Wine 2.0-rc1 Arrives, Prepares For Wine 2.0

[AsuMagic]

I did pro audio some time ago.. jack is shit for pro audio.. sorry..

routing is great,
latency and stability is horrible

last time i tried using jack using lmms to use my midi keyboard i had excellent latency. it's your problem if you don't know how to configure software.

[starshipeleven]

I did pro audio some time ago.. jack is shit for pro audio.. sorry..

JACK is just a sound server, on average if there are issues it is because the sound driver is crap.
This is true also for Pulseaudio. People trash talk Pulseaudio when the issue is that the sound driver itself is crap.

[cj.wijtmans]

yep audio is pure shit on linux. terrible latency especially with pulseaudio. not good for "pro audio work" not good for gaming and recording.

[timofonic]

JACK is just a sound server, on average if there are issues it is because the sound driver is crap.
This is true also for Pulseaudio. People trash talk Pulseaudio when the issue is that the sound driver itself is crap.

To me, the whole audio infraestructure is CRAP:

- There's ALSA with it's many design issues: Instead improving it, new layers of crap get added that more or less improve it. Those are the sound servers.
- About JACK, there's JACK1 and JACK2.
- About Pulseaudio, it has *TOO MANY* interesting features but these very rarely are supported by PulseAudio GUIs so they are practically USELESS.

I think systemd is an interesting approach, despite it's shortcomings. Despite of that, I think the developers attacked a less important issue than others for desktop computers:
- Audio driver and infraestructure.
- Semantic file systems and related stuff to become standarized and improved: Tag file systems, Tracker, the KDE and Gnome ones, locate stuff, etc etc etc and a very big etc including giants amounts of kitchen sink solutions. It's not easy to have a standarized solution to convert chaos to order, but ReiserFS TODO had some very interesting ideas too.
- A real multiplatform but also lightweight approach to Gstreamer by improving FFMpeg API.
- A real multiplatform but also lightweight approach to libpurple/Telepathy/Empathy, making it very dependency constrained and having extremely portable and good C++ code. Please add Video and Audio support too.
- Modernize Terminals: Notty, DEC Terminal features and others synthetized in a totally new STANDARD.
- Modernize and incentivate LFS: Make user configuration files more standard in a .config directory, promote it and create a standarized parser for tons of programming languages.
- Make PRIME be less of a mess and make Nvidia clean their shit, even better if adopting Nouveau as 1st class project: I know this is impossible, just a hope.

[schmidtbag]

I'm interested in what the wine devs constitute as stable. Though I don't wholly agree with birdie, he has a valid point regarding that wine has never really had a completely stable release. Personally, I think it should work a lot like Mesa, where each stable release is determined by which version of Windows applications are fully functional in. In other words, if 100% of Windows 98 programs work, that deserves a stable release. If 100% of XP programs work, that deserves a stable release. Obviously, this is based on reasonable doubt (because not everything can be tested) and is based on what version the application itself is recommended for. So, if a program recommends Windows 7 but just happens to run on XP, that doesn't count as an XP application.

But, let's say wine gets a new version number whenever it reaches a major milestone - what would that milestone be for 2.0?

Wine makes a lot of progress but that progress is made in every direction. It seems no matter which way you look, there is something in wine that isn't 100% complete or stable. I'm not criticizing, I'm just confused about how the next version of wine can rightfully be called 2.0.

[AsuMagic]

I'm interested in what the wine devs constitute as stable. Though I don't wholly agree with birdie, he has a valid point regarding that wine has never really had a completely stable release. Personally, I think it should work a lot like Mesa, where each stable release is determined by which version of Windows applications are fully functional in. In other words, if 100% of Windows 98 programs work, that deserves a stable release. If 100% of XP programs work, that deserves a stable release. Obviously, this is based on reasonable doubt (because not everything can be tested) and is based on what version the application itself is recommended for. So, if a program recommends Windows 7 but just happens to run on XP, that doesn't count as an XP application.

But, let's say wine gets a new version number whenever it reaches a major milestone - what would that milestone be for 2.0?

Wine makes a lot of progress but that progress is made in every direction. It seems no matter which way you look, there is something in wine that isn't 100% complete or stable. I'm not criticizing, I'm just confused about how the next version of wine can rightfully be called 2.0.

stable = not breaking 50% programs because of a regression, I guess. CSMT works decently for D3D9 but breaks D3D11 on the few games I've tried (if they ever do). GTK theming is broken and most rendering and gui elements are broken.

[starshipeleven]

To me, the whole audio infraestructure is CRAP:

ALSA is crap, the sound servers on it are sound servers and can't fix ALSA as it isn't their job. Even Windows has sound servers.

