Well, it took me, uh, 352 days from September 4, 2018, to August 22, 2019, to complete the 100daysofcode challenge, but I did it. 100 days where I worked on my game. Granted, nowhere near the 100 days in a row, but it was really useful to keep logging things and keep track of what I was doing and where I was going. Plus, reading about my failures and/or challenges now is just funny.
This log entry might be one of my favourites. I am so frustrated, you can practically hear the fatigue in my text.
Anyway, in my last blog post, I figured 62ish hours of work to get things done for Labour Day. Well, my alpha launched at 2am the morning after Labour Day (so close enough), and it was close to 130 hours of work, from June 30-September 2. Note to self: double all estimated hours going forward.
Throughout the last year and more, I have learned a great deal about PHP, MySQL, Amazon Web Services, Docker, Composer, sessions, CSVs, importing CSVs, SSL certificates… A lot of stuff. And now that the game is launched, the adventure is kind of only starting. I’m going to have data! Actual, real, user data!
I also desperately need a LOT more questions, so that’ll be my focus for the next couple of weeks. Why do I need a LOT more questions? Well, the game is currently, shall we say, heavily weighted towards the nerds and geeks of the world. There are 300 Star Trek questions. 100 Red Dwarf questions. Over 100 Doctor Who questions. 100 Back to the Future questions. It’s, uh, not great for a well-rounded trivia experience, so when I looked at the real-time “Correctly-answered questions percentage” at some point today, it was 38%! Like, what?!? Sure, even I don’t get 100%, but I get in the mid-70s, typically. Granted, I wrote the vast majority of the questions, so I do have an unfair advantage, but even taking that into consideration, I thought for sure we’d be looking at closer to 50%.
Well, my dad signed up and played and, because we have a small sample size and because it seems he got a bunch about the periodic table, he brought the overall correct question percentage up to 42% and change. But, the wonderful thing is that I can call up all these stats and find out which questions are hard. Or too hard. Or too easy. Like, okay, I have a question about a hockey player named Blake Geoffrion, who’s descended from not one, but two famous Montreal Canadiens players. Sure. That’s hard. I don’t expect a lot of people to get that.
Should a non-Canadian know the capital of our Nunavut territory? Well, probably not.
But you’d think most people would know Brandon Routh played Superman in Superman Returns, no? Well, still very geeky, I guess.
Anyway, I’m watching the data and will be writing many more questions that are less nerdy in scope. The goal is to get like 5-6000 questions total for the main launch. Then the Star Trek questions would only be five percent of the total questions instead of a whopping 25% or so right now.
That said, props to whoever knew Deanna Troi’s mother’s name, whoever knew who played the Seventh Doctor and whoever knew what game Queeg challenged Holly to play!
Well, I took a really, really long break from #100daysofcode. Like, 31 days. And it was all because of National Novel Writing Month. For the 8th time ever (and seventh time in a row), I achieved the goal of writing 50,000 words in the 30 days of November. NaNoWriMo is something I have attempted sixteen times. I started in 2002 and I skipped 2011, but I’ve at least attempted it 16 times.
Finally, this year, I’ve pulled even. 16 attempts, 8 victories, 8 losses. And, in the process, I crossed the 500,000 lifetime word mark. Solely during November, over these last 16 years, I have written 500,000 words of various novels. It’s astounding to me. That doesn’t count my other writings, it doesn’t count how I’ve continued on with some of these novels. It’s literally just 480 days of writing. That’s just over 1000 words each and every one of those days. Some days, of course, I’ve written as many as 17,000 words (for real) or as few as 0, but the average is about 1000.
So another November comes to an end and another December begins.
And now, I can go back to coding.
I had figured I could absolutely write for NaNoWriMo and code for 100 days of code at the same time.
Wrong! Just wrong. Maybe I had time, but my brain did not like the shifting back and forth from writing to coding, so I did no coding at all during November. Oh God. What is code?
