SGI: Development

I don't think so, according to the manpage "cpp is a K&R C language preprocessor (not an ISO/ANSI C language preprocessor)"...

WxWidgets is a cross-platform application development framework, a pretty good one actually. Write once, run anywhere as long as "anywhere" is Windows, Linux/Unix or a Mac...

You can scratch the "/Unix" off that description, too

Okay le, explain this to this donkey : wxWidgets is "cross-platform." But what's the point ? It's not cross- toolkit . So you are still stuck with the same damned problem. The C part is already cross-platform (if the C code is actually up to snuff.) The part that's a hassle is the ridiculous amount of toolkits. So wxWidgets "uses the native toolkits" but there's already implementations of the native toolkits for almost every platform you might want to run.

You can't write wxWidget Base code then have wxWidgetMotif libraries turn it into "native" Motif on a native-Motif system, or get wxWidgetGTK2 output to "native GTK2" on a native-GTK2 system. (Which is what I thought it did and why there isn't a project to do just this astounds me ? Yes, of course it would be limited to the simpler functions but 90% of programs only use simple functions, so why not ?)

So what's the damned point ? GTK2 is the same across platforms, so is GTK, so is Motif, so is fltk. What do you get by adding another layer ?

Well, as an application development framework wxwidgets has a ton of classes that take a lot of the grunt work from the coding. With regard to cross platform toolkits, I think Firefox is pretty close with XUL, though I really like the Motif add-ons they used prior to that. Ramiro Estrugo I think did most of the heavy lifting on that, and Chris Toshok and some others. One of the guys put "No more Motif widget programming for me" in his farewell post at ex-mozilla.org. Their comments about the forehead slapping weirdness of the Motif internals are hilarious. That was back in the day when TOG charged 50,000 dollars for Motif source code licenses. Before Mozilla went open source the PTB at Netscape made their engineers purge all the cussing from the code base. They weren't entirely successful. Another pretty good XP toolkit is TCL/TK and it's variants, Perl/TK for one. Also, Fox is starting to get really good. http://www.fox-toolkit.org/

"Brian Dougherty describes PC/GEOS’s user interface:-

' The object oriented flexible user interface technology in PC GEOS is to this day the most sophisticated U.I. technology ever built into an OS. The team at Sun that developed Java studied it and stole some of the concepts but in my opinion did not achieve the same level of sophistication.

Applications in PC GEOS contained a generic tree of objects describing the user interface features the app required with the ability to provide hints for how to realize those elements. The operating system then had a specific user interface library that would map those generic UI objects to specific UI elements like menus or dialog boxes.

The same binary of an application could be made to run under an entirely different look and feel. For example, at one point we wrote a Mac UI that turned a PC running GEOS into a machine that was almost indistinguishable from a Mac. You could go to preferences and select either the Mac UI or the Motif UI (Windows-like) and the system would restart and all of the applications would come up under the look and feel you selected. You almost have to see this live to believe how cool it was.

We actually got into extensive discussions with Apple about developing a low cost notebook that would run GEOS with the Mac UI. It got killed by the hardware group doing Mac notebooks, but it went all the way to a board meeting we attended with Scully et al before it died'."


http://camendesign.com/writing/geos

Modern interface programmers do nasty things with the rotting portions of a deceased male donkey

Huh! GEOS... To quote Corporal Upham in "Saving Private Ryan," I never heard a that...


Speaking of which on my Solaris box at work they took away CDE and started making me use the Java desktop, what an unadulterated POS! It periodically just stops communicating with the monitor, keyboard and mouse and the only way to recover is to telnet in and reboot the sucker...

But it never locks up ! It's Unix ! Nothing to worry about, that's just the window manager freezing.

Doesn't that make you feel all warm and fuzzy ?

Ha ha, yup! I have to eat my own words, window manager freezes are not kernel crashes...

It does! Underneath those non-functioning console elements Slow laris is still sailing right along...

