Many different optimizations to improve performance and capabilities.

[drakmor]

Sorry about the PR, but it's getting really hard to break it down into feature blocks.
I suggest we discuss it in the comments and remove the unnecessary bits.

[john-tornblom]

I will comment here on this PR, and suggest you start from a fresh clone of current upstream master and make feature branches that you send individual PRs for.

[john-tornblom]

On second though, could you provide a bullet list of the new capabilities first, and why it is needed. A lot of the ftp protocol is scattered acroos several RFCs, and many features are not really needed. Also, if a new command is not actually implemented, I would prefer to not have it at all, e.g., the authentication commands. All clients I have tested work without those commands.

[john-tornblom]

FYI: I've started integrating some of your improvements. Some questions:

• whats up with the pthread_attr_* shenanigans in srv.c? Seems like this is not really needed?
• There are some changes not needed, e.g., if(env.data_fd > 0) ---> if(env.data_fd >= 0). fd 0 is stdin and does not need to be closed
• ftp_prefix_ieq et.al can be replaced with libc alternatives (I've done so in srv.c)

Its gonna take a while for me to integrate all of this. I will get to cmd.c eventually

[drakmor]

Hello.

1. pthread_attr_ remains from the battle with stack leakage; you can either leave the creation with PT_DETACH as is or remove it.
2. env.data_fd >= 0 for unification, as opposed to -1.
3. ftp_prefix_ieq yes, thx.

I am currently collecting the issues that arise, and towards the end of the week I will fix what I find and add edits for working with SELF, for example, so that the process is not killed when trying to read certain files.

Added the latest commits to PR