- There's ALSA with it's many design issues: Instead improving it, new layers of crap get added that more or less improve it. Those are the sound servers.

No, sound servers are the right choice for that job, unless you want everyone to support that stuff in the driver.

I think systemd is an interesting approach, despite it's shortcomings. Despite of that, I think the developers attacked a less important issue than others for desktop computers:

I'll tell you a closely-guarded secret: Red Hat's bread and butter is servers.
Systemd is the first, second and third coming of jeezus combined, for modern servers (and not just them).

laundry list of stuff

I agree with most points apart from:

• Modernize Terminals: Notty, DEC Terminal features and others synthetized in a totally new STANDARD.

completely pointless. Command line is still alive as it allows to interact with devices to do diagnostics and mainteneance. While I think Notty is awesome, I don't need it to do diagnostics and mainteneance.

• Modernize and incentivate LFS: Make user configuration files more standard in a .config directory, promote it and create a standarized parser for tons of programming languages.

You mean Linux From Scratch here? also, wtf is a "standarized parser for tons of programming languages"? Why is that even needed?

But in general I think these won't happen until users on opensource start teaming up and make some crowdfundigs and/or hire people to make it happen.

[M@yeulC]

I'm not criticizing, I'm just confused about how the next version of wine can rightfully be called 2.0.

The article says that it's their first release wit a time-based versioning scheme. Which would make sense, if none of the releases can be considered completely "stable" anyway

And stability has a lot of definitions...

[schmidtbag]

The article says that it's their first release wit a time-based versioning scheme. Which would make sense, if none of the releases can be considered completely "stable" anyway

And stability has a lot of definitions...

Somehow I missed that part.

But as stated before, if wine's stable releases were based on which versions of windows were fully supported, the versioning would be a lot more logical. For example, if version 3 meant "compatible with all applications designed for Windows XP and prior", that would be a very easy way to know what works and what might not. I think the improved organization would be easier for developers too, where they could focus on completing one OS at a time.

The way wine is built now is kind of like a skyscraper, where there are a handful of floors that are fully furnished with working utilities, while there are floors below and above them that are nothing but the support beams and an elevator shaft. Then, there are other floors where they have furnished offices but one of the walls to the outside are missing, or maybe there's no plumbing. It's a functioning system, but it's far from efficient. Currently, they build floors and rooms as needed, and it's no wonder why regressions happen. If they built the floors one level at a time, it would make for an easier transition as they keep stacking up. Sure, that would mean some desirable features would be deliberately held off for a while, but it would (in theory) yield better results.

[timofonic]

ALSA is crap, the sound servers on it are sound servers and can't fix ALSA as it isn't their job. Even Windows has sound servers.

No, sound servers are the right choice for that job, unless you want everyone to support that stuff in the driver.

I'll tell you a closely-guarded secret: Red Hat's bread and butter is servers.
Systemd is the first, second and third coming of jeezus combined, for modern servers (and not just them).

I agree with most points apart from:

• Modernize Terminals: Notty, DEC Terminal features and others synthetized in a totally new STANDARD.

completely pointless. Command line is still alive as it allows to interact with devices to do diagnostics and mainteneance. While I think Notty is awesome, I don't need it to do diagnostics and mainteneance.

• Modernize and incentivate LFS: Make user configuration files more standard in a .config directory, promote it and create a standarized parser for tons of programming languages.

You mean Linux From Scratch here? also, wtf is a "standarized parser for tons of programming languages"? Why is that even needed?

But in general I think these won't happen until users on opensource start teaming up and make some crowdfundigs and/or hire people to make it happen.

About Redhat: Yes, I got your sarcasm. The main *ISSUE* is that Canonical is the only Linux-desktop company that does it best not because they are really good (they suck a lot), but because others are totally pathetic at maintaining a user base and unable to have enough resources for it. I prefer Linux Mint as a better version of Ubuntu, for example.

Why do we need a zillion of sound servers? What about having one of parts of it standarized and provide a modular design for different use-case scenarios?

And why not finish to merge all the RT stuff in mainline? Please. And what about using it in desktop systems so we don't have to suffer certain "temporal" freeze issues? :/

About Notty-like features:

People blamed systemd, but it really has many advantages. It needs more polishing (not Polish despite I like that language, English is easier) and maybe improve in certain areas, but systemv-type stuff is too ancient even compared to Microsoft Windows.

Notty-like features can look not so important at first, but it could make the processing of complex data much easier. Not only because this way GUI features could be merged into terminals, but because better text rendering and able to see graphical data or even stuff like statistics and graphs.


About "LFS": I'm sorry, I confused the acronym. I did mean Linux Standard Base, it needs a simpler and stricter directory and configuration structure. I already suffer too much pain after trying to understand that totall bullshit called Android.