SGI: Development

I just uploaded dillo 3.0.4.1 and libXft-2.3.2 to incoming. You can find both packages at ftp://nekoware.de/diegel already.

This time dillo is build with fltk 1.3.3 and xft enabled. This makes dillo slower, but you have much better fonts available. It is still fast on a R5000 Indy as long as you don't use sub pixel rendered fonts. I build it with the current libXft, which is an dependency now and is available as a separate package.

Please make sure that you have installed neko_fontconfig-2.6.0, there are older packages out there with the same package version number that will crash dillo.

I disabled some of the dillo ssl dialogs and it comes with desktop icons now.

I like to see this version of dillo in current, please test both packages.

Is this built with the latest gcc from current? Erm, I meant g++? Since the liberal use of variadic macros in the dillo code have the same effect on MIPSPro as a land mine...

vishnu wrote: Is this built with the latest gcc from current? Erm, I meant g++? Since the liberal use of variadic macros in the dillo code have the same effect on MIPSPro as a land mine...

Yes it's gcc/g++ like all my builds. I don't see any sense in porting software that was written for gcc to mips pro. In my experience gcc and mips pro output differs in speed +/-10%. Sometimes mips pro is faster, sometimes gcc. The most important difference is it compiles and works with gcc wihout much work. Many problems I had with the firefox3 build caused by broken mips pro nekoware packages. I think using our SGIs with current software is more important than an arbitrary mips pro fetish.

diegel wrote: Many problems I had with the firefox3 build caused by broken mips pro nekoware packages

which isn't mipspro's fault.

anyhow as long as no c++ libs are involved things should work together. you could however add something like -gcc to the file name or so so that people don't have to download first check the readme for finding out whether it's gcc or not.
the important question tho is how many of the programs that don't work with mipspro do actually make sense on irix? the vast majority of the so called "modern" stuff is made on and for linux and in many cases is bloated as hell to somehow compete with the commercial stuff available for windows and osx which they don't accomplish in the end anyway
in fact a fair share of stuff doesn't even compile with anything but gcc on linux which is so poor. i had hopes that the increasing popularity of clang would change that a little but that didn't happen yet ...

anyway getting too far off topic now. any kind of contribution is welcome of course, even gcc builds

foetz wrote:

diegel wrote: Many problems I had with the firefox3 build caused by broken mips pro nekoware packages

which isn't mipspro's fault.

If I want to write my own software I would choose mipspro also. But I can't see any reason to build programs written for gcc with mipspro, when I have a working gcc. If I remember correctly, a lot of the original SGI freeware was also build using gcc.

diegel wrote: I can't see any reason to build programs written for gcc with mipspro

not just gcc but gcc/linux/x86. what might make sense for gcc/linux/x86 could be not so great for let's say gcc/osx/ppc or gcc/solaris/sparc ... you get the idea. the compler is only one part of the whole thing.

either way you nailed the problem, writing stuff for one specific compiler (and os in worst cases) is pretty much the worst case except for very special situations.

digel wrote:

vishnu wrote: Is this built with the latest gcc from current? Erm, I meant g++?

Yes it's gcc/g++

And that's the version that has the recently made MIPS code generation improvements yes? Might have to give that a try...

foetz wrote: either way you nailed the problem, writing stuff for one specific compiler (and os in worst cases) is pretty much the worst case except for very special situations.

The only way to change this is paticipating in open source development. Don't blame a community that don't has access to our compilers/systems.

vishnu wrote: And that's the version that has the recently made MIPS code generation improvements yes? Might have to give that a try...

Yes, it's build with gcc 4.7 and please try it.

diegel wrote: The only way to change this is paticipating in open source development. Don't blame a community that don't has access to our compilers/systems

not at all, this isn't limited to mips or any specific platform. the opensource community has free access to clang, sunpro for linux and solaris and the intel compiler for non commercial use and pgi had a free offer for osx as well not so long ago ... and others i forgot about right now
there's no reason for limiting themselves to gcc and certainly nobody forced them to "abuse" the *features* of gcc. by simply running the code through a serious compiler one can discover many flaws the source might have and start delivering an actually proper product.

diegel wrote: The only way to change this is paticipating in open source development.

In some rare cases this is true. A very few projects are concerned with portability and standards. Those projects also go out of their way to get access to oddball hardware. The guy at Ted, Graphics Magick, and XPDF come to mind.

Most of the others dismiss anything other than the latest generation of x86 hardware and gcc as "legacy" to be despised. Hardware is cheap, disk space is cheaper, integer is the be-all and end-all, gcc works perfectly on all platforms, whatever the latest buzzword is, is perfect. Remove that old cruft from the configure script ! Irix, HP-UX, NeXT, all worthless old trash from before we were born ! Old coots know nothing !

In too many cases the "open source community" is a bunch of ignorant self-righteous no-talent buffoons. Hell-ooo Mozilla Corporation !

Don't blame a community that don't has access to our compilers/systems.

If they were remotely responsible they'd get off their dead worthless asses and buy some of this "exotic hardware." If they actually cared about the quality of their work they'd do this on general principles because no monoculture is robust. They'd do it from pride in their work and a desire to build a quality product.

But they don't. They don't even test their shit on their own beloved loser platforms. Look seriously at gtk2 - it starts as garbage and gets progressively worse. Every two or three versions they made some insanely stupid decision, such as removing all the default icons from a toolkit ! A couple versions before they removed them, they renamed them. For what purpose ? Just to fuck people up ? What kind of braindead imbeciles do such things ? What does that tell you about Open Sores in general ?

I love the idea. And some of the projects really are great. But in far too many cases, these are five year olds playing grownup in mommy and daddy's clothes.

hamei wrote: Remove that old cruft from the configure script ! Irix, HP-UX, NeXT, all worthless old trash from before we were born ! Old coots know nothing !

We have to accept that we are legacy. If you don't drop support for old systems we still would deal with 64k memory bounderies. Compared with NeXT or HP-UX our situation is still quite comfortable.

hmm, what's wrong with hp-ux? last time i checked it was fine and recently hp said they wanna keep it up for another 10 years

foetz wrote: hmm, what's wrong with hp-ux?

Possibly I don't know the right source. Is there any freeware library for hp-ux comparable with nekochan?

foetz wrote: hmm, what's wrong with hp-ux?

Would you like that list alphabetized or by priority?

diegel wrote: Is there any freeware library for hp-ux comparable with nekochan?

indeed and even a professional one. but see for yourself: http://hpux.connect.org.uk

ClassicHasClass wrote:

foetz wrote: hmm, what's wrong with hp-ux?

Would you like that list alphabetized or by priority?

I'm delighted to see that they are primarily PA-RISC packages. That makes me happy. I just don't have a place to put the C8000 right now or I'd grab a few (my HP-UX jones is satisfied by the 9000/350 for the moment).

Decided to give this a try on my Indigo2. It installs fine and from a performance point of view, it's still nippy enough for me and makes browsing Nekochan much more of a pleasant experience than Firefox or SeaMonkey.

Only problem I've noticed with it so far is that the menus appear to be a bit screwed up. Any idea what could be causing this?

Trippynet wrote: Only problem I've noticed with it so far is that the menus appear to be a bit screwed up. Any idea what could be causing this?

I'm still examining this problem, it only happens if you are using antialiased fonts. You can avoid it by enabling a theme in your dillorc. By default there is no dillorc in your .dillo directory. You can copy or edit the system dillorc file: /usr/nekoware/etc/dillo/dillorc.

There is an updated dillo package in beta now.

Awesome! I'll try that when I get time, but might not be for a week or two as my computer room is in the process of becoming a guest room for a few days