Back to dillo, vish ... if you can make up a template for fixing the variadick bs, I can do a bunch of the gruntwork of altering the code, location by location. I'd like to have a newer dillo, don't mind putting in some time if it will help ... I could always poop only once a week, that would save an hour here and there. If people gone say I'se full of crap, may as well be full of crap

Actually I'm working on that right now, I tried a couple of shortcuts that failed miserably or I might already be done. I definitely want to get dillo working as firefox 2 is getting beyond tiresome. I'm still not sure what the scope of the effort is going to be since they've done almost nothing to either describe or centralize their message system, but it does look like every fix is going to be custom so a one-size-fits-all template won't fit...

EDIT: Anybody know the MIPSPro compiler argument to get it to ignore empty If-else code blocks?

I'm not sure dillo will be a lot better but it's worth having.

About fireflop, I was serious about reviving OS/2 solutions. I'm now running smartcache with a huge list of refused addresses. It does seem to help. Killing javascript is also a big help. Much of the trouble I have had with fireflop seems to originate in crappy javascript track-you scripts. If the capitalist idiots insist on making the Internet unusable, I'll just restrict myself to the portion of the 'net that hasn't been corrupted. Screw you, google. And the horse you rode in on.

Well, I'm not sure a billion pages of tracking data is useful to even the most sophisticated data mining program, if there is such a thing. But even though google has all this information about what their users do when they go where they go, they still don't know who their users are. That's why they're so PO'ed at Zuckerberg, he did what google couldn't do, which is figure out a way to get the users to willingly identify themselves...

Google data is not what I was concerned about ... it's the googleads and googleapi and all that javascript garbage that people embed in their web pages that totally screws up Fireflop. At least, it does on a slow connection. Try checking out "view page source" and then searching for .js. It's ugly out there ! Killing those addresses (and a bunch more) has dropped my fireflop crashes in half, or maybe more.

Back at the OK Corral, was doing a quick attempt at mxshowfont. It seemed useful. Compiled fine except that

Grr. Something missing in the Makefile ? libXt.sa.4.10 should be a stub ? So I tried butchering things by making a link to libXt.a and renaming it to libXt.sa.4.10 which caused the app to compile. The binary opens but then crashes immediately with a "bus memory error", which seems to indicate that was not a very effective fix

Suggestions ?

run the application under dbx or gdb. when it crashes type soemthing like:

In gdb:


or just..



This should start you on the path to something called "debugging"... It might give you a few hints..

You might want to read the dbx or gdb documentation, find other fun things to try like printign out values of variables and things like that.

R

It would give me a hint if I knew what I was doing, but still cool - the window actually opens now

gdb is a no-go but dbx says :


Cool, thanks. Of course, now I have a couple processes I can't kill but what the heck ! Progress is our most important product ...

You no gots ProDev Workshop? Type `cvd programname` to start it in the graphical debugger, much more efficacious...

EDIT: Well alright, I did it myself, it's crashing in main.c in a call to malloc, which, for the non-C-programmers among us is a system call to allocate memory to the program from the free store:

Spiffy ! thanks ... if I can't fix anything, at least I can help to point the way to people who can ...

New Year's eve tonight, have fireworks and dinners to attend so no C program messing for a couple days but ... progress is good, thank you

Gung hay fat choy!

We always said that as "Gung hay, fat boy !"

Happy draggin' year to y'all ....

Now that's just freakin' beautiful!

Which is not something I can say about the Dillo code...

Just kidding! Still slogging through it, one c++ file at a time. Gotta couple of things I need to post to their development list about, but my subscription attempt seems to have failed...

Found this post while I was just compiling fltk 1.3 and dillo 3

Fltk compiled ok. Same issues with ellipsis and c++. Ran out time to learn why. How's the patching? Got it to compile yet?

Did you run the fltk fluid? It works for a bit but the I managed to choose a menu item that made my O2 clear screen and play the down tune before offering me a chance to restart. No crash dump no log (savecore is on).
Second time I quickly made the fluid windows corrupt in redraw before I quit it.
Doesn't bode well for running fltk app like dillo. O2 is stable with native apps so I don't believe it's hardware issue.