I’m planning to get back to it tonight or tomorrow. In the meantime, I haven’t been posting here about coding because I’ve been posting about coding in my 100 days of code log on GitHub. I do want to spend more time writing about my challenges and problems here, though, in more long-form writing.
Highlights include my doing a Docker course by Bret Fisher, doing #Hacktoberfest, and generally getting through portions of my game. I’m still in the intro in the game, but 90% of the code I’m writing now will be used in the actual game, so… this is good.
That’s about it for now, but that’s what I’ve been up to. :)
I started doing the #100daysofcode challenge on Tuesday morning. It was a modest start, just did about an hour of stuff before work, since I was inexplicably up at 4:30am or something dumb. And yet, I was able to actually tackle true email uniqueness.
So I had two concerns here:
Prevent people from creating multiple accounts with the same email address. Like, user @ gmail.com is the same as u.s.e.r @ gmail.com or user+game @gmail.com. So I didn’t want that to be possible.
Not mess up people’s given email addresses so I can actually send them communication about the game. What if someone’s not on gmail and, I don’t know, the period between their first and last names actually matters? What if someone set up a mail filter for the game using the plus sign? I definitely wanted to send mail to the address they had provided, but I didn’t want to have that address be the unique address.
What I did is save the given email address, but also normalized it (strtolower). Then, I made it unique by virtue of, well, here’s the code and I’ll explain it below.
Breaking the email up into its component parts of username @ and domain by exploding on the @ sign made things easy for me to modify just the username and then reconstruct the email address as a new variable called $uniqueEmail. So I then explode on any + sign. Then, by focusing on the first part of the resulting array ($sanitizingUsername), which is to say anything before any + sign that might exist in it, I’m now dealing with just the username. Then, I replace anything that isn’t a-z, A-Z or 0-9 with nothing (“”). Then, I rebuild the email address with the sanitized username I just created, concatenated with the @ sign and the previously-split apart domain.
This allows for an email address like julie.m.a.r.t.i.n+game @domain to be viewed by my program as juliemartin @domain for the uniqueEmail() check function, but, I’m still emailing that specific julie.m.a.r.t.i.n+game @domain address for any game communications.
Bonus: julie.martin @domain works fine, as does juliemartin @domain. Exploding on a + that may or may not exist has no ill-effects if it doesn’t.
So that’s how Day 1 of #100daysofcode went.
I’m logging my daily progress in a log file on GitHub, so you can check it out here:
Guess what else I forgot to change?
The password field. I’m validating the password too, so I made two fields, one called password1 and one called password2 but at no point in time did I change my snagging information from the POST to snag from password1, so it was still pulling from the (non-existent) password.
Do you know how much fun it is to try to log in with a password that doesn’t actually exist when you don’t know it doesn’t exist?
What’s hilarious is that the issue with the password storage and hashing comes on the heels of resounding success on Day 3 where I basically coded a whole function blindly and, to my utter shock, it worked perfectly on the first try.
Day 3 was dedicated to getting SendGrid to send out emails for validation and for blacklisting and I’d set up the validation workflow on Day 2 and finally got the link mailed out on Day 3, whereupon I then coded the blacklisting function without testing it at all until the end and it worked.
But Julie, you may ask, why the hell would you use a blacklist function? Surely people who are signing up for your game are, you know, actually interested in playing your game, no?
I have a Gmail address. I’ve had it since April 29, 2004. Unfortunately, being an early adopter means that I got my first choice of email address. As such, everyone else who wants that email address has to modify it, like if it were julie @gmail (it’s not), then everyone after me had to do julie1 @gmail or juliea @gmail or julie01 etc, etc, etc.
Do you know how often people forget that they have appended something to their email username?
Like, multiple-times-a-day-often, sometimes. Certainly multiple times a week. In the past, I have received emails from banks, airlines, universities, family and friends of other Julies, real estate agents and, the worst of them all, REPUBLICAN CAMPAIGN NEWSLETTERS.
All because these… these… these bozos, can’t properly understand what the hell their email address is.
