Unknown reporter writes

I have my Recoll configured to examine only PDFs via the mime type in recoll.conf. Also, the search is restricted to only my home directory which contains no external mounts.

When Updating the index, it seems to take just as long as rebuilding it from scratch. Especially, the update hangs on "binary blob" pdfs, ones that are mostly very inefficiently encoded image data.

It seems that Recoll would be checking some sort of modification time against a record of it, instead of digging through each file everytime.

Any ideas on this behaviour? I have attached a file that it has been particularly grinding on (This is the update, not the rebuild. This file has previous been ground upon too).

medoc writes

What recoll version are you using, on what platform ?

Of course recollindex checks a modification date.

Retries are usually caused by indexing errors. We need to check the log file. Please take a look at https://bitbucket.org/medoc/recoll/wiki/WhyIsMyFileNotIndexed for setting up the log file

FYQ, the attached file causes no problem on my test system (ubuntu 14.04, recoll 1.21.3), indexing is instantaneous, and no further processing is done during ulterior indexing passes.