dig·i·tal/ur·ban·ite

Well, I’m kind of surprised it hasn’t been longer than nearly four months since I sat down to write here! Go me. That feels like progress of some sort. Maybe.

And speaking of progress, there has been very little progress on River of Kurn of late. This is for a number of reasons. First, November is National Novel Writing Month, in which I participate annually. Once again, I managed to finish 50,000 words of a novel in the 30 days of November! And that meant zero coding, because I also helped plan my own high-school reunion (never you mind just how many years we celebrated), plus work was wild. December arrived and work was even wilder, plus those pesky “holiday” things ate up time in terms of thinking, planning, shopping and preparing.

I did make some attempts to code in October, but failed miserably because I got stuck on a totally optional thing. You see, I wanted to build myself a tool to add more questions to the game easily. Of course, adding them isn’t that hard in the first place, but it was infuriating to add the questions to the questions table, then add in rows for those questions in the statistics tables for the blue game as well as the green game. And if there was an issue importing, I’d have to dump the numbers out of the question table and reassign numbers to the questions before attempting again, or I’d go from, say, question 1331 to 1381.

So I was stuck. I’d never coded something to import from a file before. I can do database dumps and backups, but I had never tried to bring data into the database programmatically before. So I got stuck and then I got busy.

I’ve been off work for a full six days now (and have another 10 to go!) and the holidays are now over for me, so when I woke up at 4am today for no apparent reason, I decided to reorganize my plans for my game and I decided to tackle this import issue.

Four hours of testing, testing, testing and testing some more later, I now have an import tool built into the admin section on my game’s website. I mean, it’s not live yet, I’ll do that when I finish adding another 100 questions (about Star Wars this time) and add a couple of other things to the game. I’m going to rework an introduction question, add text to explain the validation link will be sent to the email address you provide, etc. Anyway. I am super jazzed about this, because it makes adding questions so much easier. I literally browse for the CSV file on my computer, upload it and it does everything for me. Victory! (Another victory: managed to renew the SSL cert for the game site and am feeling very proud about this because it was NOT simple.)

In not so great news, my EVGA GTX 1080 Ti video card died on me just over two weeks ago. It lived about 16 months. I saw my holiday plans of gaming and coding vanish before my eyes. While I could always use the integrated graphics to at least use a monitor and actually use my computer, that’s no way to game. I also have a pretty crappy video card that’s been sitting in my old computer, so that’s what’s in my new computer now. I can’t do much in the way of gaming, but I can power a second monitor, so I can at least get some coding done! (As evidenced by my building my import tool.) The other piece of good news is that I was under my manufacturer’s warranty, so I got approved for an RMA (return merchandise authorization) to send EVGA my card and they’ll send me another one. The bad news there is that even though I paid for 1-day shipping, they only attempted delivery on Tuesday, December 24 (and not Monday, December 23), which means that EVGA was closed. Obviously, they were closed on the 25th as well.

Thankfully, the delivery was accepted on the 26th. But now we’re probably looking at them shipping a replacement next week. And, well, their site says “shipping for online/RMA orders will stop completely on Tuesday, 12/31/2019 at 2pm Pacific Time” and will resume the following Tuesday. So if they don’t ship it out today (Friday) or Monday, I probably won’t get my video card until, uh, mid-late January. Even if they ship it today or Monday, I probably won’t get it before the end of my vacation time. This is the universe’s way to tell me to catch up on TV and coding, I guess.

So that brings me to my plans for 2020 for River of Kurn:

• 0.2.0 release scheduled for January 4, consisting of: 200+ more questions, the tool I created, other bits and bobs and hopefully a way to keep sessions active or allow them to fail gracefully.
• 0.3.0 release scheduled for January 25, consisting of making things work properly on mobile.
• Beta release scheduled for February 29, consisting of 37 separate issues… hahaha oh boy.
• Full release scheduled for July 1, consisting of an additional 15 issues.

