>>1284257

>for i in grep -E '^htt.*\.pdf$' pdflist;do wget $i; pdftotext -r 300 ${i%.pdf}; done

It isn't equivalent, since basename strips off the leading http URI element. ${i%.pdf} is OK for a filename but not a URL.

You could do a more complex match, but might as well use basename since it will handle any unusual cases more smoothly.

sK:dd

>>1657831

Ask the BO they might have a better idea.

At a guess, it might be a frequency of caching algorithm. The older and less requested files 404.