It's been a while since my last blog post, but I've still been working on it. Nothing important to mention other than the fact SpyderFly 0.5.0 contained a bugged log file that crashed my entire server for a few days and I'm typing this after getting the server back online. The bug caused a giant log file to be created. It ballooned to 635 GB and crashed the entire server. I promise that bug will not be in the final release.
So the proxy and HTTPS stuff is harder. Because OpenSSL is third party, I don't have a good reference like The Linux Programming Interface and this is confusing. I imagine this will take me less time than the sites thing. Remember, a release will only happen when this website and the others are running SpyderFly. That includes Nextcloud LEMP stacks, this site, my other mini sites, my main site, code server, Jellyfin, and probably something else that's irrelevant. I'm a little burnt out on DaemonCraft too.
BIG NEWS! Port number auto import... thingy... now works! Next step, https and proxy. It feels good to finally figure it out. The fork C command in DaemonCraft made me realize the solution.
I said I hoped to get a release before October and that it probably wouldn't happen; I hate being right. Currently stumped on proxy configuration, root directory, and import sites. So I'll be taking a litle break for now. I decided to take the build for SpyderFly and work on a Minecraft server management panel called DaemonCraft. I took the foundation for SpyderFly, and some code from the auto config and here we are. Features setup so far: it can check the LAN IP (code stolen from the auto config), it can check the public IP (also stolen from the auto config but not exact code; it was the code to verify PHP), and it can check open ports on your router. It already can check the public IP, it can also check if you have ports 25565 and 19132 open. Just like SpyderFly, the port is not hard coded into the code, it can be changed if need be. Eh, lemme create a site for DaemonCraft. The earlier link was the github, click here to go to the site. I haven't set it up yet, I will eventually.
My brain is so fried right now. I don't know what I'm doing. I honestly wouldn't be surprised if I never get SpyderFly finished and released.
I decided to start fresh on the code again and changed the version names. It's likely the first public beta release will be 0.3.x or 0.x.0 because 0.1.0 and 0.2.0 are the early code I scrapped. I'll save the code for those two and put them somewhere for people to look at. Anyway... I'm struggling to figure out how to run SpyderFly to auto import, until I had a mad science idea. The one managed by the daemon service will be called the upstream. It will bind to a port, it will have a folder directory, it will have a visual to verify it's running, and a way to disable and change the port number if desired. You create the config in the sites directory, and then the upstream scans that, and takes a template code and writes an upstream binded to that config. It'll create mini web servers. This is a crazy idea to me and for memory purposes it'll probably be horrible, but most likely version 1.0 will ship with that and subject to change in later versions. In other news, my copy of The Linux Programming Interface arrived. I thought about going to the library today to work on programming, but it's raining. I ain't walking... how far is it? Hold on, lemme look it up... half a kilometer. It's not pouring, it's just I don't want to walk in rain. I'll do some work on the code today, I just have a few things I need to get done for college first. It's piling on me, help. I could use some food. TO THE DINING CENTER! Love the tunnels. My ADHD is going wild this morning. Anyway... SpyderFly is unlikely to be released within the next week, I'd say November, but I still don't wanna jinx myself. December is still the latest, ideally. Side note: reading the NGINX source code got me nowhere and I stopped after I had a dream where NGINX came to life and posessed my body.
Well... idk what I'm doing. I'm going to take a break from working on the sites for a bit and focus on getting a beta 1.0 release where it has a site on HTTPS with a domain, but only one site and no other features like the sites directory. I'm struggling with that one. For now, that's on hold. Beta 1.0 will hopefully be released soon.
I started this project five days ago, and I haven't done much progress since then. Initially I got the web server to run if binded to a specific directory and port, but that's not what I want. Trying to figure out how NGINX does it has been... challenging. NGINX has three source code repositories that I've looked through: the main source code, the package OSS or whatever that means, and the Debian APT repository or whatever it's called. It took me a while to find how it works. I finally did. I don't know how it works for other operating systems, but NGINX works like this for Debian and Ubuntu based systems: depending on the OS, it creates /etc/nginx/nginx.conf and a lot of--if not all of--NGINX runs through that. Depending on the OS, it depends what's generated in it. Debian creates /etc/nginx/sites-enabled written inside the config. I looked through the code for how it manages the config file. It's code was annoying to look through. I found the file ngx_conf_file.c with how it manages the config file. It uses some headers. I tracked down those headers and realized one was not in the code, because you have to execute a file called configure. It creates a folder called objs. That folder includes the missing headers... Which have told me NOTHING!