Author stefan.richthofer
Recipients darjus, stefan.richthofer, zyasoft
Date 2017-05-07.10:48:44
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <>
So, in a Jython sprint we investigated this issue and found the regression point. It's actually a regression in setup tools:
(added darjus to nosy for this)

Take a look at the discarded method as_header:
It contained the shebang correction that is vital for OSX.

In Jython-side this commit pulled the regression into Jython:

So, how should we fix it? Restore the relevant code in setuptools?
Or maybe better monkeypatch setuptools from Jython, so we have better control over this aspect, wich might be convenient if launchers change over time etc...
E.g. I suppose this problem would go away if we could introduce LiJy-launch as default launcher:
(It is constrained to OpenJDK's license (GPL with classpath exception) because it inherits from OpenJDK C-code. Don't know what trouble it would make to have a stricter license just for the launcher. The launcher is usually not subject to library binding etc.; Also, I will take a look if I can maybe remove or rewrite the relevant code, or maybe we can ask OpenJDK for permission to use the code snippet in question under PSF2 license.)
Date User Action Args
2017-05-07 10:48:45stefan.richthofersetmessageid: <>
2017-05-07 10:48:45stefan.richthofersetrecipients: + stefan.richthofer, zyasoft, darjus
2017-05-07 10:48:45stefan.richthoferlinkissue2570 messages
2017-05-07 10:48:44stefan.richthofercreate