[Raion]

I think we're arguing over fine details too much. That all said, if someone agrees on a name that's not:

irisware
irixware
sgiware

or anything like that, I have no qualms myself. I think because of Nekoware's current recognition we only hurt ourselves by rebranding to a name that nobody knows.

[Irinikus]

I think we're arguing over fine details too much. That all said, if someone agrees on a name that's not:

irisware
irixware
sgiware

or anything like that, I have no qualms myself. I think because of Nekoware's current recognition we only hurt ourselves by rebranding to a name that nobody knows.

Agreed!

[Axatax]

We have to get this right, because every package we make depends on this --

/opt/nekoware, is OK. I don't like it, philosophically, but I can roll with (and accept) that.

I can't see something like /usr/freeware (or /usr/nekoware) being viable for this project. Historically, there's already stuff there that isn't ours, and isn't in our interest to try to fix.

[Axatax]

The MPV code base is *MUCH* simpler than MPlayer.  Thanks to Google ( O_o ), we have an ffmpeg and libraries optimized for MIPS now (but not for R14/16K - for cell phone processor-type CPUs), and those optimizations are alien to MIPSpro.  They're done with inline MIPS ASM with LLVM.  If I was a suicidal person, I'm sure I could come up with all kinds of pithy laments.

If you're really enthusiastic, and want to build MPV right now, start from the Ubuntu ffmpeg build guide and the VideoLAN Wiki.  You'll need a very recent CMake, mercurial and meson/ninja, which needs a recent Python, which needs, etc.... There's **ALOT** of dependencies to build MPV.  Like seriously, a lot.  That's why we need to do a new distro, essentially, if we want more modern apps.

Ubuntu ffmpeg build guide:  https://trac.ffmpeg.org/wiki/CompilationGuide/Ubuntu

VideoLan build guide:  https://wiki.videolan.org/Category:Building/

---

Did you transplant in all the MIPS-specific improvements ?

I had to move over the Video and Audio I/O for SGI. 

To clarify -- It leverages my SDL2 port, which is where the SGI/MIPS specific stuff is.

[hamei]

If you're really enthusiastic, and want to build MPV right now, start from the Ubuntu ffmpeg build guide and the VideoLAN Wiki.  You'll need a very recent CMake, mercurial and meson/ninja, which needs a recent Python, which needs, etc.... There's **ALOT** of dependencies to build MPV.  Like seriously, a lot. 

Christ :( Can we strip it down ? Much of the stuff "modern" software wants serves no purpose. I guess that's fine if you are running a four-core 3 ghz x86, but for us .... three features that work well would be better than twelve that serve no purpose

[Axatax]

If you're really enthusiastic, and want to build MPV right now, start from the Ubuntu ffmpeg build guide and the VideoLAN Wiki.  You'll need a very recent CMake, mercurial and meson/ninja, which needs a recent Python, which needs, etc.... There's **ALOT** of dependencies to build MPV.  Like seriously, a lot. 

Christ :( Can we strip it down ? Much of the stuff "modern" software wants serves no purpose. I guess that's fine if you are running a four-core 3 ghz x86, but for us .... three features that work well would be better than twelve that serve no purpose

The whole thing is obviously insane.  Welcome to the asylum.

Don't worry about any performance issue with this -- this is software where our machines are in their elements.

[Raion]

Per further discussion we are calling the project Neko II / Neko2 /Nekoware II.

The prefix will be /opt/neko2

This change was made simply because of path length and to also help distinguish us from the legacy project while still giving homage to it.

You should be seeing some further developments quite soon.

[Raion]

https://wiki.irixnet.org/nekoware_ii_design

A little update. Axatax and I hammered out a basic design doc here. We're looking for community feedback.

We should have some more information out soon on development efforts. Moreorless, we're currently at the point of: We have most of the major stuff ported, now we're making decisions on packaging. To remind people how this works:

Nekoware II will have foundations distributed similar to IRIX 6.5 foundations. These will contain base libraries, runtime dependencies etc. for bigger apps.

Then, we'll have individual packages for bigger things. The scope of what is and isn't a foundational distribution is the current discussion. Ideally, we'd have something like:

Nekoware II Foundations (Base)
Nekoware II Foundations (Media)
Nekoware II Foundations (Development)
Nekoware II Foundations (SSI) SSI - Sources, Specs and IDBs (This will lower the size of our other distributions)
and possibly a few others. This depends on the size. Ideally, I want to limit most Foundations to a size of about 600M, excluding SSI. This is so we can put them on professionally made CDs as a way for people to support the project.

Once you have those, to for instance install as a user mplayer, you'd grab:

Base and Media, Mplayer.

For GTK2, you'd grab:

Base and Media, GTK2

Ideally, that is. I'll investigate repackaging some nekoware apps that we're not updating as well (i.e. Firefox, I don't have high hopes though) and maybe see if we can revisit FF 3.5 at some point? I dunno. It depends on a LOT of factors, including how much time solidcore, Axatax, me and others have. We don't want to overload our plates.