SGI: Development

fltk 1.3 compiles fine, just needs gmake. vish is a bit stubborn sometimes but we're kind of stuck with gmake, it works okay and most fossy apps depend on it. There's a recent one in nekoware.


Fluid seems to work okay with fltk 1.3. The fonts on the menus are a bit weird but that may be the color combination they chose. Probably has a resource file, maybe can fix that, I haven't looked for it. On an O350 it hasn't crashed or been a problem, but I only messed with it a little to see if anything was obviously wrong.

fltk 3.0, on the other hand, doesn't even try to build. Three or four object files in, ka-poof !


Which Fluid was this ? The old one or 1.3 ? 1.3 seems better. Ime the O2 is also a touch shaky. Lots of things an O2 doesn't like, that don't seem to bother an O3xx series machine. Or even an Octane.


The prior Dillo seems reliable, just limited. I think I have another couple apps that use fltk, they are not a problem. Maybe the screwup is with Fluid, more than with fltk itself. The elderly Dillo doesn't crash much ...


What level Irix and so on are you at ? My O2's have always been more fragile than other SGI-mobiles. Fun little boxes but still, a touch fragile.

Thanks for jumping in to this ...

edit: just built a little fltk app to test, an snmp traffic monitor, seems to be fine, if you want to try ...

I wish I'd thought of using Guh-Noo make on my IRIX box when I was struggling with this problem: http://unix.derkeiler.com/Newsgroups/comp.sys.sgi.misc/2004-10/0039.html

fltk 1.3 is in nekoware? That must be new; the most recent version I saw when I started looking at this Dillo stuff was 1.1.7, IIRC.

Anyway, yes I am ifdefing out all the variadic macros in Dillo and replacing them with overloaded functions that, at this point, do nothing. I'm just trying to get it to compile. It shouldn't be an issue because Dillo only uses variadic macros for it's internal message system, which the Dillo authors say is optional. So the current status of the effort is: All of the C files are compiled and about half of the C++ files. As I mentioned earlier in the thread there are (so far) three C++ files with code outside the variadic macros that cause MIPSPro to puke. It's on my todo list to interface with the Dillo authors about those, on their development mailing list. So anyway, I'll update more status when there's more status to update...

Sorry. I meant that there was a newish gmake, not fltk.

Oh yeah! I installed that after you convinced me to stop hacking on the Dillo makefiles trying to get them to work with IRIX make...

Once I get Dillo compiled I'll tardist it up along with fltk1-1.3 and upload them to nekoware's incoming...

Don't intend on hijacking this thread just want to see what the solution might be. I'm getting this exact same issue on a totally unrelated piece of software I'm trying to compile.

I removed the ... in foo(...) But still cc-1055.

Yeah, you can't remove the variadic macro syntax (the three dots between ellipses) without replacing it with a series of overloaded C++ functions that can handle all the argument numbers and types that the source code is trying to pass in to them. Painful, I know, but that's what I'm trying to do with the Dillo code again right now. After nearly a year hiatus from when I first started working on it...