So today, our "senior tester" logged a bug because he "Viewed Page Source" on the web based report he's testing and the Year values are hard coded! So 2017 will have to added manually!
I burst out laughing so hard, all my colleagues think I'm having a mental breakdown.

Python. Ok, so it's a really cool language, as a scripting language it's awesome, quick to write.

When it's been used to make full fledged oop programs that you suddenly have to maintain things like duck typing become problematic. Looking at an object fuck knows what methods are available. Worse still when some bastard that thinks he's being clever doesn't bother declaring any object attributes and instead overrides the __set_attr method to dynamically add them as they are used there is no hope for the poor sod that has to maintain it later.

I've also now worked out I'm at least the 3rd person having been given the task of maintaining it, i spent a day changing CLI options wondering why they didn't do anything but occasionally crash the app. I've now found a few thousand lines deep that someone had hard-coded these values because they couldn't work out where to get the CLI args from!

I've gained a new appreciation for nominative, strictly typed languages.

Gaming community of mine launched their slick new website with their new "ticket system" where people could put in tickets to get help by volunteers.

2 hours and an approval by one of the admins later I managed to inject forge http request into literally every form on that side. Modify permissions, delete users, edit tickets, put invalid values into every attribute of them... In other words break everything.
Turns out the whole thing was coded as a first time project by a person who has no clue about web development and noone is in charge of anything really. There are no requirements, no beta testing, no version control or backups, but at least they had a hard deadline. 🤣

Still not sure if I wanna fix their shit and do it properly or just enjoy seeing it crash and burn.

Thanks to the developer who decided to release his library with no documentation and hard coded values .. 10 hours well wasted !

My best code review experience?

Company hired a new department manager and one of his duties was to get familiar with the code base, so he started rounds of code reviews.
We had our own coding standards (naming, indentation, etc..etc) and for the most part, all of our code would pass those standards 100%.

One review of my code was particularly brutal. I though it was perfect. In-line documentation, indentation, followed naming standards..everything. 'Tom' kept wanting to know the 'Why?'

Tom: 'This method where it validates the amount must be under 30. Why 30? Why is it hard-coded and not a parameter?'
<skip what it seemed like 50 more 'Why...?' questions>
Me: "I don't remember. I wrote that 2 years ago."
Tom: "I don't care if you wrote it yesterday. I have pages of code I want you to verify the values and answer 'Why?' to all of them. Look at this one..."

'Tom' was a bit of a hard-ass, but wow, did I learn A LOT. Coding standards are nice, but he explained understanding the 'What' is what we are paid for. Coders can do the "What" in their sleep. Good developers can read and understand code regardless of a coding standard and the mediocre developers use standards as a crutch (or worse, used as a weapon against others). Great developers understand the 'Why?'.

Now I ask 'Why?' a lot. Gotten my fair share of "I'm gonna punch you in the face" looks during a code review, but being able to answer the 'Why?' solidifies the team with the goals of the project.

You know what kind of devs I hate?
- The "Oh I never worked with it so its shit"
- The "I dont wanna learn something new"
- The "You can use JavaScript for this and everything else"
- The Pro ++ Ultra Dev who never heard of modulization and layering
- The hard coded values guy

> do you feel sorry for freelancing contractors
> whose previous client abandoned them
> they ask you to help them fix some trivial bugs in the shitty code
> you believe you can change the world by going overboard by also improving the code quality, along with fixing the bugs
> initialize an empty file where you'll translate the shitty code into a more organized one
> start creating variables and generic functions which can be used in a modular and organized fashion
> meticulously document the first function you write
> realize this is not worth your time
> insert some glue code into the original code which fixes the trivial bugs
> glue code has hard coded values so it adds to the shittiness of the code
> submit the work
> get $$$

When you gotta learn those hard coded values for a practical exam cause your professor doesn't accept other values:)
#feelsEngineeringInIndiaMan

Spent seven hours reading source code at work yesterday. The little documentation I was able to find alternated between English and Spanish. And some of the things I saw... Straight out of a horror novel.

For example: NUMBER_2 * NUMBER_60 * NUMBER_60 * NUMBER_1000 to get the number of milliseconds in two hours.

Or this super contrived method which capped the registration age at 100, which now caps it at 102 anyways because they use hard coded values for the current year. Took me 15 minutes to find out what "fixYear" (this method) did.

No wonder I got home and crashed in bed till nearly midnight after that... I swear that was harder than a university Calc final...

Another terrible rant from the inhereted Hydra source code. So deep in the dark dungeon of that code I noticed something interesting. They declare this INT32 array with an incredibly long (like 200 values) list of hard coded magic numbers. Something along the lines of:

INT32 array[200] = {-1,0,1,21,4,7,19,33...};

However, the resulting output was incorrect. After spending a fort night and a good chunk of my remaining sanity I had overcone the 437 levels of indirection left by the previous programmers, and narrowed it down to this line. But it looked perfectly fine.

I pull up the diffs and notice someone had checked in a change to the source. I track it to this line and find what the original data had been.

INT32 array[200] = {-1,0,1,2l,4,7,19,33...};

In VS the default font shows l and 1 as fucking identical. Someone had accidentally made that change to 21 from the original 2l and checked it in. I mean I can't really blame them. Who the fucking hell inatantiates a fucking int32 array and peppers in a fucking 2l (long) for no fucking reason?!

Hard coded values everywhere give me nightmares. WHAT IS THIS CRAZY FUNCTION USED FOR?

Spent all not fixing someone's errors because they used XML and hard coded values, couldn't even work in the application I was going to work on

Does anyone have any advice for making api integration tests data independent? We have a few hundred tests firing against our seed data and they all check for hard coded values. But the problem is that this means adjusting seed data is really really slow as you have to fix integration tests too, which can be thousands of lines of values needing to be corrected.

Idea - possibly a bad one - visual studio extension that hits a database to display possible values to supply as a method argument. With caching of course..

I'm thinking along the lines of permissions in a database that at some point have to be hard coded against code to enforce them. Stuff like that.

Possible or beyond stupid?