China DOS Union

If the font library is version 5.0, the following piece of code should be removed
if(j==1){ //If it is a Chinese character, adjust the address offset value hz.offset-=0x10000000; }

Hope everyone can like it！！！！！！！

#include "d:\borland\user\src\graphics.h" #include "d:\borland\user\src\graphics.c" #include <string.h> #include <stdio.h> #include <conio.h> #include <math.h> #include <mem.h> #include <alloc.h> #include <dir.h> #include <dos.h> #include <stdlib.h> #define baseSize 250 //The larger this value is, the smaller the font is. The original value is 168 #define portion 100 //100 FILE *fpfh,*fphz,*fpe,*fp; int X0,Y0,xRange,yRange,curX,curY; int newX0,newY0,newxRange,newyRange; unsigned char *wordCont_0=NULL,*wordCont_8=NULL; unsigned bytesNum_0,bytesNum_8; unsigned char myread(void); void control(int *xx,int *yy,unsigned char controlWord,int nowxRange, int nowyRange,int nowX0,int nowY0,int call_8_0); void showhz(char *fontfileName,unsigned long col); void b3_4(int *x,int *y,unsigned long col); void b3_3(int *x,int *y,unsigned long col); int readbyte(void); int readcount; static unsigned char next; //Used to read the flag of the high four bits and low four bits of a byte, 1 for reading the high bits, 0 for reading the low bits typedef struct _hz{ unsigned long offset; //4 bytes unsigned int length; //2 bytes }hz_; hz_ hz; unsigned char hzdata; //Used to read font data void main(void) { unsigned char *hzstring="然字Mm骤啊H骤g9酆8G字,汉，然。!！戆";//戆您原回骤然燃";//"共有汉字个骤";原骤戆啊汉字 long color; initgraph(VBE480_16M); X0=Y0=1;xRange=yRange=152;//The original value is 152 showhz(hzstring,rgbcolor(255,0,0)); getch(); cleardevice(); closegraph(); } void showhz(char *hzstring,unsigned long col) { unsigned long oldcolor; unsigned long wordOffset; unsigned char qm,wm; int bx,by; //Used to record the coordinates of the starting point of the stroke int xx,yy,controlWord,i,j,m=0; char ch,cha,chx,chy,chxx,chyy; char mark1_x,mark1_y,mark2_x,mark2_y,mark3_x,mark3_y; char *fontFileName="HZKPSKTJ";//HZKPSXKJ";//HZKPSST.GBK";//"HZKPSHTJ";//"HZKPSFSJ";//"HZKPSXBJ";HZKPSKTJ;HZKPSY3J;ASCPS;HZKPST;HZKPSDHJ,HZKPSYTJ oldcolor=getcolor(); setcolor(col); if((fpfh=fopen("HZKPST","rb"))==NULL) { //Open Chinese symbol curve font library printf("Cannot open %s file\n","HZKPST"); exit(0); } if((fphz=fopen(fontFileName,"rb"))==NULL) { //Open Chinese curve font library, if it is a GBK font library, then open the GBK font library printf("Cannot open %s file\n",fontFileName); exit(0); } if((fpe=fopen("ASCPS","rb"))==NULL) { //Open English curve font library printf("Cannot open %s file\n","ASCPS"); exit(0); } i=0; while(i<strlen(hzstring)) { readcount=0; next=0; if((hzstring&0x80)==0x80){ //Judge whether it is a Chinese character, if it is equal, that is, the highest bit is 1, then it is a Chinese character qm=hzstring-0xa0; //Area code wm=hzstring-0xa0; //Bit code j=1; //As a flag of Chinese characters, used to adjust the address offset if(strcmp(strrchr(fontFileName, '.')+1,"GBK")==0){ //First find the position of the last '.', add 1 to make the pointer point to the first character after '.', then compare, judge whether it is a GBK font library fp=fphz; //GBK font library still has some missing Chinese characters, such as the character "镕", which does not exist in the font library wordOffset=(((long)qm-1)*94+(wm-1))*6; //GBK font library has included the Chinese symbol font library before 16 areas. When calculating the address, the area code should be subtracted by 1 instead of 6 } else { //Not a GBK font library if(qm>=16){ //Chinese characters after 16 areas fp=fphz; wordOffset=(((long)qm-16)*94+(wm-1))*6; //Offset of Chinese characters after 16 areas } else{ //Chinese symbols before 16 areas fp=fpfh; wordOffset=(((long)qm-1)*94+(wm-1))*6; //Offset of Chinese symbols before 16 areas } } i+=2; //Chinese characters are two bytes } else{ //It is an English character fp=fpe; wordOffset=(0*94+(hzstring-32)-1)*6; //There are ten types of English fonts 0~9. Let the font number be N, and the ASCII code of the character be CC, then the offset of the English character is (N*94+(CC-32)-1)*6. The ASCII code value of the space is 32, but it does not exist