Anonymous 10/07/2023 (Sat) 20:23 No.33239 del
>>33234
1. Let's make this clear, you aren't trying to extract hidden meaning from files which are still online and can be downloaded with gallery-dl or similar tools without problem?

2. Two strings are of even length and consist of [0-9a-f] alphabet, so it's clearly some binary data encoded as hexadecimal. Whether it's md5 of the file you've shown, you can clearly see yourself, it is not.

It could be some sort of 32-byte hash sum of original file, before stripping metadata and cropping to 1280px. Similar to what endchan is doing, if you download the file again, it no longer will have an md5 of f139d1e286740281e720c023a2c65ce0 as it says in the url path. There is no way to reverse this if unknown amount of metadata has been stripped.

Additionally, it's no longer possible to download "_raw" resolution images since mid-2018.
https://github.com/mikf/gallery-dl/issues/64
https://stackoverflow.com/questions/16832963/what-is-the-url-schema-of-tumblr-images

Here are 2 other scripts I came up with while writing twitter name decoder:
- One does same for unix timestamps of common old-school imageboard filenames.
- The other is for making quick list of search queries for multiple file hashes on 4chan archives.

>>33103
Curiously, all file modification times in your archive are skewed 2 hours and 2 minutes (and in some cases, additional 1 second) behind calculated "true" post dates. Perhaps the zip format or your archive app doesn't preserve universal time. But where does the 2 minute skew come from? If you, say, downloaded them from Twitter with gallery-dl, it should have written minutes exactly, but all files have the same 2 minute skew.