Serious code quality issues

[bradenbest]

• - 1 file (which wouldn't be so bad if not for...)
• - 3 kLOC (should be ~500 SLOC per file at the most)
• - functions are hundreds of lines long (should be < 15)
• - functions each take 20+ arguments (should be < 5)
• - max nest depth averaging 7-8 tabs (cyclomatic complexity = fucked)
• - tons of global variables
• - using namespace (you clearly didn't care about naming collisions)
• - no cohesion
• - no encapsulation
• - ugly
• - repetitive
• - bleeds memory
• and many, many more...

This is some of the absolute worst, most irresponsibly written code I've ever read. No wonder the project's been dead for years. There was no way you would have been able to maintain this.

Also, just because the following was one of the first things I saw while speed-scrolling through the source file...

val[0] = *(((char*)&(xf))+7);
val[1] = *(((char*)&(xf))+6);
val[2] = *(((char*)&(xf))+5);
val[3] = *(((char*)&(xf))+4);
val[4] = *(((char*)&(xf))+3);
val[5] = *(((char*)&(xf))+2);
val[6] = *(((char*)&(xf))+1);
val[7] = *(((char*)&(xf))+0);
write(of,val,8);
val[0] = *(((char*)&(zf))+7);
val[1] = *(((char*)&(zf))+6);
val[2] = *(((char*)&(zf))+5);
val[3] = *(((char*)&(zf))+4);
val[4] = *(((char*)&(zf))+3);
val[5] = *(((char*)&(zf))+2);
val[6] = *(((char*)&(zf))+1);
val[7] = *(((char*)&(zf))+0);
write(of,val,8);
val[0] = *(((char*)&(yf))+7);
val[1] = *(((char*)&(yf))+6);
val[2] = *(((char*)&(yf))+5);
val[3] = *(((char*)&(yf))+4);
val[4] = *(((char*)&(yf))+3);
val[5] = *(((char*)&(yf))+2);
val[6] = *(((char*)&(yf))+1);
val[7] = *(((char*)&(yf))+0);
write(of,val,8);

There is a reason we have functions like memcpy and fwrite.

[TroZ]

Hi,

You are only the third person to contact me about the code to DF2MC, with one of the others being the author of DFHack which DF2MC uses.

I can't disagree with you.

I'm not a C coder, I know some C but am far from an expert in it. Java is the language I know best.

I wrote this mostly to see if it can be done, and to attempt to convert one of my fortresses. It worked well enough for me, and apparently a few other people, but I know there are many issues in the code.

Interest in this seemed to disappear, mine and others, a short wile after I released it, although when the next version of DF came out (about a year later), there were a few requests for an update. I attempted to update the code, but the way DFHack worked had changed significantly, and I was having issues just getting it to compile without modifying any code. Additionally, by that time, Minecraft had changed how it was dealing with saves, and I couldn't find any C libraries for writing the new save format, and getting the old save format correct was one of the hardest issues I had with the original version. With not much progress in updating the code, and with no offers of assistance up to that point, I lost interest and moved on to other things.

If you would like to rewrite DF2MC, feel free. I can discuss the general algorithm with you if you want. I actually haven't played Dwarf Fortress in a few years, so there are probably a few new situations that will have to be handled (off the top of my head, trees used to occupy only two locations, one on the ground and then space above that, but that is no longer the case as far as I know). The general algorithm can probably still work, but there will probably have to be new materials / items added, and maybe some code for some new special cases. Additionally, Minecraft has changed a bunch as well, and that will likely allow more accurate representation of the Dwarf Fortress world. The biggest change is a higher build height, which allows more layers form the DF world to be copied over, possibly eliminating the need to remove some layers (the original DF2CM removed 'uninteresting layers', ones that didn't have 'airspace' other than stairs, so that more interesting layers could be kept in the limited area that the Minecraft world allowed (only about 42 layers, 126 Minecraft blocks tall (3 MC blocks per DF layer))). However, Minecraft now has many more objects such as armor stands, pistons, more plant and wood types, etc., that would make for a more interesting and accurate conversion from Dwarf Fortress.