Disable automatic type conversion in order to continuously calling java method?

[passerbythesun]

I would like to do some string processing in a continuous way, e.g., java.lang.String(" some string").trim().replaceAll("regex", "replacement"), so that re module is not needed and code consistency with java side is better.

But I find that trim() returns unicode type, which result in an AttributeError: 'unicode' object has no attribute 'replaceAll' error.

To my understanding, Javatrim() returns java.lang.String type, so this should be the effort of automatic type conversion of java string to python unicode. Is it right?

If so, how to disable automatic type conversion in this situation? Or if not, is there another way to do so?

[Thrameos]

There appears to be a bug in version 0.6.2. There is supposed to be an option that allows you to select if you want conversion or not. However, no mater which way I tried the switch it gives a buggy behavior.

import jpype
import jpype.imports
import _jpype

jpype.startJVM(jpype.getDefaultJVMPath())
_jpype.setConvertStringObjects(True)  # True gives unicode, False gives JString

from java.lang import String
f=String('foobar').trim().replaceAll('bar','foo')
print(f)

Unfortunately JString is only useful to pass to a java method and doesn't have any String methods exposed.

Thus the best I can recommend is either use another string cast to put the wrapper back as a string
f=String(String('foobar').trim()).replaceAll('bar','foo') or use the devel branch which we will release as 0.7.0 shortly.

[passerbythesun]

@Thrameos Thanks for answering. I just used the way you recommend. Waiting for 0.7.0~

[Thrameos]

Fixed and tested in 0.7.