SGI: Development

Moving right along here, thought I'd try something way over my head. A Motif word processor might be nice to have, he says ...

This thing is so old, is there a directive "CC=c79" ?

try:


then check for the function calls and make sure you always specify match_case because now we don't have the default to TRUE anymore

You probably already found out that setting the default value again, as in
is the problem. just remove the = TRUE, and you'll be able to continue to the next error, which is the simple fact that memem() doesn't exist on IRIX, you'll have to provide a replacement, e.g. from gnulib, just as we've been doing for getopt_long() for ages.
http://www.gnu.org/software/gnulib/manu ... emmem.html

from http://www.gnu.org/software/gnulib/manu ... emmem.html


hehe they might rather list where it's not missing
why do people use stuff that's not available anywhere?

This thing is really old. I think it uses a dinosaur for an icon

It's over my head, too, but would be nice to have (maybe) so I'm willing to plug away if y'all are willing to put up with stupid questions ...

sure go on, would be a nice addition for sure

Where'd you get the source code? Their sourceforge page says "this project has no files"

Two cows has it :

http://www.tucows.com/preview/9775

feel free to dig right in !

And that's the same project as this? Link: http://maxwellwp.sourceforge.net/

EDIT: Nevermind, asked and answered; it is. This is some serious abandonware, most of the files haven't been touched in 11 years. Anyway, you can dl a more recent version than the one at two cows here: http://maxwellwp.cvs.sourceforge.net/viewvc/maxwellwp/maxwellwp/maxwellwp/?view=tar

At least it's got autoconf and doesn't require setting arcane environment variables prior to running make. But unfortunately the bozos didn't generate the configure script or the makefiles, so you need to run autoreconf at the top of the source tree yourself. So, I just did that and ran make and oy is it broken!

Don't snivel, it could have wanted xmkmf


Wait until you get to the fun stuff ! All kinds of weirdass permissions problems and a few ranlib errors and a script or two that isn't executable, really good !

But I got way farther with your version than the thing I had, gracias. It actually went kind of okay more or less until here :


edit : okay, seem to have moved past that by changing the NULL to a 0 ?

?? http://stackoverflow.com/questions/4174 ... nition-why

Now we have


Now, the oddest thing here is maybe not the error but the fact that this happened twice before with other libraries but just re-running gmake over it beat the process into submission.

Even if this thing compiles, I bet in the end it crashes like a mofo all the time


edit 2 : Okay, I'm blind. But I can count to two ... wth ?

Looking at CREDITS I see that Danny Backx worked on it after it's GPL release, he's one of the huger names in the Motif world. Given how far along this project obviously was I can't imagine why everyone abandoned it.

Let's see, the stuff I've had to do so far, copy the file tmp/mx_language.d to the directory shared/util, create the lib directory, make the script build/movedep executable, then it started to compile, then I had to copy the file tmp/mx_stdio.d to shared/util, then I had to copy the file tmp/mx_string.d to shared/util, so I copied all the files in tmp to shared/util, then it can't create static libs because IRIX doens't have ranlib? I know I should remember this but what does IRIX use to do that instead of ranlib? In my defense my days of full time development on the IRIX platform ended about a decade ago...

it is not needed; the symbol index is always updated.

You'd think I'd have been able to remember that!

Anyway if I'd read the thread a little more carefully I'd have seen hamei already had a fix for the shared lib issue, now I've come as far as the "poorly formed" errors...

This command in the Makefile appears to be a problem though ?

If I put an empty ranlib statement in the environment (per freeware compiling hints) or the Makefile, then make complains that it can't find ranlib. Removing this directive from the Makefile does not work either. Is there something one could replace it with that would satisfy the make process ?

how about a dummy script doing nothing named ranlib?

I don't know if it actually fixes the problem, but yeah. Look toward the bottom of page two of this thread, Shade's suggestion does the job.

This will happen to you several more times. I just used the same idea to repair the broken dykes and it compiled. Eventually it gets to where it stops because there's a missing makefile. I saw this in the original ./configure process but figured "fix that when we get to it." So now I'm to-it but it still isn't fixed ...

Did you have a problem with a missing config.h ? If so, what did you do about it ?

No I'm not that far yet I'm still lagging a bit behind your pace.

Very unusual architecture to this program, are you having the same problem as me with the Makefiles continually removing the dot-d files from the directories where you have to manually distribute them from Maxwell's tmp directory? Pretty annoying. Also, it's compiling everything so far into static libs which I assume will link into an eventual binary, but if it's compiling the individual source files into object files it's doing it very surreptitiously because I have yet to see one!

Oh this just can not be good:

They're in a hidden directory in each source subdirectory. This is normal for an autotools-generated makefile.


It looks like a dependency line in the makefile got truncated somehow. Have you set the ncargs systune variable to 262144 before running ./configure and all that? If you haven't, increase it and run "make distclean" then "./configure" etc.


Learn LaTeX, you will never want to use anything else

(La)TeX rocks and I produce my user manuals with it, and while you can practically program every dot on the page, create chess diagrams, music scores, write books with indices, TOCs, etc... Even though I use it all the time, I think TeX is a bit of an overkill for a letter to grandma.



R.