for me.  Which means the .py files would
>>>> have to
>>>> ship as well as the .exe and the two .dll files (IronMath.dll and
>>>> IronPython.dll).  And substituting a .pyo file does not work, so what
>>> would
>>>> ship in my app is the source code and for this particular app that
>>> would
>>>> not
>>>> be a good thing.
>>>>
>>>> The PYC program lets you specify multiple .py files, but I don't
>>> understand
>>>> what that means.  One file can be marked as main, which is what gets
>>>> executed.  What I don't understand is how the other .py files get used.
>>>> The
>>>> one example of this in the documentation was not terribly helpful for
>>> me.
>>>> Anyone have any suggestions?  Did I miss something in my understanding
>>> of
>>>> PYC?  How are others packaging their IronPython apps for
>>> distribution to
>>>> customers?
>>>>
>>> We asked about IronPython supporting compiling Python modules so that
>>> they can be imported from, and were told that it probably wouldn't make
>>> it into IronPython 1.X.
>>>
>>> That means (my understanding) that PYC can only be used for compiling a
>>> single script with no dependencies (unless you're happy about shipping
>>> those dependencies as source files).
>>>
>>> Luckily for us at Resolver, I think we're happy to ship source files. If
>>> you're not happy with this you'll need to resort to other means for
>>> obfuscation (like encryption of source where the decryption is in the
>>> compiled part). This will confound casual observers, but IL is easily
>>> (relatively) decompiled and so will not stop the determined attacker
>>> from breaking your encryption.
>>>
>>> Michael
>>> http://www.voidspace.org.uk/python/articles.shtml
>>>
>>>>
>>> ------------------------------------------------------------------------
>>>> _______________________________________________
>>>> users mailing list
>>>> users at lists.ironpython.com
>>>> http://lists.ironpython.com/listinfo.cgi/users-ironpython.com
>>>>
>>>>
>>>>
>>> ------------------------------------------------------------------------
>>>> No virus found in this incoming message.
>>>> Checked by AVG Free Edition.
>>>> Version: 7.1.409 / Virus Database: 268.16.0/610 - Release Date:
>>> 30/12/2006
>>>
>>> _______________________________________________
>>> users mailing list
>>> users at lists.ironpython.com
>>> http://lists.ironpython.com/listinfo.cgi/users-ironpython.com
>>>
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> users mailing list
>> users at lists.ironpython.com
>> http://lists.ironpython.com/listinfo.cgi/users-ironpython.com
>>
>>
>> ------------------------------------------------------------------------
>>
>> No virus found in this incoming message.
>> Checked by AVG Free Edition.
>> Version: 7.1.409 / Virus Database: 268.16.1/611 - Release Date: 31/12/2006
> 
> _______________________________________________
> users mailing list
> users at lists.ironpython.com
> http://lists.ironpython.com/listinfo.cgi/users-ironpython.com
> _______________________________________________
> users mailing list
> users at lists.ironpython.com
> http://lists.ironpython.com/listinfo.cgi/users-ironpython.com
> 
>