Sarah Eggleston wrote:
>
> Santiago Gala wrote:
>
>> Sarah Eggleston wrote:
>> (...)
>>
>
> If this isn't general interest just shout and we can take it offline.
>
> The XML data isn't stored in cache, I got hold of this by running with
It is. All the JetspeedDiskCacheEntry stuff means it is being taken from the cache.
> no stylesheet and saving the XML result manually. It's basically
> identical to the working XML data (uses the same ISO and DTD file)
> except that the entry-count value (and number of entries) changes.
>
> <?xml version="1.0" encoding='ISO-8859-1'?>
The problem is a bug (my code :( ) The current version of URLFetcher.java (line 353)
checks for an encoding declaration as "encoding\"". This fails to take into account
that you can use
"'" to delimit encoding, and ALSO, that whitespace could be between name, "=" and
value.
The code needs fixing. It needs time to make it right and I'm rather flooded right
now. It should check for
"encoding" optional whitespace "=" optional WS "['|"]" to pick up the value, and
remember which separator
was used to find the end of the value.
You can work around it by using double quotes to separate encoding. I will take a look
to it ASAP.
The error comes because the stream is "marked" to "rewind" it after the
encoding has been read. Dirty, but it was the only way to handle XML
files lying about the encoding in the HTTP Headers. :(
As it takes too long to find a couple of "\"", the mark is not big enough.
--
--------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Search: <http://www.mail-archive.com/jetspeed@list.working-dogs.com/>
List Help?: [EMAIL PROTECTED]
