Day 140: Map and Jumping

I finished #155 today, and named my blog post after it. The total explanation for how to use the map and how to jump out of the sector was maybe two paragraphs, and it took the entire two hours to do. Thankfully, it's not because I'm a slow typist.

No, the reason it was so problematic was #162. The previous tutorial mission which tells you to land on the planet only ends when you get the map-and-jumping mission. In order to do this, map-and-jumping sets a flag upon receipt, and land-on-planet checks this flag to see if it's ending. If it is, it removes itself.

I have a condition check for this. It had three fields: is_set, number_is and note_is. Those last two are to check if the flag is set to the given number or has the given note attached. Is_set was, I thought, to merely check to see if the flag was set.

I checked the unit tests to find that its purpose was originally the exact opposite. If you didn't specify is_set, it would return true if the flag was not set. Otherwise, the note or number had to match.

So there was some refactoring. Now is_set does what I originally wanted, and there's a new is_not_set which, if selected, returns true if the flag hasn't been set yet.

That was working, the land-on-planet mission was vanishing at the right time, but the map-and-jumping mission wasn't working at all: it was supposed to pop up an informational box about how to, well, pull up the map and jump the ship around. It wasn't. I started having flashbacks to my two-day multi-hour coding binge of doom. I really didn't want there to be something I'd have to re-code the entire notification system for.

And, lucky me, there wasn't. In fact, it was the other mission, the land-on-the-planet mission, that was causing the problem. The loop which checks all the missions looks like this:

@player.missions.each do |m|
  m.check
end

The problem being that check can end the mission, and thus remove it from the @player.missions array! This messes up the loop, and it ends up thinking it's done and not checking the other mission.

Thus, @player.missions.dup Life was surprisingly easy after that!

Day 139: Miscalculation

"Hey Roger, when did you start this thing, anyway?"
Yesterday, you might have thought I started it sometime in early February, which would have made it day 119 and been a rather belated New Year's Resolution. I actually started keeping track of everything in a spreadsheet on January 12th. I just today discovered a mistake on my spreadsheet that wasn't counting 20 days of work. So the answer to your question is no, watching the Lost season finale did not catapult you 20 days into the future.

Today I got #120 and #131 done, which means the new market dialog is live! I didn't take any 'before' screenshots, so imagine something that's worse than this 'after' screenshot:



Also, those numbers in the 'status' bit haven't changed, but they're important because I made a huge change to the way cargo was handled. All the tests still pass, though. Ship it!

Wait, no, don't ship it yet. I was kidding! Put down the packing tape!

Day 119: Lost

My excuse for today's entry being short is that I'm typing it during the LOST season finale. I recall, when I first started this blog up, I feared that all my entries would be short. A friend of mine reassured me: "This isn't NaNoWriMo. You don't have to pad your word count." She makes an excellent point!

Today I finished #154, which is the tutorial telling people what you can do on a planet. It goes into some detail on trading, which is special because that doesn't actually work yet. This brings me nicely to my next topic: I didn't so much get things done today as create a whole slew of things to do:

Moved #120 and #131 to this milestone, because I'll need them in order for the tutorial to be accurate. Created #162, #163, #164, #165 as cleanup tickets to make things look better and more functional.

In short, my total number of tickets for this milestone went from 7 to 11, and my completed number only increased by one :)

Day 118: Medicated, again

If you glance back at the blog there, you'll see the last time I was medicated was at the very beginning of the month. Believe it or not, this is a huge improvement.

Regardless, this will still be a short entry.

#125 is done. You can scroll to your heart's content, and all will be well.

Before:


After:

Day 117: Caret

Short blog today, as I'm still not feeling so hot. And I ought to, considering that it's 90 freaking degrees outside.

Today I got 'home' and 'end' to go to the beginning and end of line. I got the editor to correctly recognize more than one return character (this was also causing off-by-one errors when it wasn't working). It turns out that String::split will ignore trailing whitespace unless you tell it not to. Yes, I know the docs I just linked to say that, but I hadn't considered looking there until things weren't going right.

Programming Tip of the Day: Read the Fine Manual.

Other work includes interpreting mouse clicks and positioning the cursor appropriately. After that I did arrow up and arrow down; the clicking was needed because I simulate moving the caret up by pretending there's a click on the line above the current one. This works remarkably well.

Back to sleep!

Day 116: Line10k

I finished my compulsory two hours of work today two lines of code short of 10,000, so I went on to work for another two hours. If you look at the summary to the right, you'll note that in that two hours, I wrote a whopping 10 lines of code. This was because there were tons of bugs and it required a lot of subtle fixes that were also wrong.

All four hours were spent on #125, the ticket to put in saner text editors. As it was before, you could only change the last bit of anything you were typing. So if you'd made a typo a paragraph earlier and just now noticed it, you'd have to delete your way to it, fix it, and then re-type everything.

Today marks the second time I've had to sit down and do some work on paper. It initially seemed easy to figure out where a cursor should be drawn; I was iterating through each line of the input, adding the size of that line to a running total, which I would then compare to the caret's position. This was error prone and hard to tune.

Then, after writing down an example, I figured out that I could determine which row of the input the caret was on by counting the number of newlines before it.

Later, I kept getting these strange off-by-one errors. I'd been thinking that the caret - an index pointing to where in the original string any changes will be made - was off from the displayed string because the display had extra newlines put in there from word wrapping. It wasn't until I wrote down a sample situation that I realized the extra newlines were replacing spaces (because that's where the words would wrap) and my intricate correction code was causing the off-by-one problems.

Additionally, because I want 'home' and 'end' to be valid text editing keys, I had to change the "emergency end the program" key away from 'end'. It's the backslash now. I accidentally ended the program twice today anyway.

For those interested, the 10,000th line of code is:

      <child name='flags'/>

Day 115: Tutorial

Programming work was done yesterday, but by the time I finished it was no longer day 114. 1.75 hours of programming was enough to finish up the Missions milestone! From now on, it's module development.

The problem with tracking my progress is that every file saved by Kuiper was just one line. One gigantic line with no linebreaks - it's XML, so this actually doesn't matter, but it makes reading it with an ordinary text editor a pain. It's this way because any whitespace was interpreted as a text node, and my code wouldn't check for that, instead trying to re-create a ship from whitespace. This is even more impossible than it sounds :)

So the first part of my day was spent changing the way the XML serialization works. It now skips over any text nodes (I don't use them; all text ends up in attributes, which is a bit messy but works) and once more passes the tests. With that out of the way, I can now use Cloc to count work on my module as lines of code.

The first module I'm making is, in my opinion, one of the most important: The Tutorial. One of the reasons I'm doing this now as opposed to later when the editor is more stable is that I'm trying to shake bugs out of the editor. Later one of my goals is to make a scenario-making tutorial, and I'll likely rewrite the original tutorial then, so if I mess up now it's not a huge deal.

Two bugs fixed today in support of the tutorial:

#159 If you ever lose sight of the planet, it's very easy to get lost and not find your way back. Even for me, and I wrote the darn game! Now, if the center of the sector starts leaving sight, a green arrow appears on the radar to show you which way to head back.

#132 Is the first part of the tutorial, giving basic info on how to move around and land on planets. This one helped to reveal not exactly bugs, but things that would be nice. For example, I've had #125: Better Text Editors open forever, but I've moved it to the tutorial milestone because now I'm typing a lot of text and it'd be nice to be able to edit it in a sane manner.

My wrist hurts from typing; I'm taking the rest of the night off!