And, of course, since adding questions are so much easier now, I’m hoping to add more questions much more regularly.

So that’s what’s up with me. Hoping to post more frequently here, too!

Earlier this year, I made the decision that I needed a new computer. I mean, this decision has been in the making for a while. I bought my last computer from Dell in 2011 and have upgraded it a few times — added RAM, swapped video cards and such. But it takes forever to load programs. Even loading Chrome can make the computer temperamental. Multitasking? Mostly out the window.

The CPU is an Intel i7 930 quad core and, at cpu.userbenchmark.com, it ranks at 330th of 1122 CPUs ranked. I mean, not terrible, right? Still, I can’t do most of what I want to do as well as I’d like to do it. So I decided, with the help of a friend of mine, that I’d not only buy a new computer, but I’d assemble it myself.

I bought all the parts on Tuesday night and they’re starting to arrive today (Thursday).

The new CPU I’m getting is ranked 8th/1122. Just a bit of an upgrade, eh?

I went all-out on this. We’re talking 16GB RAM (expandable to 64), 500GB SSD and 1TB HDD, GeForce GTX 1080Ti graphics card, SLI-compatible motherboard — and I suspect the 1080Ti will drop (eventually) so it won’t be quite so expensive to buy a second one later on, since the 2080 and 2080Ti have been announced. No peripherals, since I’m just going to use the ones I already have. And I also want a new monitor to replace one of my two (and I’d give one of them to my brother). So the new rig will be able to handle video card upgrades, RAM upgrades and such.

My only worry is nuking my CPU by accident, hahaha oh god. Anyway, I should be assembling everything late next week and I’m super excited to be able to open programs like Photoshop while Spotify is open, or open Discord while playing a game with my brother and not have my ENTIRE COMPUTER FREAK OUT. Even opening my writing program, Scrivener, can make everything chug and then I’ll realize that oops, I left Docker open from the last time I was coding and maybe THAT’S the problem.

So, coding.

When last I wrote, I was dealing with registration validation of things like unique usernames and such. I’ve moved from the PHP validation (mostly — still something left there) to JavaScript/form validation. I’ve used regular expressions (oh, regex, I did not miss you) to ensure that usernames are only 5-16 characters and don’t begin with a number. I’ve made sure that if your email1 and email2 fields don’t match that it alerts you to that. I’m currently working on an implementation of password strength based on the zxcvbn library by Dropbox. I’m having trouble with it and, of course, if something is wrong with the JavaScript, the user gets registered anyway because the validation won’t occur and, since I’m only doing uniqueness checks in PHP, it won’t stop “test” from getting registered unless I’ve already registered “test”. (Yes, I’ll have to ensure that JavaScript is running on the user’s browser in order to play the game, good times.)

It’s slow going with JavaScript. It’s definitely the language I’m least familiar with after my web programming diploma. Like my current error is, I’m sure, a variable scope problem, but I’m not really sure how to fix it. Guess I’ll figure it out going forward.

So once the JavaScript validation is done, I want to go back to the PHP checks and look at making sure I have a unique email, which will mean stripping all the dots and plus signs (and everything after a plus sign) from addresses, to ensure that Gmail users, for example, can’t sign up for multiple accounts. In truth, it’s not the end of the world if there are multiple user accounts, but I’m hoping for some measure of cooperation in the game and that would make things too easy for people.

So I figure I need to hold on to two email addresses — the original one and then a unique check one, and run the unique checks on … both? I guess? I still need to figure out the logic there, which I’ve been putting off, if I’m honest, because I haven’t quite thought it all the way through yet.

Also to do:

• add validation email with validation link
• ensure validation email has opt-out/blacklist link

And then, I think I could move on to login. Which is also terrifying to me, but I’ve ostensibly done it before, in PHP II, so we’ll see where that goes.

