compiling Kalles Fraktaler 2.7.3 on Linux with mingw

[claude]

So I tried to compile Kalles Fraktaler on Linux targetting win32, made some progress (mostly adding const all over the place) but now I'm stuck, and the error log is a manageable size now (at the start it was 1.2MB including warnings and notes, now it's 13kB) so I'm asking for help!

The first error I'm having trouble with is:

Code:

/usr/lib/gcc/i686-w64-mingw32/4.6/../../../../i686-w64-mingw32/include/comip.h:12:2: error: #error No __uuidof support for this target

If I comment out this code in common/getimage.cpp:

Code:

#include <comdef.h>

It gives these errors:

Code:

common/getimage.cpp:25:2: error: ‘IPicture’ was not declared in this scope
common/getimage.cpp:25:12: error: ‘pPicture’ was not declared in this scope
common/getimage.cpp:36:44: error: ‘IID_IPicture’ was not declared in this scope
common/getimage.cpp:36:74: error: ‘OleLoadPicture’ was not declared in this scope
common/getimage.cpp:43:5: error: ‘OLE_HANDLE’ was not declared in this scope
common/getimage.cpp:43:5: error: expected primary-expression before ‘)’ token
common/getimage.cpp:104:2: error: ‘IPicture’ was not declared in this scope
common/getimage.cpp:104:12: error: ‘pPicture’ was not declared in this scope
common/getimage.cpp:120:40: error: ‘IID_IPicture’ was not declared in this scope
common/getimage.cpp:120:70: error: ‘OleLoadPicture’ was not declared in this scope
common/getimage.cpp:149:5: error: ‘OLE_HANDLE’ was not declared in this scope
common/getimage.cpp:149:5: error: expected primary-expression before ‘)’ token

I'm guessing the code is supposed to turn raw image data into something that can be displayed into a window.  I did some searching and found suggestions to use VS to create header files for mingw to use on Linux, but I didn't find enough detail to know if it'd be easy.


The other error is more of a C++ thing, and I might be able to solve it if I search enough online:

Code:

common/StringVector.cpp:282:35: error: ‘void* operator new(size_t)’ may not be declared as static
common/StringVector.cpp:286:35: error: ‘void operator delete(void*)’ may not be declared as static
common/StringVector.cpp:290:41: error: ‘void* operator new [](size_t)’ may not be declared as static

EDIT: think I just solved this, by putting them as class members.

[youhn]

 

You got my support for the Linux version! I'm no programmer, but if I can help in any way, please let me know. I can read and write like a few-years old can handle written normal languages, but I do know the whole configure-build-install drill.

[quaz0r]

when i first got turned on to fractals i endeavored to port kalles to lunix.  you are basically wasting your time.  the code itself is so windows-centric, and i think intentionally obfuscated by karl so as to discourage anyone actually being able to read it or work with it, its just a monumentally losing battle.  in the code he posts...every variable is named a b c d e f, there is no whitespace or formatting, etc, he clearly did not intend for any human to actually do anything intelligent with his code.  not to mention the outright dubious design decisions he made with his program.  i quickly realized it would take many more lifetimes to successfully port kalles than it would to simply go back to the drawing board and learn the latest stuff about programming, learn the latest stuff about fractals, and just spend some time writing something from scratch.  what a great decision that was.  and with his hardheadedness about his dubious design decisions, like not supporting rendering animations to a series of stills like a normal sane renderer would do, building with visual studio 1.0 for windows 3.1 32bit even though everyone has had a 64bit processor since 20 years ago, forcing ill-conceived functionality on the end user like his "merge frame count," where you have no choice when making animations but to use his autist-inspired functionality of "merging" previous frames into the current frame to "anti-alias" ... omg how ridiculous and terrible that is, and it taints every single video ever made with his program.  you can see plain as day, boxes within boxes, in every single frame of every single video made with his program.  clearly made with rendering as-is no-supersampling videos for the purpose of uploading to youtube which will further degrade the quality of the already tripe-pile mess you have produced in mind, so by the time youtube has mangled your already steamy pile into a furtherly steamy pile maybe you wont notice the boxes within boxes autist-inspired faux-anti-aliasing........ewww. eww, and ugh.  ugh, and eww.

[claude]

It's not a Linux version as such.  I'm compiling on Linux, and the result (if successful) will be a win32 exe, so you still need wine32 to run it.  Fully porting kf to Linux would be an epic amount of work, and it wouldn't be possible to easily merge in changes from Karl's version when he releases new source zips ( hint hint ).


quaz0r - pretty harsh!  in an ideal world code would be beautiful and documented.  some things I don't like about kf, but some parts I do (automatic glitch corrected perturbation rendering with series approximation).  Ideally the parts I like would be a re-usable cross-platform library, but they're not, so I do the best I can with kf which currently is using kf to render kfb sequences and colouring them / zoom assembling them with my own programs.

