SGI: Development

I would like to get up to speed about the state of software packages on Nekoware and if it is still being used. It's been such a while since i used all of this, i have to relearn/rediscover stuff.
I've already seen that there is another distribution site/flavor for mips3 software, good.

The problem is porting more recent software requires more recent C/C++ compatibility and with the MIPSPro compiler this is going to be increasingly difficult. You can go gcc/g++ of course, but that would possibly impact performance and moreover i'm not sure about compatibility of libraries with Nekoware stuff.
Also, more recent software requires modern openGL versions, which isn't available on many systems. Software performance is nowadays less of an issue with current Linux systems but what runs fine on PC's does not run fine on old SGI iron, save for the lucky few who have Fuel's, Octane2's and Tezro's.

I think we should look at porting more generic software packages like programming languages and stuff which isn't in IRIX. I agree with Hamei in that if we have good tools in IRIX, there's no need to replace stuff with opensource software.

I believe jan-jaap (and others) did some testing that indicates gcc is now on par with MIPSPro for code speed, so it shouldn't be an issue performance wise. Library compatibility is a separate issue, but I think most things will work.

Nekoware is still the most used freeware distribution for Irix 6.5, but there are very few users left. Currently we have two main issues: get firefox 3.5 or newer running, get a newer gtk2 than 2.12 working. Both packages compile without major problems, but are not working. Firefox 3.5 stops with the same problem you have seen many years ago and gtk2 stops working if you open the file chooser.

Since we are so less active users, the beta testing process isn't working anymore. We have a very good gcc 4.7 running, but many nekoware users still don't like to use gcc compiled mips code. You still can't mix gcc build c++ objects with mips pro c++ objects, but otherwise there is no problem with gcc code from my perspective.

nekonoko wrote: gcc is now on par with MIPSPro for code speed, so it shouldn't be an issue performance wise

that's wrong. just recently i discovered that something as basic as the string class is multiple times slower with g++ (yes, even using ShadeOfBlue's options).
also diegel started replacing some of the libs for his firefox with mipspro builds because of the speed difference.

gcc got better but mipspro is the first choice as it always has been

foetz wrote: that's wrong.

Hey, I'm just passing on what's been argued to death on this very forum for years now. I'm certainly not the originator of those claims.

nekonoko wrote:

foetz wrote: that's wrong.

Hey, I'm just passing on what's been argued to death on this very forum for years now. I'm certainly not the originator of those claims.

oh no offence intended in any way. just wanted to set dexter on the right track

I have started going in a different direction and so far, pretty happy with it. Same goal, just a different path ...

To me, when nekoware was started the gnu stuff was either pretty good or had a lot of potential. But since then, much of it has turned to crap.

We aren't going to change those people. gtk2 for instance, is just garbage. Not only is it awful to build, what you end up with is still pointless trash. I really really really need thirty gigabytes of html docs in Tibetan -- not !! The nitwits hardcode gcc into everything, the code is often stinkypoo, new-age gnu apps are repulsive. If you work your way up the gtk2 chain reading the release notes, the shit they did is appalling. Kindergarteners should not be allowed to play with software.

But there is good old stuff out there that could be rescusitated. Or in some cases, maybe the okay level of a gnu app can be rescued, bugfixes from newer versions backstalled, and Irixicisms inserted - kinda like what you guys did with MPlayer. If Inkblob or Scribus actually worked that would be cool. There's Maxwell. There's Axene. There's Ted. There's some games. There's some 3d graphics stuff. There's utilities. There's a ton of older software that would be useful.

The trouble is, the coding knowledge is here - there's at least five guys in nekochan who really know what they're doing - but the support at user level sucks. You can't get the "I wuvs my sgi ! it's the kewlest machine ever and I wanted one ever since grade school, why don't those meanies at sgi Release the Source ?" group to even try out the most basic apps and give feedback. Developers can't do everything.

And why should they, if no one even bothers to try it and report back ? It's like pissing into the wind.

