This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.

Bug 16513

Summary:	Filesystems: Filesystems are ordered after moumting new one
Product:	platform	Reporter:	Marian Mirilovic <mmirilovic>
Component:	-- Other --	Assignee:	Vitezslav Stejskal <vstejskal>
Status:	CLOSED FIXED
Severity:	blocker	CC:	jzajicek
Priority:	P1
Version:	3.x
Hardware:	All
OS:	All
Issue Type:	DEFECT	Exception Reporter:
Attachments:	ide.log

Description Marian Mirilovic 2001-10-12 13:25:36 UTC

[nb_dev](20011012),[jdk1.3.1](rc1)

Steps to reproduce:
- run IDE ( now is mounted sampledir )
- mount some jar ( xx.jar )
- mount some Local directory ( some/directory )
-> now Filesystems ordered and order is :
	1. sampledir
	2. some/directory
	3. xx.jar

If you had expanded some filesystem, that one is collapsed.

Comment 1 Jan Zajicek 2001-10-12 13:38:55 UTC

Adding log for following scenario:

1. start 3.3
2. mount javacvs
3. mount jar
4. mount localdirectory

Final order:
1.localdirectory
2.jar
3.cvs

Comment 2 Jan Zajicek 2001-10-12 13:41:05 UTC

Created attachment 2988 [details]
ide.log

Comment 3 Jaroslav Tulach 2001-10-12 14:41:49 UTC

Not reproducible without projects module.

Comment 4 Jan Chalupa 2001-10-12 15:27:16 UTC

Affects classpath -> P1.

Comment 5 Vitezslav Stejskal 2001-10-16 19:15:12 UTC

This problem was discovered when I have added FilterFileSystem 
implementation into the core and used it in SessionManager as a 
project layer. The FFS delegates to another filesystem and uses some 
FileObject from that filesystem as a root point. All FOs accessible on 
the FFS are shifted in comparision to original FOs on the delegate 
filesystem.

Current implementation of MultiFileObject.getAttribute doesn't expect 
this and uses precomputed path when asking delegates. Precomputed 
path isn't valid for delegates from the shiftig file system and thus 
no attribute is found. I have change it to ask delegates without 
passing precomputed path, thus each delegate computes its own path 
again now. It fixes functionality but causes performance regression. I 
have filed PERFORMANCE issue #16618 against the filesystems in order 
to measure this regression and correct it.

Comment 6 Marian Mirilovic 2001-10-17 09:25:27 UTC

verified in [nb_dev](20011017)

Comment 7 Quality Engineering 2003-07-01 16:50:23 UTC

Resolved for 3.4.x or earlier, no new info since then -> closing.