W3C home > Mailing lists > Public > www-jigsaw@w3.org > September to October 1997

Re: VERY IMPORTANT: Bug in getParent()

From: Benoit Mahe <Benoit.Mahe@sophia.inria.fr>
Date: Thu, 30 Oct 1997 17:08:53 +0100
Message-Id: <199710301608.RAA25780@www43.inria.fr>
To: www-jigsaw@w3.org
 sorry I forgot to cc the mailing list :

>Hello all again!
>It is confirmed!  The directory resource (the parent) was getting unloaded
>and then the next access to the child resource would fail with the stack
>trace that you see below related to another problem identified (see below).
>This seems to be a bug to me.

 I think I know what's happen.

 The loaded ResourceStore are puted in an LRUList. Each time a
 resource is accessed, its store is puted at the head of the list.

 The manager start to collect store when there is more than 16 stores
 loaded. So, When the manager try to collect stores (unload), 
 it starts at the tail of the list. In Jigsaw the resources always 
 accept to be unloaded, so we are sure that a resource which has just been 
 accessed will not be unloaded. but, if there is many resources which 
 refuse to be unloaded, this system become not very reliable, because 
 it is possible to unload a resource which has been accessed recently.

 The only reliable solution is to refuse the unloading for all the resources.
 You can disable the StoreManagerSweeper in ResourceStoreManager.java. But
 you will need more memory to run jigsaw (java -ms<number> ie: java -ms16m).

 This bug will be corrected in the next version of Jigsaw (2.0).

- Benoît Mahé -------------------------------------------------------
                      World Wide Web Consortium (W3C)
                    Architecture domain - Jigsaw Team           

  http://www.w3.org/People/Mahe - bmahe@w3.org - + 
Received on Thursday, 30 October 1997 11:09:11 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:41:23 UTC