| astippich |
| Frameworks |
Currently, both XML and SVG documents are indexed as plain text due
to mimetype inheritance. This fills the content index with meaningless
data (tags, attributes, attribute values ...).
Use QDomElement::text() for generic XML documents and <text/> nodes
for SVG to extract the content. Also try do find Dublin Core metadata
and add the relevant properties.
Depends on D16488
| Automatic diff as part of commit; lint not applicable. |
| Automatic diff as part of commit; unit tests not applicable. |
| src/extractors/xmlextractor.cpp | ||
|---|---|---|
| 25 | seems unused | |
| 28 | same | |
| 72 | Can we agree on using a static qstringlist if the mimetypes are fixed? I prepared a patch to convert all other extractors to this scheme when applicable in D16554 Also, using an initializer list is faster, see https://www.angrycane.com.br/en/2018/06/19/speeding-up-cornercases/ | |
Reorder mimetype check to use SVG(+XML) special case if applicable and use
generic implementation for all other mimetypes inheriting from
application/xml.
Use static QStringList for mimetypes.
| src/extractors/xmlextractor.cpp | ||
|---|---|---|
| 25 | I think that gets pulled in via "kfilemetadata_debug.h", but nevermind | |