Description Vlad Klicnik 2001-10-10 22:50:21 EDT

We should ensure that we do not use String.getBytes() anywhere in the core except where we 
call the core dll to do file work.  In general, marshaling of strings etc should be done using UTF8
(readUTF and writeUTF) as this will preserve the data in the face of locale changes.

NOTES:

DJH (8/13/01 12:11:35 PM)
	In boot.update, are these problems?
		LogStore.save(Log, URL)
		XmlLiteStore.save(XmlLite, URL)

RTP (8/15/01 4:40:28 PM)
	I suspect this code belongs to update as well:

	org.eclipse.core.internal.boot.LaunchInfo.checkUpdateEnabled() {
	...
			byte[] b = EOF.getBytes();   <<<<<------
			uos = new FileOutputStream(f);
			uos.write(b);
	...
	}


LC (8/16/01 4:54:29 PM)
		- XmlLiteStore.save(XmlLite, URL) is never called
		- LogStore.save(Log, URL)	writes out a log that is not read in by the UM again, it is only 
			meant as	a record of what has taken place.  The user can read it (with something like notepad) 
			if he/she's interested.  The translated msgs in this log file is written in native format.   I think 
			it is appropriate to leave this unchanged.
		- LaunchInfo.checkUpdateEnabled() checks if the file system is r/w.   The file written contains no
			useful info as far as i know.

RTP (8/23/01 2:51:07 PM)
	Also, LaunchInfo.chekpoint() writes to disk using PrintWriter and the default encoding.
Not sure if this needs to change or not.

VK (8/28/01 2:21:41 PM)
	The temp files written by the update fns are part of the local install tree (not part of the workbench that can be
	exchanged) so local encoding is appropriate. Moving this PR to InBox to see if we want to change to use
	UTF-8 (for consistency) as part of the 2.0 rework.