Recently, I bit the bullet and tried a conversion.

It didn’t work. My hard drive was full.

Except it wasn’t - it supposedly had about 60 GB free. The problem? 40 GB of that was “purgeable” storage - storage that was used, but that could be freed up.

After looking online, I found a couple possible causes:

• Some articles suggested turning on iCloud Drive’s “Optimize Mac Storage” option, which removes copies of files stored on your computer that are also on iCloud. However, I’m on the free iCloud plan, which only gives me 5 GB of storage. That left 35 GB of files.
• Some articles suggested creating a few giant files, forcing macOS to clear out space for them. However, that was exactly what I had been doing - trying to create a huge file. My computer wasn’t letting me do that, either.
• Some articles suggested downloading a cleaning program. I don’t recommend those - I don’t trust them, and they might break some files.

After looking around, I found the culprit hidden in some comments: Time Machine. Even though I had told it to ignore my virtual machines (I back up any data separately), it somehow managed to grab a partial copy of my virtual hard drive in the middle of the conversion, filling up all that extra space. While it technically is purgeable (Time Machine eventually sends it to an external backup file), it wasn’t being cleared out automatically.

A user by the name of Andrew posted a comment on an existing blog post giving some terminal commands to fix it. I adapted them to my own use case, and ended up with the following.

1. Go to System Preferences > Time Machine and turn off “Back Up Automatically”.
2. Open Terminal.
3. Type the command tmutil listlocalsnapshots / and press Enter.
4. You should end up with a list of Time Machine backups. Each one starts with com.apple.TimeMachine. and ends with a date and time.
5. For each backup, type tmutil deletelocalsnapshots "datetime", replacing datetime with the date and time at the end of the backup text. For example, to delete one of my snapshots, I typed tmutil deletelocalsnapshots "2019-02-07-091157". Then press Enter. (If you know the time when you were messing with the large files, you may only need to pick the backups around that time.)
6. Repeat for all the snapshots in the list.

At this point, you should be able to see the amount of purgeable storage space decrease. In Finder, select the name of your computer, then right click “Macintosh HD” and click “Get Info”. The amount of space you have will be listed near the top.

Enjoy your new free space! (And remember to turn Time Machine back on in System Preferences when you’re done with the extra space.)

When I changed the address, GitHub removed the old one from my account and then unverified all my past commits (stating “The email in this signature doesn’t match the committer email”). GitHub still identifies the commit as mine, and provides the same PGP key; it just doesn’t recognize the email address as verified.

In case anyone cares, I declare that all Git commits prior to January 2, 2019, which are linked to JBYoshi on GitHub and signed using the PGP key with the fingerprint E726 40D0 9D0C A9EE A098 79C5 468C 8081 1BCC 9A6A (abbreviated 468C80811BCC9A6A) shall be considered as valid as those created on and after January 2, 2019 which otherwise meet those requirements. As proof, I have signed the commit originally creating this message with my new email and updated PGP key, and GitHub now shows it as verified.

My updated PGP key can be downloaded from the SKS key pool in the same way as my original PGP key once the update propagates. The fingerprint and key ID will be identical. If both email addresses show up in the search results, you should be good to go. As of the time of writing, the update is available on agora.cenditel.gob.ve:11371, and will make its way to the rest of the servers during the day.

When I first got Minecraft, I set up a private minigame server for it. It started as vanilla with a few command blocks and a bunch of repeaters (probably several hundred). I eventually discovered the power of a server mod called Craftbukkit to trim down the number of repeaters and make everything more powerful by using actual code. Things were pretty stable for some time.

Then an interesting situation occurred in September 2014: Craftbukkit was bloodily murdered. First the developers tried to shut it down, but the Minecraft developers stepped in and saved the day (in the opinions of many server owners) - it owned Craftbukkit, and had done so for a long time. This began a long controversy that shook the entire community - one of the Craftbukkit developers got upset. He found a hole in the legal side of things and used it to shut down Bukkit, throwing the Minecraft community into chaos. I decided to cover this shaky state of things as a series of blog posts on my old, private blog.

In the research for my first post, I stumbled upon Sponge for the first time, and kept a careful watch. I got more and more interested, until June 20th, when I finally got a GitHub account. I submitted my first pull request two days later - SpongePowered/SpongeCommon#72 - and it was accepted three days after that. I was very excited about joining what I hoped would be a source of stability for years to come.

I continued contributing for some time, until I was contacted privately by @Zidane, one of the leaders of the Sponge project. It took some jumping through hoops, but I ended up being accepted as an official contributor. Technically, all you gain is access to a few new channels to contact them through; in practice, however, it means you become well-known among the internal Sponge team. (You can also get custom badges on the forums and, more recently, their Discord channel, though I never actually requested a forums badge.)