So I don't know. Maybe it really is dead. If you do anything, dexter, it will be great. But you better do it for your own satisfaction 'cuz there's about five people here who really use their best-beloved coolest-ever SGI computers.

diegel wrote: Currently we have two main issues: get firefox 3.5 or newer running ...

Another way to approach this problem, and perhaps one that would be more maintainable over the long term, would be to recreate Proxomitron. There is nothing in the 'modern internet' worth having.

Do you think it might be more practical to create a proxy that would insulate us from the sewage that overlays today's web ?

hamei wrote: I have started going in a different direction and so far, pretty happy with it. Same goal, just a different path ...

To me, when nekoware was started the gnu stuff was either pretty good or had a lot of potential. But since then, much of it has turned to crap.

We aren't going to change those people. gtk2 for instance, is just garbage. Not only is it awful to build, what you end up with is still pointless trash. I really really really need thirty gigabytes of html docs in Tibetan -- not !! The nitwits hardcode gcc into everything, the code is often stinkypoo, new-age gnu apps are repulsive. If you work your way up the gtk2 chain reading the release notes, the shit they did is appalling. Kindergarteners should not be allowed to play with software.

But there is good old stuff out there that could be rescusitated. Or in some cases, maybe the okay level of a gnu app can be rescued, bugfixes from newer versions backstalled, and Irixicisms inserted - kinda like what you guys did with MPlayer. If Inkblob or Scribus actually worked that would be cool. There's Maxwell. There's Axene. There's Ted. There's some games. There's some 3d graphics stuff. There's utilities. There's a ton of older software that would be useful.

The trouble is, the coding knowledge is here - there's at least five guys in nekochan who really know what they're doing - but the support at user level sucks. You can't get the "I wuvs my sgi ! it's the kewlest machine ever and I wanted one ever since grade school, why don't those meanies at sgi Release the Source ?" group to even try out the most basic apps and give feedback. Developers can't do everything.

And why should they, if no one even bothers to try it and report back ? It's like pissing into the wind.

So I don't know. Maybe it really is dead. If you do anything, dexter, it will be great. But you better do it for your own satisfaction 'cuz there's about five people here who really use their best-beloved coolest-ever SGI computers.

Scribus is Qt based last time I checked. Now, good luck getting Qt 5 to work on IRIX - closer to compiling then Qt 4 ever was for me, but last time I tried it choked on the webkit section (or is it KHTML? Something about web rendering anyway) and I am not a clever enough person to fully work around it.

Aside from the nostalgia factor, what's in it for most users anyway? My Octanes, while I love them, aren't really too useful in the modern era. Sure, I rebuilt a lot of software on them and used them for web/db servers for a bit (ran a Drupal 7 site on there actually, and had it authenticate against Active Directory on my Windows Server), and the UNIX setup is ideal in my opinion, but the machines themselves are old, noisy, and power hogs. If I wanted to, I could accomplish the same thing using less electricity and with more horsepower on a Raspberry Pi or MIPS CI20 (actually I want either an IBM p520 or HP Integrity RX2660 next, because they were the first AIX and HP-UX systems I worked with. The only problem with the IBM is I'd need the HMC to go with it (I've also worked with i/OS on the IBM 520, that was a lot of fun when Nessus locked out all accounts with SECOFR rights))

hamei wrote: I have started going in a different direction and so far, pretty happy with it. Same goal, just a different path ...

To me, when nekoware was started the gnu stuff was either pretty good or had a lot of potential. But since then, much of it has turned to crap.

We aren't going to change those people. gtk2 for instance, is just garbage. Not only is it awful to build, what you end up with is still pointless trash. I really really really need thirty gigabytes of html docs in Tibetan -- not !! The nitwits hardcode gcc into everything, the code is often stinkypoo, new-age gnu apps are repulsive. If you work your way up the gtk2 chain reading the release notes, the shit they did is appalling. Kindergarteners should not be allowed to play with software.

well that's the other important point and indirectly solves the gcc question as well. the point being: which "modern" freeware is actually worth having?

foetz wrote:

hamei wrote: I have started going in a different direction and so far, pretty happy with it. Same goal, just a different path ...

