Unity 8, Mir Changes Landed Last Week

[Siekacz]

Wishful thinking indeed sucks...

- Mir is doing nothing faster, it's lagging behind wayland in every aspect and in all likelyhood it won't be done in time.
- "Testing" has nothing to do with it, it's basically just a buzzword you picked up from Jono Bacon's blog or something.
- Better language? GTFO. There's no such thing as a "better language". There are only tools that are more or less suitable for the task at hand. And there's absolutely no reason why C++ would be any better for writing a display server.

Myths, Stupid Linux myths.

Testing catches most of the stupid bugs. X11 doesn't have any and you get what you get - unstable piece of s... software. Wayland will be doomed too by hidden bugs.
C++ makes writing much easier by having some abstraction that C doesn't have. And C++ still is very fast. Of course it is more sophisticated than C, but gives you more control. C is simply primitive, perfect for kernels and drivers, but nothing else. Abstraction makes C++ win this competition. And please - when majority of so called Linux community beliefs in something that doesn't justify that belief as true. Major beliefs out of nowhere hurt Linux a lot.

[bwat47]

The question is will Wayland become ready before or after more and more distros are saying "fuck this, Mir is good enough lets take it".

This is never going to happen:

1. NONE of the major upstream desktops (gnome, kde etc...) have expressed any interest in Mir. The two major desktops (gnome and KDE) already have plans in place to use wayland, and there has already been much work done (unlike an untrue comment I saw earlier in this thread that insinuated that no one had plans to use wayland). Gnome plans to have partial wayland support by 3.10 and a full port by 3.12, KDE is also actively porting Plasma/kwin to wayland.

If the upstream desktop does not support mir, than the distros won't either. Distros will use whatever upstream uses for a display server, no distro would want to maintain a fork of gnome or kde that has mir support, it would be insane.

Mir will likely only be used by ubuntu and possibly a few ubuntu derivatives (but so far no ubuntu derivatives have shown any interest in mir.)

Its pretty obvious how this will play out. Ubuntu will use Mir, and other distros will either use wayland or X.

[nightmarex]

Myths, Stupid Linux myths.

Testing catches most of the stupid bugs. X11 doesn't have any and you get what you get - unstable piece of s... software. Wayland will be doomed too by hidden bugs.
C++ makes writing much easier by having some abstraction that C doesn't have. And C++ still is very fast. Of course it is more sophisticated than C, but gives you more control. C is simply primitive, perfect for kernels and drivers, but nothing else. Abstraction makes C++ win this competition. And please - when majority of so called Linux community beliefs in something that doesn't justify that belief as true. Major beliefs out of nowhere hurt Linux a lot.

Said no programmer ever.

C++ has it's share of quirks and faults and is somewhere between a true OOP and C. It's an abstraction but I think you're using that word wrong because c itself is an abstraction layer ( from assembly which in itself is an abstraction from machine code).

The amount of "MOAR MIR, WAYLAND SUX RICHARDBAGS" and "Ubuntu = Fail" makes for comically reading, keep it coming.

Edit: To clarify if abstraction was what defined a good programming lang we would all use javascript.

[garegin]

Not really the question. Canonical clearly stated that Mir is aimed at Unity and that the API can break at any given time, which is no problem for Unity, but for any other DE. So no DE will make the insane decision to support Mir, only to see it break in the middle of their development cycle, so that they have to chase after the API changes. This implies that distros that use any other DE than Unity will not likely go for Mir, because it would be a nightmare to support.

but Unity is not a DE. Don't they need Gnome to make Mir work with Unity? This is not a rhetorical question.

[Delgarde]

So your only source is an unaffiliated blog which doesn't even cite where it got that statement from?
The link it has in that section doesn't state anything about breaking compatibility.
Long shot and full of FUD.

It's the blog of a Canonical employee, one of the lead developers on Mir. Are you really trying to question his credibility on the matter?

[intellivision]

It's the blog of a Canonical employee, one of the lead developers on Mir. Are you really trying to question his credibility on the matter?

He points to a folder with no documentation then goes on about how Canonical is going to always break stuff without so much as a shred of evidence to back it up, so yes I question his credibility.
Where's the evidence to back up this claim?

[Delgarde]

He points to a folder with no documentation then goes on about how Canonical is going to always break stuff without so much as a shred of evidence to back it up, so yes I question his credibility.
Where's the evidence to back up this claim?

I'd have thought the evidence was from the guy who's going to break stuff, telling you that stuff is going to break.

[intellivision]

I'd have thought the evidence was from the guy who's going to break stuff, telling you that stuff is going to break.

Then where are the full technical reasons with source code citations?

[siride]

Then where are the full technical reasons with source code citations?

The technical reasons are that he is deciding to follow a policy of breaking the API at will. Can you be any more dense?

[smitty3268]

Then where are the full technical reasons with source code citations?

#lolfail

The lead developer is telling you his plans, and you apparently think he's lying...

That's got to be the biggest case of sticking your head in the sand that i've seen in a long, long time.


And the technical reasons are obvious - by letting them break api and backwards compatibility, they can fix bugs and problems in the protocol as they are discovered without having to worry and plan about them ahead of time. It's kind of the point of not going with wayland - letting them change things as they happen, rather than trying to get it perfect up front. They can add new features whenever they want something in Unity too - which they are free to do by breaking the protocol if that's the easiest way, since they only care about Unity which they can fix at the same time.

As for the source code citations - well, just look at the git repository. I'm sure you'll see some things that have changed over the last few months. And kind of the whole point is that nothing is done yet. You can't break compatibility before it's even done.

Seriously, i can't believe we're actually having this conversation....