Bug 132576

Summary:	Default package has no subpackages
Product:	[Eclipse Project] JDT	Reporter:	Markus Keller <markus.kell.r>
Component:	Core	Assignee:	Jerome Lanneluc <jerome_lanneluc>
Status:	RESOLVED WONTFIX	QA Contact:
Severity:	normal
Priority:	P3
Version:	3.2
Target Milestone:	3.3 RC4
Hardware:	PC
OS:	Windows XP
Whiteboard:
Bug Depends on:
Bug Blocks:	125486

Description Markus Keller

2006-03-20 13:38:01 EST

I20060315-1200

The IPackageFragment of a default package should always answer false to #hasSubpackages(). See JLS3 7.4.2: "Note that an unnamed package cannot have subpackages, since the syntax of a package declaration always includes a reference to a named top level package."

Currently, it returns true if the package fragment root has packages other than the default.

Comment 1 Jerome Lanneluc

2007-06-21 11:44:55 EDT

Unfortunately, the spec for IPackageFragment#hasSubpackages() is slightly different. It says that it "returns whether this package fragment's name is a prefix of other package fragments in this package fragment's root".
Since the name of the default package fragment is "", it is always a prefix of any other package fragment's name (if any).
This cannot be changed without breaking the IPackageFragment's spec.

Comment 2 Victor Toni

2007-06-21 12:03:35 EDT

(In reply to comment #1)
> Since the name of the default package fragment is "", it is always a prefix of
> any other package fragment's name (if any).

I disagree since packages have a separator which is ".". If other packages should be treated as subpackages of "" their names should be something like ".my.own.subpackage" which is not a valid package name.

Comment 3 Jerome Lanneluc

2007-06-21 12:12:34 EDT

(In reply to comment #2)
> I disagree since packages have a separator which is ".". If other packages
> should be treated as subpackages of "" their names should be something like
> ".my.own.subpackage" which is not a valid package name.
Yes, this is what JLS3 7.4.2 says. However IPackageFragment#hasSubpackages() is not compatible with JLS3 7.4.2 (since it is using the word 'prefix' and an empty string is always a prefix of any string).
So if you wish, you can enter a feature request to ask for a new API that corresponds to JLS3 7.4.2.