From there, I’ll certainly want an account page so people can:

• change their email address on file (but not their username)
• reset their password
• invite a friend…? (This may only be a v2 thing, but we’ll see)

THEN, I might be ready to actually create a flow for people to actually play the game. Part of the issue here is that I’m still working on story. So far, I can pull trivia questions from a database and keep score. This is awesome. It was hard work to get to that point! But there’s this whole story I vaguely have in mind, which is going to mean keeping track of different variables and that will affect how the gameplay works. I also want to build in a maintenance function (one that runs automatically but also one I can run manually if need be) and I want to add in all kinds of admin functions like being able to add, delete and modify users through an admin frontend rather than mess with the database directly. Same with adding, deleting and modifying questions. I also want to code something to back up my databases on command and reset the game.

Still, once I get the login going and allow for a flow for people to get to the game, maybe I’ll send out a few invitations to some friends to try it out, just to see if they can break the (very basic) functionality I have. If they can break it, I’ll have a lot of work to do before I can think about bringing in the actual gameplay I have in mind.

Whew. So that’s my update.

And while I’ve been writing this, I got notification from Intelcom that my delivery consisting of my new:

• CPU
• SSD
• HD
• power supply
• motherboard

is arriving in the next ~3 hours. If everything that should be delivered today does arrive, I’ll have had deliveries from UPS, Purolator, Canada Post and Intelcom. In a single day. And I feel for Purolator, because they’re deliverying three packages to me, apparently, including my CPU cooler and my RAM. UPS is bringing me my video card, while Canada Post is bringing me the last part of my nephew’s birthday gift, which is a Spider-Man t-shirt.

Anyway, I should get my day started, despite the fact I just want to go back to sleep. I’ll have pictures of stuff next time, no doubt!

(That title, just so you know, should absolutely be read in the voice of the voiceover dude from The Muppet Show‘s Veterinarian’s Hospital.)

In the week and a half or so since my last entry, I’ve been working on my registration branch, trying to do All The Things ™. Primarily, I wanted to check the username for uniqueness (well, a close approximation thereof) and I wanted to check the email address against a list of domains known to be used by spammers. The username check went pretty smoothly, all things told. I’m using this:

function checkUniqueUsername($normalizedUsername) {
echo "
Unique Username Check
";
// Get the database instance
$pdo = database::getInstance()->getConnection();
//Construct the SQL statement and prepare it.
$sql = "SELECT COUNT(username) AS num FROM users WHERE username = :username";
$stmt = $pdo->prepare($sql);

//Bind the provided username to our prepared statement.
$stmt->bindValue(':username', $normalizedUsername);

//Execute.
$stmt->execute();

//Fetch the row.
$row = $stmt->fetch(PDO::FETCH_ASSOC);
if($row['num'] > 0) {
return false;
} else
echo 'Username is unique.';
return true;
}

Basically, this is a function that is called from the registration page that grabs our database connection info and then asks the database to perform a count of usernames on the users table where the username is the bindValue of :username (meaning $normalizedUsername).

In MySQL, this will return a single-column table, where the column is named “num“. And it will say 0 if there are no entries with that queried username and it will say 1 (… or more?) if 1 (… or more?) exists. Then I’ll return false if there are any entries and return true if there aren’t.

Perfectly logical, perfectly reasonable, works just fine.

So imagine my utter shock and consternation when I tried almost identical code to check the domain of the submitted email address and nothing was flipping showing up in my results. Like, absolutely nothing. Dumping out $row with a var_dump($row) returned a boolean false. What on earth? From my Instagram:

You can see in the image that I tested out the query on the MySQL command line and I got 1 row. So why was I getting a boolean false and why wasn’t I getting my 1 count in the row?!

Obviously, the MySQL query, when I typed it out manually on the command line, was fine. And I had logs running, so I could see that the bindValue was working and was fetching the domain in question. So what was the issue?

I took the weekend off because I went to my friend’s wedding in Ontario — I was a bridesmaid — and, rather than sleep like the dead on Sunday night, I coded. I was getting ticked off. Something wasn’t right.