So well-known, in fact, that it seeped into their booth at Minecon 2016 back in November (of 2016, duh). I showed up and introduced myself to @Mumfrey as someone who had been following and contributing to the Sponge project for some time. He asked what my GitHub username was, and when I told him, it became instant access to being introduced to a bunch of the Sponge developers in person (plus some extra Sponge-related loot). Zidane even asked me if I wanted to become a core Sponge developer then. However, I was somewhat encumbered outside Minecon - specifically, a really tough AP Latin class, and couldn’t accept the responsibility at that time.

The other difficulty was that although I occasionally went on IRC, it was incredibly rare to see me on there. However, when I discovered the “official unofficial” SpongePowered Discord server, I joined immediately. To date, I’m pretty much online all the time. (I’ll blame that on my phone, which is usually connected to the Internet 24/7. However, I am on “Do Not Disturb” at night, and the Discord app doesn’t sync the DND settings, so I’m not always available even when I’m online.) That all changed when @Grinch set up a Discord-IRC bridge. Now the Discord server has reached the "official" status dropped the “unofficial” status, and I’m on IRC whenever I’m on Discord, disguised as the Spongie bot (it will indicate when it’s from me).

That opened the door that someone stepped through today. @gabizou DM’ed me about becoming a core developer, and I accepted. Becoming a core developer (as far as I’ve seen) gives you direct access to modify the repository without having to go through a pull request (though the developers sometimes do use pull requests), manage issues and pull requests, and access to more internal communication channels. It’s pretty cool. And I finally get the title of “Spongineer”.

1: End a comment, if it exists

//*/

If there is a multi-line comment before this, it will end the comment here. If not, the rest of the line will simply be skipped.

Here’s a basic example of this:

int startOfMultiLineComment = 0;
/*
Multi-line comment
//*/
int endOfMultiLineComment = 0;

//*/
int notCommented = 0;

2: Toggle comments

/*/

If there is a multi-line comment before this, it will be ended. If not, one will be started.

Here’s an example:

int startOfComment = 0;
/*/
Commented out
/*/
int endOfComment = 0;

3: Not the start of a comment

//*

I use this to say “maybe start a comment, maybe not”. This is the disabled state; only this line is commented. To enable it, just delete one of the /s.

An example:

//*
int maybeCommented = 0;
//*/   // See #1

Then I can quickly switch to:

/*
int maybeCommented = 0; (This is commented out)
//*/ // See #1

So I use this to switch between two variations of a file quickly:

// Variation 1

//*
int variation1 = 0;
/*/
int variation2 = 0;
//*/

// Variation 2
/*
int variation1 = 0;
/*/
int variation2 = 0;
//*/

I do have some work-in-progress ideas, but to implement them fully, I’ll need SpongePowered/SpongeAPI#1370 to be merged. It’s been put on hold, as SpongeAPI version 6 is going to be for Minecraft 1.11 changes only. All other breaking changes (such as that pull request) will have to wait for API 7. Once it’s merged, I’ll get my ideas finished and implemented.

My current code is here: SpongeAPI | SpongeCommon.

I’ve pushed a few minor API changes to my SpongeAPI fork. I’ll post here as I work on it.

1. Length is more important than complexity. This xkcd comic I saw some time ago demonstrates this concept pretty quickly.
2. Although length is more important than complexity, complexity can help.
3. Use different passwords on different websites. If you use the same password on multiple different websites, anyone who figures out one password can easily log in to all the others.

So here’s the approach I thought of. It’s surprisingly simple.

1. Think of a random, easy-to-remember phrase tied to the website you use it on. In my opinion, funny is the best way to go for this. Just make sure nobody else thought of it first.
2. Type it in to the password box unaltered. Keep the capital letters, spaces, punctuation, everything.

Yep, it’s that simple.

Most password strength meters will give this a high rating. The capital letters and punctuation are also helpful for using it on websites that have requirements on complexity. And making the phrase related to the website you’re using it on helps with duplicate passwords (#3) and memory.

1. Set up the repositories individually, as you would normally do.
2. Choose a location to place the submodule. Place exactly one checkout of the submodule there.
3. Delete the submodule directory from all other locations. In their place, add an empty directory (for compatibility reasons).
4. Run git config --local submodule.SubmoduleName.path /path/to/submodule in each repository you want to share the submodule between. The path should either be relative to the repository root or absolute.

Now you have a single checkout of the submodule shared among all repositories. I made a sample repository using SpongeForge and SpongeVanilla; you can view it here. See the README file for usage instructions; the init script contains all the logic.

I decided to start on a new program. Right now, it’s just called “Robot Game”, but that’s the best I could think of. It allows you to write Java programs to control virtual robots. I’m almost ready to push out the first commits, but I’m trying to add in a few more features first:

• An automatic updater. Every few minutes, your client will check for an update, and download one if it exists. I’ll need a CI setup, though, (although it could just be a .git/hooks file), and I’m still working on writing the updater code.
• Right now, the program just looks for a file named RobotGameScript.java in your working directory. I’d like to change that up before release, although it’s not a priority.

I’ll hopefully get it pushed out sometime in the next couple days. Stay tuned!