next »

[oahda]

but does it handle emoji
/sad joke, kill me

Wouldn't count on it!

Today I discovered the -jN flag for Make, which enables a multi-core build (set N to the number of cores + 1).

Since I have an 8-core (FX-8320) my build is lightning fast now  

What happens if you try to run that executable on fewer cores?

[Richard Kain]

I was experimenting with running an external program from Unity last night. I wanted to be able to have the user choose the path to the program, and then somehow confirm that they had chosen the correct program. I was having a little trouble because I couldn't think of any element of the program that the user wouldn't be able to change that I could test on. When I run the program without any arguments it doesn't return anything.

Then I tried to see what error information might be returned. Bingo! When you run the program without arguments, it returns a unique error message. I was able to parse that error message and perform a test based on that information.

[Boreal]

What happens if you try to run that executable on fewer cores?

It doesn't affect the executable itself, only the Make process.  So for me, it will kick off up to 8 concurrent compilations instead of having to do them all serially.

[oahda]

Was just about to go download good, ol' TinyXML to put into my project, only to find that it now has a 2nd edition that's much faster and easy on the RAM!

http://www.grinninglizard.com/tinyxml2docs/index.html

Best library ever just got even bester.

What happens if you try to run that executable on fewer cores?

It doesn't affect the executable itself, only the Make process.  So for me, it will kick off up to 8 concurrent compilations instead of having to do them all serially.

Oooh. Shiny.

[oahda]

My TinyXML wrapper is becoming the sexiest.

[Layl]

Got mono working embedded in a little test project!

[Sik]

This fix in SDL2. Fixes all the problems there were getting it to work on Arch (it worked fine on Ubuntu though). Not a problem for 2.0.3, but was a problem for the later development versions which added IME support to Linux.

[itsyourpalmike]

HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY HAPPY HAPPY JOY JOY

[InfiniteStateMachine]

I got in the Perforce cloud beta (sort of a github for perforce) and it's heavenly. Development and project management is such a pleasure now.

[Garthy]

First bug prevented by C++11's "override" keyword. I'd put the code in the wrong class and the compiler caught it. Yay!

[TheLastBanana]

C++11 is just so fantastic. I love how much it's helped me to make code simultaneously more readable and less buggy.

[oahda]

I think it it actually makes my code less readable now that I can do stuff like someFunction({0.0f, 0.0f}) instead of someFunction(Vectorf(0.0f, 0.0f)). But eh, I guess that just comes down to assigning good names to functions.

[Schrompf]

C++11 is beautiful. Helps alot with the daily tasks. Now my only remaining wish is a proper module system that gets rid of all that #include stuff and the huge compiling times coming with it.

[oahda]

I kinda like how header files of classes work like tables of contents tho. But maybe smart IDE's that work with Doxygen or something can keep track of that.

[Sik]

I think it it actually makes my code less readable now that I can do stuff like someFunction({0.0f, 0.0f}) instead of someFunction(Vectorf(0.0f, 0.0f)). But eh, I guess that just comes down to assigning good names to functions.

If I see something like that I automatically assume it must be some kind of vector. Context gives it away (especially the function name, as you said).

[TheLastBanana]

Yeah, as with anything, it's certainly easy to write unreadable code. As long as you're careful about it though, I've found it cleans things up a lot. For example,

Code:

for (std::vector<int>::iterator it = vec.begin(); it != vec.end(); ++it) ...

has always sucked to write and clutters up the screen with really long lines (especially if your vector is named something better than "vec").

Being able to reduce it to:

Code:

for (auto it = vec.begin(); it != vec.end(); ++it) ...

or even better,

Code:

for (auto &element : vec) ...

Admittedly that does make it less clear what exactly is happening. But as long as you're aware of how C++11 range-based loops work, it makes the code so much more terse and clean to look at.

[oahda]

Yeah, the last option should be obvious to anybody who knows C++ and probably even some people who don't, but know, say, Java or C#, so I think it's a great option. You can just write a comment on the line above saying something like // Loop through the joints. and you probably don't have to anyway as the vector is going to be called joints or so.

[Dacke]

Yeah, for-each is pretty standard in lots of languages. Some use the in-keyword (JS, Python) some use a colon (Java, C++) or ":= range" (Go). It's common enough that I doubt you'd need to comment specifically on it.

[InfiniteStateMachine]

The foreach colon. The one time where java is actually terse :D

[MorleyDev]

I tend to use a functional-esque version when available (.forEach in Javascript, a custom .Each extension in C#, std::for_each in C++ etc) but it's nice to have the loop when I need it

(Reasoning is that a loop is possible to abort via break, so whether or not it gets applied to every item can be ambiguous without further reading, whilst the functional version is always applied to every item in the collection. I like that guarantee since it's one less potential code path to think about).

In other news: Yay, managed to convince team that we should use Typescript for the latest project. modules, classes, interfaces, static typing, still support those legacy browsers that the banks want, yay.