So many of them are important messages so I try to write back with a canned response going “hey, wrong person” basically.
Just slightly above the Republican campaign newsletter crap I get, in terms of annoyance, is the stuff other people have signed me up for. Not actually me, obviously, but like an acquaintance, friend, family member of someone else out there has signed up my email address for Blue Apron in the past. Multiple times. This woman Susan signed me up to Blue Apron in March of 2016. Three. Separate. Times.
Thankfully, they listened to me and blacklisted my address, but the point here is that I don’t ever want anyone to have to deal with that pain in the ass when it comes to my game.
Hence, a blacklist system. My blacklist function does the following:
looks up the user
snags the unique email
adds the unique email to a voluntary blacklist
deletes the user
deletes the token associated with that user
You can only get to the blacklist page by virtue of clicking the link in the email sent. The URL adds the token and the user ID in its parameters, so it’s like domain.com/blacklist.php?t=TOKENHERE&user=USERIDHERE. If the token doesn’t match the user ID, it won’t work. And I do refresh the tokens once someone validates, so they can’t accidentally delete themselves.
Gotta say, I’m pretty pleased with that functionality. Plus, if someone wants to register and uses a blacklisted email, they get asked to email me to remove the email from the blacklist. If the email doesn’t match, then I won’t remove it, simply.
So there’s been a lot of progress. Registration is effectively done. Login works. Time to clean up and refine All The Things and then work on:
identifying an admin user
adding menu options for logged-in users (admin and not)
working out what the hell logged-in users should be able to do! hahaha
Some planning to do, for sure, but this stuff is getting interesting. Keep up to date with me on Instagram, where I’m posting stories when I sit down to code and post a pic when I’m done working for that day.
I’ve been thinking about it for at least a couple of months and I’ve decided to embark upon #100daysofcode. Why? Because it’s there!
No, truthfully, I’ve wanted to do this since I first saw the tags on various Instagram posts that I’ve seen. The thing that’s been stopping me is that I literally cannot commit to 100 days straight of code. Probably the soonest I could do that is December 1.
So while I’m starting the #100daysofcode challenge today, September 4, there are two major changes I’m making to the rules for myself. The rules state you need to code one hour a day for 100 days and you can only skip one day every two weeks, adding that skipped day to the end of the 100 original days to ensure that you are actually doing 100 days.
The first rule change here is that I am not going to code from September 25 until the 29th (may or may not be inclusive). Why? I have a work trip. I definitely do not think I can get away with sitting in my room coding for an hour a day when the rest of my colleagues are hanging out and chatting and getting to know one another. As such, instead of this challenge ending for me on December 13, it’ll end much closer to December 18. Or possibly later.
The second is that every November, I participate in National Novel Writing Month. That means that I should be striving to hit 1667 words of creative writing during every day of November. It rarely actually turns out like that — I’ll sometimes go several days without writing and then will suddenly write 7,500 words and will catch up. However, the goal is 50,000 words in 30 days and so because of this, I’m giving myself permission to skip coding 1-2 times a week during November.
So that’s an additional 4-8 days, which beings me to December 22 to December 26. Obviously, you have Christmas in there, too, which needs to be accounted for, so here’s my thinking: in the 119 days between today and December 31 (inclusive), my aim is to reach 100 days of coding for at least one hour a day. That’s totally doable. I may even come in ahead of that projection, depending on how well the writing goes in November.
19 days is a bit of a buffer, definitely, but I think this is an achievable, albeit challenging goal. I did think about putting this off until December 1, when I could conceivably actually do the challenge as intended. But what’s the point of putting off something I want to do that will encourage me to do something I should be doing? While I’m a life-long procrastinator, it seemed silly to me to hold off on this challenge for another three months.
I’ll be primarily posting to Instagram, but those posts cross-post automatically to Twitter. Follow me on either: the username for both is juliebugmtl.
And now, to get a cup of tea and start my first hour of coding in this challenge, since it’s 7:14am and I’ve inexplicably been awake since 4:30.