Closed GerHobbelt closed 5 years ago
Suspected to be related to #96 / #98 : now that we have taken out Sorax for PDF page count calculus (SHA-1: fb775d57b7e29d3726ad52fbbdf7f1a411de005b) and introduced the Three Strikes Heuristic for Badly Behaving PDF Documents (SHA-1: e353006a1795cfe1ceef42b61db5dfc79b10a491) Qiqqa stibility appears to be much improved, at least for very large libraries such as the one tested here (41K+ articles and growing as I merge the various historical versions of this library as those versions were left to rot, waiting for better times as they invariably caused Qiqqa to crash in spectacular ways, e.g. blowing up its Lucene index and render it non-working, crash on application startup, etc.)
Closing and decluttering the issue list so it stays workable for me: fixed in https://github.com/GerHobbelt/qiqqa-open-source mainline=master branch, i.e. v82pre3 experimental release and beyond.
Here's the Qiqqa log file from the time of crash - the Sniffer session had been running for about 2 hours straight.
In the background I'ld triggered a global OCR+indexing action by opening Expedition and hitting Refresh just about, what was it, 15 minutes before the crash? This of course loaded the Intel i7 machine to the max (as intended for the test), while I also downloaded a few PDFs from Qiqqa Browser sessions that were shown in the main application window - these we started from the Sniffer as well.
Qiqqa-20191005.101203.log