[youhn]

I think Karl focussed on just a few main ideas, and was rapid-prototyping those. The rest (readable code, manuals, howto, good program design, gui, etc) was less important. I think he's clear on that, so I don't understand why you're being so harsh.

I just forgot the Windows-centric thing. Also misread the first post here, I see the target is still win32. I just don't understand this effort then.

[quaz0r]

i dont like to be a downer on anyone, especially if they are simply interested in cool things for the fun of it.  this guy makes no bones about it though, hes not in it to make a good program that makes any sense, hes simply in it to win speed records or whatever his original vision was.  i got no love for that.

[claude]

I just don't understand this effort then.

I mainly want to tweak the colouring, so I can get pseudo-de view in kf main window, without having to save as kfb and convert with my tools.


I attached the changes I made so far, they're in a few categories:
* fix case sensitivity issues and windows path separator in #include
* make most (T &x) arguments (const T &x), especially constructors
* make T-returning (as opposed to T&-returning) operators const
* replace __try __except with try catch(...)
* move static new operators into classes

[Kalles Fraktaler]

Hmm...
Unfortunately the time I can spend on fractal programming is too limited.
Rendering requires so much less concentration, it is actually relaxing, so I try to program as little as possible and render all the rest of the time I can spend on this.

I suggest you learn the theory and write your own code, and use the source of KF only as a reference.

Quaz0r, you are wrong and youhn is right.
No further comments.

[claude]

I suggest you learn the theory and write your own code, and use the source of KF only as a reference.

Ok.  I'll work on mightymandel instead.

[Chillheimer]

...I think intentionally obfuscated by karl so as to discourage anyone actually being able to read it or work with it,....he clearly did not intend for any human to actually do anything intelligent with his code.  not to mention the outright dubious design decisions he made with his program.

are you serious?
how about releasing your own work of many hours as freeware to use for everyone?! and then even give away the source code for everyone to play with it!
Or even better, go ahead and do groundbreaking new work on your own instead of crapping on the work of others, only because you don't think they did it in the way you think they should have done it.
If you should ever do great work on your own, and donate it to the public domain, I hope there are tons of people who start bitching around how all you do sucks, just the way you do here!

sorry for my language, but what you say is absolutely inacceptable.
maybe I should just quote yourself:

ewww. eww, and ugh.  ugh, and eww.

[quaz0r]

i'll release my many hours of work as open-source GPL code for all to criticize and do with as they please when i decide it is good enough to release as such.  not any sooner, to break records, or promote my name, or any other dubious reason.  and i will welcome any such criticism, not respond to it in sophomoric knee-jerk fashion, because i am actually interested in producing something of worth, be it by my hand, with the help of others, or by someone else's hand entirely.

[Botond Kósa]

sorry for my language, but what you say is absolutely inacceptable.

I completely agree with Chillheimer.

When I first read quaz0r's rant I was laughing and initially thought Hey, what a great parody of some really furious comment! But by the end, as I realized it was no joke, I felt utter sadness. Quaz0r, you don't seem to realize that people on this forum deal with fractals as their hobbies. This is not our job, we're doing it in our few free hours at night, and share our findings for free. If we also share the source code, take it as a bonus. And if the code is clean and extensible, take it as an even greater bonus. If it is not, deal with it. But remember, we owe you nothing.

[quaz0r]

i completely agree with botand kosa.

[Kalles Fraktaler]

i completely agree with botand kosa.

You spelled Botond's name wrong...

[3dickulus]

people on this forum deal with fractals as their hobbies.

a fractal path driven by curiosity and fueled by sharing discoveries...

But remember, we owe you nothing.

certainly one can not put a value on illumination, it's priceless.

I think you are way off base here, quaz0r, even some of the most poorly written code ( like mine ) can express an idea in a manner that can lead to improvements or better understanding that may not have happened without a view from a different perspective.

i'll release my many hours of work as open-source GPL code for all to criticize and do with as they please when i decide it is good enough to release as such.  not any sooner

the sooner the better I hope it's as intense as Botond's 10k+ lines of assembler or Claude's intricate tapestry of GLSL and custom math code or Syntopia's awesome 3D realization a la Fragmentarium, and of course Kalles Fraktaler who's source code is free of clutter so that you can see the patterns in the coding.

I just can't wait    but I feel it may be some time yet