next »

[ferreiradaselva]

It appears the Android SDK no longer works the way it used to, and my CMake module for generating an Android APK no longer works, because it relies on deprecated tools. No luck trying to download an older SDK, and there's this weird gap in the archives available on the Android Studio website, where no SDK after 2009 is up for download. Can I somehow downgrade using Studio's SDK manager? Couldn't figure out how...

Everytime I think about making a setup for Android SDK, I give up. I feel like it's too much work for a task that should be simple. I decided to just use emscriptem when the time comes, and one of those automated tools to deploy javascript programs to Android.

[bengrue]

Who just spent 5 hours porting file formats and dumping binary images and munging them by hand so a cleanup artist could make his assets more consistent.

It was JUST long enough that automating the process more could've been a payoff, but also might not've been.  So frustrating.

(Also I should use tigsource more than once ever 4 years.)

[JWki]

I'm at a point with my project where I've done enough r&d related to tooling and build infrastructure that I properly work out how the build system should be structured as well as make some scoping decisions - this is a hobby project which means I can take all the time I want without compromises but tbh if I don't want to burn out on this as quickly as on some other projects I really have to take some short cuts to get to the exciting stuff. I don't want to stare at a cornflower blue window for another month.

Limitiations can be very freeing.

[InfiniteStateMachine]

Who just spent 5 hours porting file formats and dumping binary images and munging them by hand so a cleanup artist could make his assets more consistent.

It was JUST long enough that automating the process more could've been a payoff, but also might not've been.  So frustrating.

(Also I should use tigsource more than once ever 4 years.)

Haha I know that pain. Especially when 2 weeks later I have to do the same thing and wish I had written a script. Or vice versa.

[qMopey]

Limitiations can be very freeing.

Constraints are what define a masterpiece, much like a beautiful stone sculpture only comes to form when the original mass is cut away under the slicing planes.

[oahda]

I've tried everything now. Combining diacritics are wrong no matter the script, no matter the renderer. Same thing whether I'm rendering the text with Cairo using Pango's C API in my C++ OpenGL program or just using the command line utility to generate it with ImageMagick. It's not the font and it's not particular to any script. I've tried setting the language locale in Pango to match the script. Nothing. Pango is the latest version, as are Cairo and ImageMagick.

So in the end I filed my first ever Bugzilla issue:

https://bugzilla.gnome.org/show_bug.cgi?id=780999

[Eigen]

Not gamedev related in any way, but I need to vent. I've been stuck on a couple of terrible iOS projects recently at work. Written in Objective-C. Worst code and architecture I've ever seen. And I need to maintain and build on top of those codebases... my eye literally twitches right now.

Mis-use of table view:

1) The application needs to flip the contents of table view vertically, so that row 0 is at the bottom and so forth. This is achieved by inverting table view scale of Y-axis and then inverting the the cell content view as well, so things end up 'the right way up'. This is all, of course, achieved by leaving no comments anywhere and having to track it down over numerous classes.

2) This is a classical mistake made by many developers: you need to show some cells and a single header above those. What is usually done, incorrectly, is that the row count in table is returned as data.count + 1 and a special cell is returned when row == 0. That means you need to subtract 1 from the indexes everywhere to get the position of the data in the array. This shit is unmaintainable and difficult to extend.


Bad architecture:

1) There is an API singleton. This singleton has a single delegate property. When making an API call from a view controller, that vc is set as the delegate so it receives the completion events. Now, I thought this will lead straight to problems when you need to make two calls from different contexts. But no! They solved it by completely disabling the user interactions for the whole view while API calls are made so that you could not move elsewhere or click something else.

2) There are no models, just arrays and dictionaries of objects attached to view controllers all over the place. Nothing is shared.


Other shit:

1) Checking application state by seeing what image name some image view or button has. If it's "Closed" then that must mean it's closed.


And so on and on. Of course there are no comments anywhere to explain if some weird implementation is intended or not.

