Re: [Gzz] STORM indexing

[Tuomas Lukka]

On Tue, May 13, 2003 at 01:15:51PM +0300, Tuukka Hastrup wrote:
> 
> The STORM peer can't start if I try to publish a pool with a text/html 
> block that has an A element whose content includes newlines:
> 
> Exception in thread "main" java.lang.UnsupportedOperationException: values 
> containing newlines
>         at org.nongnu.storm.impl.DirDB.add(DirDB.java:127)
>         at org.nongnu.storm.impl.DirPool.added(DirPool.java:213)
>         at org.nongnu.storm.impl.DirPool.getDB(DirPool.java:179)
>         at org.nongnu.storm.impl.DirPool.added(DirPool.java:213)
>         at org.nongnu.storm.impl.DirPool.getDB(DirPool.java:179)
>         at 
> org.nongnu.storm.impl.AbstractPool.initializeIndices(AbstractPool.java:76)
>         at org.nongnu.storm.impl.DirPool.<init>(DirPool.java:115)
>         at org.nongnu.storm.modules.gispmap.GispPeer.main(GispPeer.java:66)
> 
> 
> It seems that for each text/html block, we try to store a list of blocks 
> referring to it - and some kind of description. Should something simply 
> remove newlines from the A tag content? 

Have you been able to make a failing test for this?

> Further, could I change the start of backlink list on the resulting web 
> pages to read "This page is referred to as:" instead of "Backlinks:". For 
> example:
> This page is referred to as:
> [the previous version]
> [The official Fenfire homepage]

Sounds good to me.

        Tuomas