Timeout parameter does not interrupt processing for a specific file

[marekhorst]

Dear Dominika, this is kind of related to #32 issue reported some time ago.

I have just found yet another file blocking CERMINE execution:

https://arxiv.org/pdf/1804.09018.pdf

where setting timeout parameter (#7 feature) does not seem to resolve the problem. As a result of this whole process triggering CERMINE exection is stuck forever.

Could you run some tests and check what could be the reason for timeout not being taken into account?

[marekhorst]

For the record: we are relying on most recent cermine release: 1.13.

[marekhorst]

Let me add few more details:

• I was able to confirm this issue by running metadata extraction locally on my laptop (also on the most recent cermine-impl-1.14-20180204.213009-17 version). Setting timeout to 5 secs or more results in advancing into the stage when processing could not be interrupted.
• we have images extraction feature already disabled, as recommended in https://github.com/CeON/CERMINE/issues/58#issuecomment-362945422
• it seems the problem is related to pretty complex figures on pages 19-20, at least one of the PDF viewers is also choking when loading those pages