Page 1 of 1
[request] GtkMathView package
Posted: Sat 10 Apr 2010, 16:29
by SpiceGuid
The title says it all, i use Objective-Caml plus i have LablGtk bindings to GtkSourceView2 and GtkMathView.
So i could develop advanced math software using Puppy Linux & Geany but one piece is missing.
I simply can't find a GtkMathView package for Puppy.
Could someone help, please ?
Supposing i build it myself should i expect a garden party or a never-ending nightmare ?
Re: [request] GtkMathView package
Posted: Sun 11 Apr 2010, 06:25
by Sit Heel Speak
SpiceGuid wrote:...a garden party or a never-ending nightmare?
Heh...depends on how many of the dependencies you lack.
I built it about a month ago, so this is from dim memory:
There's one ocaml package that takes either a "glut" or a "lib" or a "togl" parameter. Choose "lib".
When you build gmetadom, first apply the patch
gmetadom-gcc43-fixes.patch
from
http://sourceforge.net/tracker/index.ph ... tid=428476
lablgtk must be configured with the --without-gl parameter, because it suffers the unfortunate libxcb-xlib malady. You prolly already knew that.
I just checked out the trunk, revision 1185
Code: Select all
svn co http://svn.gnome.org/svn/gtkmathview/trunk
and am compiling it now.
src/engine/boxml/BoxMLHOVElement.cc needed an
#include <stdio.h>
inserted at the top, i.e. make generated the
printf not declared in this scope
error,
but after inserting that include line it continued without trouble.
Back in 9 or 10 with the exciting conclusion of tonight's episode...
Posted: Sun 11 Apr 2010, 10:54
by SpiceGuid
Wow
I certainly did'nt expect such a fast and informed reply.
Thanks to you.
Posted: Sun 11 Apr 2010, 15:12
by Sit Heel Speak
In the dawn's bleary light, it turns out that the compile fails.
Heh...now I recall, it failed a month ago too, the same way, a series of error messages having to do with the two lines in
src/common/Object.hh
Code: Select all
void ref(void) const { refCounter++; }
void unref(void) const { if (--refCounter == 0) delete this; }
So, what I ended up doing was...wimping out. I downloaded several RPM's and unpacked those.
Problem is, I never
tested it.
If you can give me an example of a thoroughgoing test sequence to put it through, to thoroughly prove that everything is in order, then I will try this installation. If it works, I'll package it into a .pet or .sfs.
Opera 10 can render Presentation MathML embedded on webpages, try
http://www.w3.org/Math/XSL/pmathml2.xml.
Posted: Mon 12 Apr 2010, 18:13
by muggins
SHS,
compiles OK on p412. Anyway, I've uploaded here:
gtkmathview-0.8.0.pet
Posted: Mon 12 Apr 2010, 18:59
by Sit Heel Speak
muggins wrote:compiles OK on p412...
Interesting. I'm using my gcc 4.4.1. Evidently, that little snippet breaks the newer gcc but not the older one in 412.
I just tried compiling a static build of geany 0.18.1 too, on my newer system, which carries gtk+ 2.20.1 (necessary for the current git head branch of the evolution email client) (and yes I did succeed in compiling evolution) (what an alligator-wrestle *that* was!). Nope, no can do geany. In this case the gcc isn't at fault, rather it is too new a gtk+ for geany's scintilla component. Had to drop back and do the compile on 4.20-rt-smp. This static geany does run OK, however (albeit without plugins) on my new Puplet (ordinary shared-libs geany had a nasty habit of vanishing).
Posted: Mon 12 Apr 2010, 19:13
by Sit Heel Speak
@muggins: Incidentally, on the other (Canon scanner) thread... the missing link library (but an earlier version, 1.0.1 instead of 1.0.4) is supplied in mp150_sane-0.14.4.pet.
The solution of the missing-models problem in older versions of scangearmp_common appears to be simply a matter of changing the format of /etc/sane.d/canon_mfp.conf from the way it looks in the newest .deb version to the way it looks in the .pet version, and then just recreate the .pet.
If you can't find the time, I'll give it a go in the next few days. On 412, hehe...
/usr/lib contents for gtkmathview-0.80.0 from .rpm packages
Posted: Mon 12 Apr 2010, 19:53
by Sit Heel Speak
Also, your gtkmathview is missing quite a bit in /usr/lib --see the optional features in the file configure.
Just a moment, I'll post the listing of the /usr/lib I got by combining AltLinux, PCLinuxOS, and Mandriva rpm's...unfortunately, I did not write down the configure line...
OK, here it is...
/usr/lib files for gtkmathview-0.80.0 from rpm's
Posted: Mon 12 Apr 2010, 20:38
by Sit Heel Speak
Here they are, all the files in /usr/lib from .rpm's. I have the static files too, if you want them.
Posted: Fri 16 Apr 2010, 15:02
by SpiceGuid
Thanks to all contributors.
Hopefully i will install GtkMathView and the Objective-Caml bindings
I suppose i need a
gtkmathview-dev package, isn't it ?
Posted: Fri 16 Apr 2010, 15:59
by Sit Heel Speak
SpiceGuid wrote:Thanks to all contributors.
Hopefully i will install GtkMathView and the Objective-Caml bindings
I suppose i need a
gtkmathview-dev package, isn't it ?
Yup...
...but...I just within the past 48 hours figured out Barry's
bones+
woof buildsystem...
A Ubuntu Lucid build (and probably any Debian/Ubuntu build) gives you gtkmathview including the Abiword bindings automatically.
Creating gtkmathview by hand-compiling is, by comparison, what in my adolescence was called "a date with rough, tough Rosie"...
Another 24 hours and I should have a working distro. Then, if you don't have it already, I can give you a comprehensive .pet package or, better, bolt-on .sfs (What version of Puppy are you running?).
Unless, of course, you choose to do a bones+woof build yourself...
I notice, on his blog, Barry recently mentioned that he is using boost (one of gtkmathview's myriad dependencies) from an .rpm. But I built it "The Right Way(tm)" using bjam not too long ago, so might have to upgrade the Puppy world to a better .pet, made from mine, in order for gtkmathview to manifest in all its glory.
Back in 24.
Posted: Sun 18 Apr 2010, 17:26
by SpiceGuid
My config :
Puppy Linux: 4.31 full install
OCaml: 3.11.2
LablGtk: 2.14 --with-gtksourceview2