Chen

另外还需要提一个点：__alltraps的地址应该是4字节对齐的。

We welcome such changes if backward compatibility is preserved.

Computing sha256 takes much disk io. It should be avoided in a large repo.

> 如果不进行比较，有可能出现大小一致但是文件被更新的情况 Yes indeed, but this case can be fixed manually. Disk IO is more precious for user experience.

> Did not see the app for iOS.. Will it work for iOS ? You can find NFSee on App Store

Please refer to README.md for a list of supported cards.

Interested in this, though I am using iStudiez Pro for managing these currently.

Thanks, please submit a pull request. I will test on different kernel versions.

Ah, I see where is the problem: Before: ``` $ ls /proc/PID/fd/ total 0 lrwx------ 1 jiegec jiegec 64 May 17 14:52 0 -> /dev/pts/0 lrwx------ 1 jiegec jiegec 64...

> Likely, see also: https://lore.kernel.org/all/20240515-anklopfen-ausgleichen-0d7c220b16f4@brauner/ which did not help. It should mimic the original `anon_inode:[pidfd]`. I am not sure what's the best solution here: 1. Only change the tests, but...