[oahda]

I've had similar experiences in web work.

But over the years I've realised that I sometimes produce similarly horrid code, not because I'm incompetent but because often on these project I get so strict timelines and there's just no time to do it better, and later on there is no time to go back and refactor because it's time for the next project. It usually gets better over the course of employment as there are meetings and I get more comfortable bringing up concerns as I get to know the others better, but yeah. I'm usually more humble about other people's sometimes not so nice codebases now after experiencing the stress of the trenches myself.

#1 on table view is still amazing tho.

[Garthy]

2) This is a classical mistake made by many developers: you need to show some cells and a single header above those. What is usually done, incorrectly, is that the row count in table is returned as data.count + 1 and a special cell is returned when row == 0. That means you need to subtract 1 from the indexes everywhere to get the position of the data in the array. This shit is unmaintainable and difficult to extend.

Facade that shit away. Life's too short to use garbage interfaces.

[gimymblert]

So basically I have this problem now with coding, I can do pseudo code fine, very precise pseudo code, but when I come to writing actual code, my brain snap out, I don't understand what's going on.

It was already there not long ago when I was coding but less severe, until I had that recursion blitz bug and I took a pause. I rethought and simplified the algorithm by making the needed recursion differently. It's not hard, I have already done things like that. SO my goal with this iteration is basically just to divide a tile into smaller tile based on their proximity to the ship (basically the distance of the ship to the tile center) with this distance being recurse based on progressive limit based on the depth of recursion and cap by minimal size.

But I just can't damn it, that's infuriating, it's like my brain nope it out.

[BorisTheBrave]

Two bits of advice, gimmy

1) Fluditiy comes with experience. You just have to stick with it for a while, and (this is also hard) stay on one platform / language. You might be frustrated today, but tomorrow you might be fine with it.

2) If you already have pseudocode for an algorithm, and you are having difficulty translating it, I'd suggest perhaps your pseudocode isn't as good as you think. One thing to aim for is to break down your code into smaller functions, that are both easier to explain what they should do, and can be coded and tested in isolation. Then you are much less risk of doing something so large you cannot fit it in your brain at once.

[gimymblert]

I don't have any problem translating it, I think it's more of mental blocking and depression, I know how to do it because I HAD done it before.

[Zireael]

Mental blocks are awful. I hate it when you have all kinds of ideas but can't translate them to something that makes sense, be it a story or code.

[gimymblert]

It's infuriating when you have been doing the same thing for a while too, I mean I'm just cleaning a code that did just that, but I have a "snap", I think I should spend time away from the computer, except I have nothing else to do and nowhere to go right now :/ ...

[Photon]

Mental blocks are awful. I hate it when you have all kinds of ideas but can't translate them to something that makes sense, be it a story or code.

I'm kinda there right now myself, even though I just took a huge step away from my project for awhile. Its like my brain just doesn't want to work through ANYTHING, though admittedly I'm at a point where I'm trying to do "grunt" work stuff like UI.

Edit: On the other hand, I sometimes forget that not every solution is apparent after thirty seconds of thinking about it.

[oahda]

Stencil shader to clip geometry not viewed through this gate works in editor scene view but not in game view or build (for any system or any renderer) in Unity... q__q

[gimymblert]

THat's weirdest shit ever, Try build, I generally only test in build

[JWki]

I keep blocking myself again by overthinking a problem. Not a good habit.

[quantumpotato]

Hello, recursion is difficult. I wrote 1 recursive algorithm in at least 4 hours, then several months later thought "Oh I'll just rewrite it in new language it will be easy" and it still took over 4 hours. but it's fun to get working!

Grumpy: Dealing with Apple is frustrating and Xcode's "let me fix that for you" warnings in Swift is schizophrenic.

[smartyMARTY]

I'm grumpy because I chose to be a programmer and not an artist and I am grumpy because I would be grumpy of the vice versa.