4Suite / McMillan Installer / py2exe

Thomas Heller theller at python.net
Wed Sep 4 21:41:13 CEST 2002

"Thomas Korb" <Doc at goodweb.de> wrote in message news:doc.m3admxkclw.fsf at goodweb.de...
> Hi! I am using the XSLT-Processor from 4Suite (0.12.0a2), i.e.
>   from Ft.Xml.Xslt.Processor import Processor
>   from Ft.Xml.InputSource import DefaultFactory
>   from Ft.Xml import ReaderException
> On Linux, I can use McMillan's Installer to create a distribution;
> on Windows, it fails (problems with a Math.py-module in the 4Suite
> package). And py2exe complains about missing modules.
> I can overcome the py2exe-problem by importing everything from
> Ft.Xml.Xslt, i.e.
>   import Ft.Xml.Xslt.ApplyImportsElement
>   import Ft.Xml.Xslt.ApplyTemplatesElement
>   etc. etc. (a lot of etc. :-)
> (Too many module to include them on the commmand line with py2exe's
> '-p' option; more than 1024 characters!)
Hm, you can also list them in the setup.cfg file.
Something like this:

And IIRC you can also use (on the command line)
  --includes Ft.Xml.Xslt.*
or even
  --packages Ft.Xml.Xslt

> After that, py2exe still complains about missing modules (some of
> which even do not exist on the system; I think at least some of them
> are Mac-specific!), but it works (as far as I could check up to now).
> But now the program takes quite a time to start (many imports), and

If you don't go the setup.cfg solution, you can probably speed it up
by only importing the modules in a dummy function (which is never
executed). Something like this:
def py2exe_hints():
   import Ft.Xml.Xslt.ApplyImportsElement
   import Ft.Xml.Xslt.ApplyTemplatesElement
   etc. etc. (a lot of etc. :-)

> I actually do not really trust this solution. Moreover, I would have
> to check and test all the above imports with each new version of 4Suite.
> Does someone know how to use 4Suite with the McMillan Installer or
> py2exe on Windows? Any help or hints would be highly appreciated!
I never used 4Suite, so no. I even heard ;-) that py2exe
doesn't work so good with _xmlplus.


More information about the Python-list mailing list