I tried all kinds of things — different bindValue. No bindValue. Different files. Taking the checks out of the functions. Putting them back in. And why on earth was the username check doing just fine but the domain blacklist wasn’t?!?

I finally figured it out on Tuesday evening. I was googling stuff, for the billionth time, and found this issue on GitHub. In it, down a bit, it says:

Running a GRANT command to allow SELECT for this user solved the problem

I froze, the realization crashing over me like a wave. But of course the user has permission, my queries worked… just… fi… I was logged in as root. OH MY GOD, I WAS LOGGED IN AS ROOT.

Fun fact: the user didn’t have permission on the email blacklist table. Just the user table. That’s why everything was empty! That’s why I was getting a boolean false! And I was consistently logging in to the MySQL database as root, which is why those queries worked! I quickly made an edit to my creation SQL file and then manually added permission for my user — voilà. It worked perfectly.

It worked so well that I added the following checks:

• email blacklist (the original one I’d been having trouble with)
• disallowed usernames (like admin, etc)
• voluntary email blacklist (for people who want to opt-out and don’t want to receive invitations to the game)

I still need to compile a list of profane words to prevent people from using them as a username, then add that check.

I also want to add in a unique email check, which will require:

• storing the email address entered and use this for any communications
• transforming it all to lowercase for uniqueness, so JULIE @ is the same as julie @
• stripping periods from the username portion of the email address (julie [at] gmail.com is the same as j.ulie [at] gmail.com is the same as j.u.l.i.e [at] gmail.com) to check for uniqueness
• stripping any plus signs and anything that follows up through to the @ sign (julie+test [at] gmail.com, for example, is the same as julie+test2 [at] gmail.com) to check for uniqueness

I also want to implement some JavaScript checks, which, shockingly, I actually know how to do:

• Username between 5 and 16 alphanumeric characters in length
• Password strength
• Email address (1) matches email address (2)

Once all the checks are implemented, I want to send a verification email to the email on file, which ideally also holds a “delete me” link that, when clicked, will trigger insertion of that email address into the voluntary blacklist. Of course, if they click the verification link, this will activate the actual account.

Plus, I probably want to ask permission for whether or not I can add them to a mailing list. That seems useful.

So, I’m feeling pretty awesome for realizing it was a permissions issue. I’m feeling slightly less awesome for having made the mistake of testing with root versus testing with the actual user in question, or having forgotten to give the user the right privileges in the first place, but I still fixed it.

And that feels pretty great.

I have made a great amount of progress on my game so far, at least as I see it. But before we get to that — and we will! — I wanted to show off my new living room workspace. I’m fortunate to live in a 4 1/2 (two bedroom) apartment by myself, so I have a spare room. The spare room has been the office for just about a year, now, and that’s where I primarily do, you know, work. My previous remote-working experiences have shown me that I’m way too likely to work too hard and too long if the work area is in my living area, so I keep work and life as separate as possible. (Although this past week, I’ve worked exclusively from my living room because I have an air conditioner in here but not in the office. As a result, I’m pretty sure I worked about 42 or 43 hours this week. Oops?)

As for my gaming, my writing, my coding, all of these things have always happened in my living room where my desktop computer is located. Recently, I got rid of the desk I’d had for like almost 20 years, in favour of an identical desk to the one I have in the office. Then, I also got a VIVO dual monitor, desk-mounted stand for my screens in the living room, as well as a boom arm for my Yeti Blue microphone. Oh, and a great little under-the-desk headphone holder, which I’m loving.

Here’s what it all looks like.

And I’m adoring it. Sure, the wires are absolutely revolting, but I’m not going to cable manage anything just yet, because I’m on the cusp of getting a new computer. I’m going to be attempting to assemble it myself from the various component parts. I’m not much of a hardware person (although I’ve installed RAM, a secondary HD, a primary HD and a video card), so my friend, Andrew, has sworn to me that he’ll FaceTime me and walk me through the assembly if I need a hand. So with a new computer imminent, cable-managing right now is dumb. Once that’s all set up, though, you can bet your bippy I’ll be doing my best to hide all those unsightly cables.

