3 Jan
2009
3 Jan
'09
8:09 p.m.
Hi Vincent,
Ok.
Not only the bridge but also the document name (against which the cleaning
is done) will need to be provided for this filter. But since the signature
of a filter is clean(Document dom), there is no other way to pass additional
information (like the document name) that is required for determining the
attachment URL. I have another way to get around this problem with
cleaningParameters we introduced to the HTMLCleaner (we can pass the
document against which the cleaning is done). Will work on it.
BTW not sure why you removed the Transformer components.
Transformers are specific to office importer. They won't be used by any
other component outside office importer. So I thought there is no point in
keeping them as components.
Thanks.
- Asiri
I'm not sure how to do this.
My comment was about the mixing, not about the need for the Document
Access bridge.
Unlike other filters, ImageFilter needs to have
access to the original XWikiDocument (via DocumentAccessBridge)
other than
the w3c DOM. This is because when cleaning image links, the src
attribute
need to be replaced by the corresponding attachment url. This is why I
introduced XWikiHTMLCleaner which takes in the OfficeImporterContext
in
addition to the w3c dom.
Make them components and have the bridge injected.