xmlgraphics-fop-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andreas L Delmelle <a_l.delme...@pandora.be>
Subject Re: FOP Memory issues (fwd from fop-users)
Date Wed, 10 Jan 2007 20:00:34 GMT
On Jan 10, 2007, at 20:39, Andreas L Delmelle wrote:

> On Jan 10, 2007, at 04:50, Manuel Mall wrote:
>
>> On Wednesday 10 January 2007 03:42, Andreas L Delmelle wrote:
>>> <snip />
>>> Property lists themselves are no longer alive in the snapshot, it
>>> seems. I don't suppose they are that much of a problem. They are
>>> meant to be in scope for only a very brief interval (best case: from
>>> PropertyListMaker.make() until FObj.bind())
>> Hmm, what are the 80K+ ArrayLists and their backing Object[] then?
>>
>
> That remains the question, of course...
> If it had been the PropertyLists, there would also be 80K instances  
> of StaticPropertyList.

BTW: PropertyLists use plain arrays of Property[] (not ArrayLists)

>
> I'd say the 80K ArrayLists are simply the childNodes lists of all  
> those FObj (TableCells and Blocks), and that those are, in most  
> cases, lists of only one element.

I have read in relation to Java performance tuning that, if you know  
in advance which type of element will be stored in a Collection, it  
almost always pays off to write your own implementation... In this  
case, we could add a FONodeList class, that implements the  
java.util.List interface, and have it backed by a FONode[].

This could at least drastically reduce the number of casts made in  
the fo package.

Hmm, OK, this is such a good idea that I might even get started on it  
right away :)


Cheers,

Andreas


Mime
View raw message