To me, when nekoware was started the gnu stuff was either pretty good or had a lot of potential. But since then, much of it has turned to crap.

We aren't going to change those people. gtk2 for instance, is just garbage. Not only is it awful to build, what you end up with is still pointless trash. I really really really need thirty gigabytes of html docs in Tibetan -- not !! The nitwits hardcode gcc into everything, the code is often stinkypoo, new-age gnu apps are repulsive. If you work your way up the gtk2 chain reading the release notes, the shit they did is appalling. Kindergarteners should not be allowed to play with software.

well that's the other important point and indirectly solves the gcc question as well. the point being: which "modern" freeware is actually worth having?

Depends on what you're doing - I built PHP, OpenSSL, HTTPD, and MySQL just to run a Drupal site. All in the eye of the user.

armanox wrote: Depends on what you're doing - I built PHP, OpenSSL, HTTPD, and MySQL just to run a Drupal site. All in the eye of the user.

As you just pointed out, you can do all that faster, cheaper, easier and quieter on a peecee.

What you can't do on a peecee is run Irix.

Now, everyone has their own reasons for whatever they do in life. But as far as I can see, the only good reason to use an Octane is to run Irix. All the gnu stuff was interesting and in some cases useful but currently, most of it is shit. It started out okay but then went to the dark side.

Any actual decisions about the direction of nekoware would have to be made by the commyewnity ! But we don't have a community. It's like pulling teeth to get one person to say "hey, the newest version of libjpeg works for me."

Damned if I know.

I know that Display Postscript is tits. I know that I can open PShop 3 in Irix and have a jpeg 'enhanced' in Irix faster than CS3 even starts under Windows. I know that I can create professional-grade CAD files that anyone in the world can use. I know that xpdf kicks ass over any version of Acrobat back to v3. I know that neko MPlayer works better in Irix than it does in Windows. I wish I were a Framemaker ace, it's ten times more capable than I will ever be. There are a few gnuey applications I'd like to have, but none that I need to have. In every case where there's an Irix application comparable to a gnu one, the Irix one is better.

To me, there isn't much point in talking about server uses. If a person wants to do that, it's cool. But the hardware is definitely not the best for that. It's on the desktop that Irix still shines. On the desktop, I'd rather see five programs that run really well than fifty that are half-ass. But if no one actually uses it, then what's the point ?

Dunno, dexter1. I hope you can get something going. But nekoware at the moment is totally stagnant. What is the reason for that ? Answer that and you'll know what to do.

hamei wrote: If there were a commyewnity

Well at least it's more active here than the NeXT Computer forum . I love the concept of that forum; I keep visiting hoping people will post something cool, but consider myself lucky if I find one new post a week.

What we should do n order to compile this "modern" shite on our SGI's is to write and compile binary patches to MIPSPro that add the functionality that C and C++ now possess, that MIPSPro doesn't. Difficult? Yes. Impossible? Probably. Now, who's with me? Into the fray!

vishnu wrote: Into the fray!

nekonoko wrote: Well at least it's more active here than the NeXT Computer forum . I love the concept of that forum; I keep visiting hoping people will post something cool, but consider myself lucky if I find one new post a week.

Yeah, the NeXT guys ... gotta feel sorry for them. The software is very cool but the hardware is just a leetle too slow. When you can type faster than an editor can display on-screen, that's not very practical. And keeping those things running must be a chore

I was surprised at how little I miss the 2x800 O350. Would still like to get four V12's running the one monitor tho. Maybe some day.

vishnu wrote: What we should do in order to compile this "modern" shite on our SGI's is to write and compile binary patches to MIPSPro that add the functionality that C and C++ now possess, that MIPSPro doesn't.

How is that going to make gtk2 apps run under Motif ?

That's the real problem. The gcc stupidities can be coded over. But what are you going to do about the retarded dependencies ? and the ghastly toolkits ? Right now, we're running two complete operating systems, one on top of the other. Sometimes three. That's silly.

