Linux 4.14 Is Up To Around 23.2 Million Lines Of Code

[phoronix]

Phoronix: Linux 4.14 Is Up To Around 23.2 Million Lines Of Code

While I usually look at the Linux kernel code size following each merge window, I am a few days late this time around due to busy Xeon/EPYC benchmarking and XDC2017. Anyhow, Linux 4.14 is showing some weight gains but nothing too bad...

http://www.phoronix.com/scan.php?page=news_item&px=Linux-4.14-Code-Size

[andyprough]

I'm running rc1 on openSUSE Tumbleweed via the kernel:HEAD repository. So far so good, no obvious regressions, and the system seems stable and responsive. For 4.13 rc1 and rc2 I lost USB and/or wifi and had to switch back to the stable kernel, but no such trouble so far on 4.14.

[edwaleni]

You can now compile a kernel in what, 34 seconds on a Threadripper now? Granted Linux hasn't taken the Microsoft approach to code bloat.

[rabcor]

So many lines, so many unoptimized lines

[Zan Lynx]

So many lines, so many unoptimized lines

As in writing English, writing code with more lines makes it more understandable. There is no benefit to stuffing as much code onto one line as you can. There is no benefit to stuffing every function call together in chains as some Haskell coders seem to want. Intermediate, named variables improve code reading so much!

[JustRob]

Maybe a good Benchmark for AI:

Decompose, recompose, and refactor the Linux Kernel. The function shall not change and improved readability is a bonus.

Any machine with that many parts would be expensive to make maintain, and repair - a Linuxean Hydra.