I perfectly understand why they don't want to merge bad/broken code. And not, that is not the same as 'unfinished code'. In case of unfinished code (like Vulkan when it was merged) most things that don't work are well understood (known to be unfinished) in this case we have an 'optimization' patch set, which fixes things in some cases, but causes crashes and other random behaviour in other.

The authors, instead of insisting on merging that as is right now, should make sure the code, when enabled, at least passes the Mesa standard test suite. Some tests may be skipped or ignored when they are known to be broken by some missing functionality, but ignoring failing tests 'because someone can fix that later' or 'it is ok, this is not a problem for the games it work with' is a very bad idea.

Unexplained test failure means some undefined behaviour with some unknown trigger. It may start affect games that work now when something changes anywhere in the build or runtime environment. Imagine: steam upgrade and game starts to crash, just because it was on the 'whitelist' and used the bad code, which just happened to work earlier. Steam developers, Mesa developers and game developers (provided they still care) won't be very happy with the bug reports coming.

View this comment - View article - View full comments

The social/legal part reminds me A Tale in the Desert.

View this comment - View article - View full comments

I don't understand why distributions like Ubuntu don't stick up with current Mesa releases. It is not like Mesa API or ABI changes with every update and whole distribution needs recompiling. The API/ABI is mostly OpenGL, which is backward-compatible practically to the very beginning. Mesa also has very solid development process and continuous integration infrastructure, so regressions are extremely rare and quickly fixed.

I see no reason for Ubuntu released last year not getting an updated to Mesa 13.0.1. Getting it compiled with LLVM recent enough for full functionality of the Radeon drivers might be a problem, though.

View this comment - View article - View full comments

There is a good side of such requirements: RAM is cheaper than a high-end GPU and you can usually easily add RAM to a laptop.

View this comment - View article - View full comments

It doesn't crash for me. It used to have some visual glitches with neck-ties of the staff at the beginning, though and now I have brightness problems.

I have just changed my laptop (Intel HD4600 replaced with Intel Iris 550, so OpenGL 3.3 with OpenGL 4.5) and now the game is much too dark in some areas. Practically unplayable without using the 'mat_fullbright 1' hack. The brightness setting in the options window does not work at all.

View this comment - View article - View full comments

Quoting: liamdaweAdded requirements in now it's properly live: Requires Nvidia 600 series, AMD 6000 series, Intel Iris Pro or better.

Works on Intel HD Graphics 4600 (Haswell) too.

View this comment - View article - View full comments

Quoting: neowiz73while in development doing a benchmark won't do any good because with the development layers in place

Development layers can easily be enabled/disabled on runtime using environment variables. Or even when they are configured via game settings, they can be easily disabled and then they cause _no_ overhead. Vulkan validation and debug layers is not a thing that you recompile your whole project with. Just enable it, or not, when loading the Vulkan driver.

There might be some other development / debugging code enabled there, though.

View this comment - View article - View full comments

Quoting: KimyrielleI am however, genuinely curious what can still done to Civ to improve it.

Fix the AI – so it is a challenging opponent without so much cheating.

It seems like the AI is as dumb in Civ V, if not dumber, as it was in the first Civilization. The only reason AI wins on higher difficulty levels is because it has more resources, gentler rules, more game knowledge and artificially aggressive „diplomacy” (everyone against you whatever diplomacy you did). Yes, computer is not an intelligent person and may need some help, but game would be much more fun if they felt a bit more 'real'.

View this comment - View article - View full comments

Quoting: adolsonThat screenshot doesn't make it look very exciting.

This looks exactly like Master of Orion should look like in 2016. I can't wait for the release.

View this comment - View article - View full comments

The time when games asked the X server to render some 3D graphics through GLX calls is long gone. Currently games communicate mostly directly to the hardware, through OpenGL libs and DRI. Desktop software interaction is only for set up – there is little to gain here.

Resolution changing? Yes, that is a problem, but has little to do with desktop performance and a lot to do with bugs in the games or engines. And it is IMHO mostly unnecessary anyway. Most people want to play in monitor's native resolution and the best games just don't change that, but adapt. And even is someone really, really want to play with lower resolution (note: I am Intel HD4600 user – if it would be worth for performance reasons I would know it), then there are ways to do that painlessly, but many games fail here. I guess most problems could be fixed by making the desktop environment or the launcher (like Steam) to fix what the games breaks (restore the original desktop resolution after game exits).

View this comment - View article - View full comments