Actions
(673.26 KB, 1474x2049)
>>/9753/ Since the text below in this post concerns itself with a similar phenomenon, I will explain the situation yet again. > such interesting things [sarcasm] as how his scraper found that one source had a URL in lowercase and another source had the same URL with capital letters. He is referring to that time that I explained how some of the data in https://poneb.in/ is invalid. Well, basically all of the data in poneb.in is invalid due to the website owner modifying newlines and deleting certain whitespace, but that's a different story (and I guess all of the source pastebin.com data/archives that poneb.in is based on is available). The malformed data that I pointed out was this: a likely small amount of paste IDs in poneb.in are invalid due to case sensitivity or insensitivity; I suspect that this has to do with a mechanism in web.archive.org >>/9613/. Two considerations here: (1.) Caring about web data being a faithful copy of the source and pointing out when it isn't - this is related to "archive-quality" data purity and the dialectic of WARCs vs. raws. (2.) Caring about what some random anon thinks is "interesting" or "annoying". Obviously, I am going to value the 1st point over the 2nd one. So, on with what I was going to point out. Looks like iwiftp.yerf.org has a Windows-friendly-filenames copy of some/all ponibooru.org torrents. Examples: > Ponibooru-All-Unrated/7177 - pinkie_pie %22original%22_character brundle rainbow_dash.jpg > Ponibooru-All-Unrated/7177 - pinkie_pie "original"_character brundle rainbow_dash.jpg the torrent only has the non-Windows-friendly-filename > Ponibooru-All-Unrated/7643 - %22original%22_character ms_paint funny_to_me skywind weanus alpine_horn comic MetaSue.PNG > Ponibooru-All-Unrated/7643 - "original"_character ms_paint funny_to_me skywind weanus alpine_horn comic MetaSue.PNG and > Ponibooru-All-Unrated/7641 - artist:wasd999 original_character.jpg > Ponibooru-All-Unrated/7641 - artist_wasd999 original_character.jpg related to >>/9754/ > Ponibooru-All-Unrated/7642 - artist:wasd999 gun.jpg > Ponibooru-All-Unrated/7642 - artist_wasd999 gun.jpg URLs look like this for duplicate images where one has a more friendly filename: . https://iwiftp.yerf.org/Pony/Art/Image%20archives/ponibooru.org/Ponibooru-All-Unrated/7177%20-%20pinkie_pie%20%22original%22_character%20brundle%20rainbow_dash.jpg . https://iwiftp.yerf.org/Pony/Art/Image%20archives/ponibooru.org/Ponibooru-All-Unrated/7177%20-%20pinkie_pie%20%2522original%2522_character%20brundle%20rainbow_dash.jpg Those are two different files on that server, and web.archive.org falsely sees them as the same file. ウェブ魚拓 correctly sees them as two different files: . https://megalodon.jp/2024-0312-2108-56/https://iwiftp.yerf.org:443/Pony/Art/Image%20archives/ponibooru.org/Ponibooru-All-Unrated/7177%20-%20pinkie_pie%20%22original%22_character%20brundle%20rainbow_dash.jpg . https://megalodon.jp/2024-0312-2108-30/https://iwiftp.yerf.org:443/Pony/Art/Image%20archives/ponibooru.org/Ponibooru-All-Unrated/7177%20-%20pinkie_pie%20%2522original%2522_character%20brundle%20rainbow_dash.jpg . (archive.ph also doesn't mess up the percent encoding, and rclone and .warc download should get it right too.) Other than being somewhat interesting to me, it is useful to know this. A reason: if you rclone those folders from that HTTP website, then you won't get a 1:1 copy of those torrents. Not only will some files be renamed, you will get extra files, which are MSWin-friendly files.