installation size

Hi Peter,

the PA size is indeed something we need to address.

This is related to the file content indexing mechanism. Some people in the team might give your more details, but so far as I know, we used to rely on the IFilter system, which in turns relies on the Windows extensible indexing mechanims through light win32 calls.

Now for some reason, it appeared recently that this system does not always work, so we introduced alternatives for pdfs based on iTextSharp and pdfbox, and the later in turns requires IKVM (since it's a Java port). I reckon there's also additional stuff for Open Office docs.

Add the fact that we leverage quite a few Lucene related official extensions (snowball, highlighter, similarity etc.), and that's a bunch of dlls to be shipped in the PA.

We surely need to make that more concise and offer an alternate lighter PA.

For now, I guess you could try to get rid of :

• bc...dll
• Ikvm...dll
• DocumentFormat...
• PDFBox....
• iTextSharp....

If you can get them into the bin folder later, then this should be fine

Otherwise, in theory the module should keep functional and even keep indexing the file content through the IFilter system provided you got the corresponding IFilters (they usually install on the server together with the regular reading applications), but we need to make sure of that.

Sorry for the inconvenience, and thanks for your interest.

There's a lot in our Lucene module, thanks to the several business projects we could throw into it, but there's still some cleaning to be done to package that gracefully.

Also, be aware that we run an unwrapped version of Lucene 2.9.2.2, with many breaking changes we had to take into account from the common 2.0.4 version that you'd usually find in similar products. It means that appart for Peter's DMX, which gracefully wrapped the 2.0.4 version it uses, and similar products, you might have troubles if any other module of yours also uses Lucene.

Cheers

thanks Jesse, I'll give it a go.

Peter