So the new workspace is pretty great for the time being and having the monitors mounted to the stand is AMAZING. I’ve reclaimed so much desk space I hadn’t even realized I was missing. I genuinely like sitting here and my HUE lights just make me happy. :)

Okay, so that’s the update on the workspace in my living room. Now, on to the game stuff!

Once upon a time, I used to chat on IRC (Internet Relay Chat) pretty consistently. Like, I’d go to bed for the night, but just leave my computer and IRC client running, and I would always have my IRC channels in the background while I was working. I mean, I was working as an independent contractor while I was the Chatting Online Guide at About.com, sooooo I’d have to say that was pretty reasonable. Anyway, I made friends online, as one does, and several of them were in Australia. Australia has a pretty significant time difference between the East Coast of North America and, well, virtually anywhere in Australia, so getting to chat with the Aussies usually meant someone was up late or up early.

One day, I woke up and stumbled to my desk and blearily tabbed over to mIRC and, in one of my channels, one of the Aussies had been freaking out several hours prior and had eventually gone all-caps in her exclamations and said “WAKE UP I NEED TO ENTHUSE!!!!!!!”

I have had this deep need to enthuse about my game for a couple of days now, and I don’t really have a lot of people in my life who, a) care about this little project of mine, and b) know anything about programming. So that’s why I’m enthusing here, I guess. ;)

Updating my personal site last week is what got me itching to code this week, because I’d played around with self-submitting forms, which is currently how I’m testing out question/answer functionality. So I sat down to try to finish getting PDO to work for me on my test page.

I’d also made a change to my docker environment in that I created three database users that are limited in their permissions — one can edit the users table, one can edit the tokens table and one can read the questions table. So that involved making sure I was using the limited questions user rather than the root user, too.

For some reason, my MySQL PDO command wasn’t asking one question in isolation, which pointed to an issue with my select statement, which, in turn, pointed to an issue with how I was “randomly” selecting a question row. Once that was sorted, my question was being asked! Hooray! Being able to pull in that question from the array felt pretty great. Of course, in order to get to the point where I was able to replicate what I’d done when I was messing around back in January, I was going to need to be able to check the answer and inform the user if they were right or not.

So my first major issue was “how to carry over the user’s answer to the page upon submission?”, but it ended up not being that hard because I remembered I could just grab the $_POST[‘playeranswer’] data and stash that in a variable. I also strtolower() it, because all the database answers are in lowercase, so for a match, the cases need to match, too.

That worked, but if I wanted to also display the last question, along with its correct answer, and the user’s own answer (plus the correct/incorrect validation), I was going to need to pass along the row from the last question. So I pass that along, currently, in a hidden form field so I can snag that from the $_POST[‘lastquestionrow’] field.

When you load the page on a GET, you don’t have any of that, but when you load it on a POST, you now get last question row and player answer stored into their proper variables. Then, the database is called and we start getting into the fun stuff:

• hey, here was the row for the last question, please display for me the question and the answer
• also here’s what the player said, so show that
• if they match, say so and grant the player a point, announce new score
• if they don’t match, say so, don’t grant a point, repeat score
• if they don’t match but player has no score, say so, don’t grant a point and let the player know they currently have no score

Now, for the most part, this worked just fine. However, I realized, after having done this, why my teachers in my certificate always recommended I go for a does not equal comparison rather than a strict equality comparison. It’s because a strict equality comparison is messed up, yo. I noticed, through my testing, that I was occasionally getting false positives while going through answers that had more than one possible answer and if one of the possible answers has more than one word. So like, what’s the capital of Canada was fine, because that’s Ottawa. But for asking about the capital of the province of Quebec, that’s where issues cropped up. Your choices are “quebec” or “quebec city”. And yet, somehow, with my earlier logic, “city” was acceptable.

