133246 -... png
(1.73 MB, 500x2157)
(1.73 MB, 500x2157)
>>/9759/ Long task to automatically verify the files by file (as opposed to by piece, which spans across files, in BitTorrent) did not complete: ended by reboot. I verified the rest of them manually: found 100 mismatches. Bright side: did a thing to more quickly check variadic various stuff manually. What I did in the past: > '$ h="something1"', '$ cmd "$h"', '$ h="something2"', '$ cmd "$h"', etc. Quicker method: > '$ read -p "Enter text: " h', '[input text]', '$ cmd "$h"', '$ read -p " Enter text: " h', '[input text]', '$ cmd "$h"', etc = 5 fewer keystrokes. Real examples: $ # older: $ h="98294 - rarity artist:brianblackberry.png" $ sqlite3 -column -header -csv -cmd "select * from images where md5sum=\"$(md5sum "$h" | sed "s/ .*//g")\";foobar" -bail $path1/meta.sqlite 2>/dev/null id,md5sum,type,uploader,tags,source,rating,date,origfn,w,h,size,filename,sel 98294,fc026712961fead7b390e667a149a3c1,png,Anonymous,"artist:brianblackberry rarity",http://www.canterlot.com/gallery/image/989-rarity-lounge-singer-final/,safe,"2011-12-14 09:41:44.510538",gallery_69_17_44320.png,642,826,448227,"98294 - rarity artist:brianblackberry.png",1 $ h="61297 - g1_backstory g1 Moondancer.jpg" $ [sqlite3...showed no results for hash of file 61297*] $ # newer: $ sqlite3 -column -header -csv -cmd "select * from images where md5sum=\"$(md5sum "$h" | sed "s/ .*//g")\";foobar" -bail $path1/meta.sqlite 2>/dev/null id,md5sum,type,uploader,tags,source,rating,date,origfn,w,h,size,filename,sel 24149,5ddacb3770921123dc417c5595620b42,png,Anonymous,"artist:sanguiniuschan rarity",http://sanguiniuschan.deviantart.com,safe,"2011-06-13 20:18:23",drink_special_by_chistery_by_sanguiniuschan-d3ip3uz.png,1032,774,265116,"24149 - artist:sanguiniuschan rarity.png",0 $ read -p "Enter a filename: " h Enter a filename: 169759 - Madame_Pinkie_Pie gypsy_magic fortune_teller crystal_ball pinkie_pie animated.gif $ [sqlite3...showed no results for hash of file 169759*] I wish that qBittorrent had a thing to only recheck missing torrent pieces. That could be programmed in, and it probably wouldn't mess up verification of the top-level infohash. Otherwise, lots of I/O gets wasted. Also, whoever made those ponibooru.org torrents like a decade ago: thanks for applying a piece size of 256K. A larger piece size would have been more of a headache to deal with, I think. Image from that website.