I'm planning to JIRA this if no one has any counter-arguments.
The current Script implementation collapses script whitespace and eliminates line-endings in scripts - even when within a !CDATA. I believe this is often bad and never good, apart from trivial optimization.
Line-ending preservation is necessary for use of // comments.
Whitespace and line-ending preservation are useful for script compiler error messages that call out a line and column position of the error.
These positions are meaningless if space has been collapsed and line-endings removed.
I happen to be experimenting with using other script languages (especially Groovy) for which line endings are meaningful. You probably don't care about that, though!
The fix is easy: in Script.read(), replace both occurrences of getTextTrim() with getText().
Any thoughts?
-Ed Staub
Yes, one thought... good proposition, file a jira issue for this.. but make sure all tests still run fine.
Even better: try to get cvs access ;-)