[IronPython] IronPython 2.7.1 & 3.x Development
vernondcole at gmail.com
Tue Mar 15 13:28:40 PDT 2011
IIRC you are at pycon. (I am at the wrong end of America and Mark is at the
wrong end of Earth. I couldn't make it.)
Can you create/find a bof session to hash out a "standard" shebang? It
should be something good enough that 2to3 could have a fixer for it.
[ reminder for the day:
Now is better than never.]
Were you thinking part of the pywin32 distribution, or CPython itself?
Either way we should have a compatible/identical tool installed with
IronPython. The thing should be Frankensteinian enough that I can have it
"cling" to my favorite distribution even when I install several others
beside it. (Perhaps an environment variable or .python file?)
What about Jython? I have been using Jython 2.5 lately (not my fault -- a
college class is making me learn Java after avoiding it for 20+ years -- but
I installed the Jython support in NetBeans, too) and discover that they are
having similar problems with how to crank up a Jython script from the
command line. The "java -jar ..." string is nasty. This starter program
could be a triple solution. Too much?
P.S. -- and when I double-click on a file named "setup.py" automatically add
"install" to the generated command line.
Tue, Mar 15, 2011 at 8:28 AM, Michael Foord <fuzzyman at voidspace.org.uk>wrote:
> On 15/03/2011 07:18, Vernon Cole wrote:
> A shebang on the first line is the usual way of specifying which scripting
> engine to use on a posix system. If introspection were to be used to select
> between python versions, that would be the most obvious way to do it.
> Windows (tm) prefers to use filename extensions, which is why .pyw is used
> to specify a GUI program. The command shell does not open the file to see
> what's inside AFAIK.
> It was suggested in the past that .ipy might be used to specify IronPython
> programs. If that idea were carried to its logical end, one would use .ipy3w
> to specify an IronPython 3.x GUI script. That is a bit frightening.
> What about the idea of a semi-standard python dispatch program which lives
> on the search path, introspects the script, and calls the appropriate
> engine? Perhaps it could even manage to capture tracebacks sent to stderr
> when a script dies?
> There has been a long discussion recently on Python-dev  suggesting the
> creation of exactly such a launcher program for Python on Windows. Mark
> Hammond (copied) said he would implement it. Including IronPython support is
> a fine idea.
> All the best,
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Users