Especially when the real one is so much better than the overlays.

"Motif is so ugly" ... umm, has anyone looked recently at the average Loonie desktop ? Bring a barf bag.

Could someone please tell me what "modern era" software brings to the table ? Warning : there's a Windows box ten feet away and software costs about a buck-fifty here. (That's $1.50, not $150.00) And the Assist bought a Macbook a while back (which is now broken, of course. Pay more, get better quality ! unh-hunh.) So I'm not speaking from a position of ignorance.

Into the fray!

Any time I get anything frayed, the Assist makes me throw it away

armanox wrote: Depends on what you're doing - I built PHP, OpenSSL, HTTPD, and MySQL just to run a Drupal site. All in the eye of the user.

of course, server stuff is relatively easy in comparison in most cases. a recent mysql not so much but you get the idea.
anyhow since opengl and gtk2 was mentioned it seems this "check" is rather targeted at desktop stuff

hamei wrote: [...]but the support at user level sucks. You can't get the "I wuvs my sgi ! it's the kewlest machine ever and I wanted one ever since grade school, why don't those meanies at sgi Release the Source ?" group to even try out the most basic apps and give feedback. Developers can't do everything.[...]

Hamei is right on this - and I'm guilty as charged. Everytime I fire up the machine in the signature, I run . After all packages have been updated, I do nothing more (my lazyness and limited time). Well I browse some websites with the new thinking "man is that slow" - even http://forums.nekochan.net/ .

So what to do? Upgrade to the latest R14000A I guess

[...] save for the lucky few who have Fuel's, Octane2's and Tezro's.

. Or I would have to buy myself a fast(er) Fuel.

On the other hand - what is fun even on older machines is commercial software. But for GNU-stuff CPU-horsepower is all I need. And that I don't have right now.

hamei wrote: But as far as I can see, the only good reason to use an Octane is to run Irix. All the gnu stuff was interesting and in some cases useful but currently, it's shit. It started out okay but then went to the dark side.

The question is - how to continue the IRIX-experience(tm)? It is common knowledge, that SGI (and IBM and the other 3-letter-Unix-companies) put tremendous work into Linux, and linux inherited a lot of IRIX' code and capabilities. Linux is for some years now capable to run on IBM mainframes and SGI's current NUMA machines, and has a lot of stuff we love (XFS!!) compiled into its kernel.

So why not put together a Nekochan-Linux-Distro called IRIX 7.0 with MaXX-Desktop as default http://forums.nekochan.net/viewtopic.php?f=8&t=16729754& ???

*****************
Of course I see and appreciate the effort of the Nekoware-maintainers - for all these years now.
But with "mobile-phones-faster-than-a-Tezro" running Android/LINUX or MacOSX/BSD, I see that there is an end somewhere in the near future -- due to moore's law.

Great discussion! Thanks for all the input.

I believe that what we need right now is some way to keep what's good and develop stuff that's useful to run and which is fun to have. Sure a decent modern browser would be cool, but i am not putting all my time in that, since debugging that Bus-Error is one of the hardest things i did. Fun and challenges are the main things which motivated me coming back and i intend to try and make stuff which is appreciated by me and others.
Also, dissing opensource projects is not the way to increase our enjoyment and productivity on our SGI's. So let's not do that, but instead focus on what does work: Do we have cool Motif stuff around? Well, get them distributed and try it out.

I think information is the main problem we are facing. The machines are getting old and knowledge is fleeting. We need to maintain this. I'll be pouring over the wiki to get some stuff straightened out.

But all in due time. My O2 is now at home, and i'll be hooking it up next week. Then let the fun begin

dexter1 wrote: I'll be pouring over the wiki to get some stuff straightened out.

Let me know if you run into any issues with it - I just did a large software update on it last week and haven't had a chance to test everything extensively.

The only problem with the IBM is I'd need the HMC to go with it

Gawd, yes. I hate that. I fortunately do have an HMC for the p520, which I have to drag out periodically when I need to futz with the master LPAR (I'll have to do this shortly because I'm upgrading the RAM).