In another example, “flash” was an acceptable answer instead of “barry allen” or “the flash”.

And, in the weirdest example, the name “david” was accepted when neither answer was even remotely close to “david”.

So there was something wrong with my comparison logic. At first, I thought it was just that I was using == instead of ===, so I thought okay, stick in a third equals sign there and I’m good! That’s what makes it a strict comparison, right? Right. Should be sorted!

Alas, no, the issues continued, but I went to bed.

The next day, the strange bugginess was in the back of my head all day long. When I sat down to code a bit after work, I really could have rewritten the logic to test for a failed match rather than a positive match, but I was curious as to what in tarnation was going on. How is “david” considered equivalent to one of two completely unrelated names? (Adam and Carter, if you’re curious.)

I sat down and did more robust testing, writing down all the weird results. So I pulled out my Programming PHP O’Reilly book and looked up comparisons, because everything I’d googled was telling me that === should do the trick.

And this is where I learned about the string compare function: strcmp()

The way I understand it is that strcmp() will compare two strings and will return a 0 if they’re equal. So, this then meant I needed to tweak my logic again. A simple “if $string1 === ($string2 || $altString2)” type of statement wasn’t going to work. I needed to funnel the player’s last answer into a use of strcmp() and compare it with the last answer as well as the last alt answer, if the alt answer existed. So here’s what I ended up with.

$mainVerificationCheck = strcmp($lastPlayerAnswer, $lastAnswer);
if ($altAnswerExists == true) {
$altVerificationCheck = strcmp($lastPlayerAnswer, $lastAltAnswer);
}
if(($mainVerificationCheck === 0) || ($altVerificationCheck === 0))

Basically, that says, here’s the mainVerificationCheck variable, which is comprised of a value returned by using the string compare function. By default, this compares the variable containing the last player answer as well as the last correct answer from the last question. Additionally, if the last question had an alternate answer (like in our Quebec City/Quebec question), create a variable called altVerificationCheck and use the string compare to test the last player answer against the last alternate answer.

Then, if the main check returns a 0 (which means a match) or if the alt check returns a 0, then do stuff. (In this case, it’s notify the user they’re right and their points have increased.)

I’ve been playing around with Instagram stories and captured my bug hunting and fix. Here are the pics from that.

So yeah, I got a lot done over the last couple of days. I worked on stuff on Thursday night and then Friday night well into Saturday morning. And I was going to start in on registration and user creation tonight — my Docker containers are up and running, I have a fresh branch of master that I’ve checked out to a new branch called registration, but I’ve had a headache pretty much all day Saturday and it’s like, four in the morning on Sunday, so I should probably go to bed. I’d have gone to bed way earlier but my headaches are largely sinus-based, so lying down is actually super painful when my sinus cavities are full. Gravity helps the head to drain, I guess!

So my headache meds seem to be kicking in and I probably won’t be able to get back to coding until at least the week of the 6th, because a friend of mine’s getting married on the 4th, and I’m a bridesmaid, so I have a lot of stuff to do this week, then I go to Ontario for the wedding on the weekend. Looking forward to finalizing my computer build either this week or next. I think the new computer will encourage me to do even more coding, and possibly not at ridiculously late hours of the night!

When I started talking about this project, back in January, I was thinking an alpha version might be available in last September.

… yeah, not sure that’s going to happen, hahaha. The base functionality of questions and answers? Sure. But not the game itself.

Anyway, the next task is going to be user registration and login, so I can start saving stuff to the user table and saving stuff to sessions. I want to not have repeat questions, so I’m going to save a certain number of questions in the session (I think?) and check that when doing the random question picker. I think.

Maybe. We’ll see how this goes. I’ll definitely keep you, whoever you are, dear reader, posted on my progress.