Hi,
After looking at the source code for the FileIndexBuilder, I'd like to request an additional setting to just Skip Image Meta Data (not Dw's meta data).
Currently the SkipMetaData flag is applied to all meta data, and with large numbers of files the index takes 3 to 4 times more. In our particular scenario, we don't need the image meta data (because we're actually just looking for PDF's but there are thousands of images in there too, so we're taking that hit) but we want Dynamicweb's meta data.
That said, having a whitelist of allowed extensions, would also benefit the time to build the index, because we'd be excluding files that don't matter, thus not spending time doing anything about it. For example, out customers use Mac and import zip files to unzip upon upload. This means we have thousands of .DS_Store files (one for each folder).
Given that we have about 200.000 files (and growing), each instance takes 15min to build. Specially during the development phase, using index builder extenders and having to download the index and use Luke to check instances, our work becomes very time consuming.
Best Regards,
Nuno Aguiar