in the font library. The starting character of the font library is the character '!', with a value of 33. Actually, it should be subtracted by 33. There are 94 characters in each type of font in the font library i+=1; } fseek(fp,wordOffset,SEEK_SET); fread(&hz,sizeof(hz_),1,fp); if(j==1){ //If it is a Chinese character, adjust the address offset value hz.offset-=0x10000000; j=0; } fseek(fp,hz.offset,SEEK_SET); fread(hzdata,1,hz.length,fp); while(readcount<hz.length) { ch=myread(); //Take the control code switch(ch){ case 0: //New starting point 0000 if(readcount>0)lineto(bx,by); //Do not connect when the first starting point of the entire character, connect the last point of the stroke with the starting point, and close the stroke xx=readbyte(); yy=readbyte(); xx=xRange*xx/baseSize+0.5; yy=yRange*yy/baseSize+0.5; xx+=X0;yy+=Y0; moveto(xx,yy); //Stroke starting point curX=xx;curY=yy; bx=xx;by=yy; //Record the coordinates of the starting point of the stroke break; case 1: xx=readbyte(); xx=xRange*xx/baseSize+0.5; xx+=X0; lineto(xx,curY); curX=xx; break; case 2: yy=readbyte(); yy=yRange*yy/baseSize+0.5; yy+=Y0; lineto(curX,yy); curY=yy; break; case 3: xx=readbyte(); yy=readbyte(); xx=xRange*xx/baseSize+0.5; yy=yRange*yy/baseSize+0.5; xx+=X0;yy+=Y0; lineto(xx,yy); curX=xx;curY=yy; break; case 4: xx=curX;yy=curY; xx=readbyte(); yy=readbyte(); xx=readbyte(); yy=readbyte(); xx=xRange*xx/baseSize+0.5; yy=yRange*yy/baseSize+0.5; xx=xRange*xx/baseSize+0.5; yy=yRange*yy/baseSize+0.5; xx+=X0;yy+=Y0; xx+=X0;yy+=Y0; b3_3(xx,yy,COLOR); curX=xx;curY=yy; moveto(curX,curY); break; case 5: xx=curX;yy=curY; xx=readbyte(); yy=readbyte(); xx=readbyte(); yy=readbyte(); xx=readbyte(); yy=readbyte(); xx=xRange*xx/baseSize+0.5; yy=yRange*yy/baseSize+0.5; xx=xRange*xx/baseSize+0.5; yy=yRange*yy/baseSize+0.5; xx=xRange*xx/baseSize+0.5; yy=yRange*yy/baseSize+0.5; for(m=1;m<4;m++) { xx+=X0;yy+=Y0; } b3_4(xx,yy,COLOR); curX=xx;curY=yy; moveto(curX,curY); break; case 6: xx=readbyte(); yy=readbyte(); xx=readbyte(); yy=readbyte(); xx=xRange*xx/baseSize+0.5; yy=yRange*yy/baseSize+0.5; xx=xRange*xx/baseSize+0.5; yy=yRange*yy/baseSize+0.5; for(m=0;m<2;m++) { xx+=X0;yy+=Y0; } rectangle(xx,yy,xx,yy); break; case 7: mark1_x=1; ch=myread(); if(ch&8)mark1_x=-1; ch&=7; xx=curX+ch*mark1_x*xRange/baseSize+0.5; yy=readbyte(); yy=yRange*yy/baseSize+0.5; yy+=Y0; lineto(xx,yy); curX=xx;curY=yy; break; case 8: mark1_y=1; xx=readbyte(); xx=xRange*xx/baseSize+0.5; xx+=X0; ch=myread(); if(ch&8)mark1_y=-1; ch&=7; yy=curY+ch*mark1_y*yRange/baseSize+0.5; lineto(xx,yy); curX=xx;curY=yy; break; case 9: mark1_x=1; mark1_y=1; ch=myread(); if(ch&8)mark1_x=-1; ch&=7; xx=curX+ch*mark1_x*xRange/baseSize+0.5; ch=myread(); if(ch&8)mark1_y=-1; ch&=7; yy=curY+ch*mark1_y*yRange/baseSize+0.5; lineto(xx,yy); curX=xx;curY=yy; break; case 10: mark1_x=1; mark1_y=1; ch=myread(); //The first four bits read are the high four bits of the six-bit data cha=myread(); //The high two bits of the four bits read the second time are the low two bits of the six-bit data, and the low two bits are the high two bits of the second six-bit data ch=(ch<<2)+(cha>>2); if(ch&0x20)mark1_x=-1; ch&=0x1f; xx=curX+ch*mark1_x*xRange/baseSize+0.5; ch=myread(); //The four bits read the third time are the low four bits of the second six-bit data ch=((cha&0x03)<<4)+ch; if(ch&0x20)mark1_y=-1; ch&=0x1f; yy=curY+ch*mark1_y*yRange/baseSize+0.5; lineto(xx,yy); curX=xx;curY=yy; break; case 11: mark1_x=1; mark1_y=1; mark2_x=1; mark2_y=1; xx=curX;yy=curY; ch=myread(); if(ch&8)mark1_x=-1; chx=ch=(ch&7)*mark1_x; xx=curX+ch*xRange/baseSize+0.5; ch=myread(); if(ch&8)mark1_y=-1; chy=ch=(ch&7)*mark1_y; yy=curY+ch*yRange/baseSize+0.5; ch=myread(); if(ch&8)mark2_x=-1; ch=(ch&7)*mark2_x; xx=curX+(chx+ch)*xRange/baseSize+0.5; ch=myread(); if(ch&8)mark2_y=-1; ch=(ch&7)*mark2_y; yy=curY+(chy+ch)*yRange/baseSize+0.5; b3_3(xx,yy,COLOR); curX=xx;curY=yy; moveto(curX,curY); break; case 12: mark1_x=1; mark1_y=1; mark2_x=1; mark2_y=1; xx=curX;yy=curY; ch=myread(); //1 cha=myread(); //2 ch=(ch<<2)+(cha>>2); if(ch&0x20)mark1_x=-1; chx=ch=(ch&0x1f)*mark1_x; xx=curX+ch*xRange/baseSize+0.5; ch=myread(); //3 ch=((cha&0x03)<<4)+ch; if(ch&0x20)mark1_y=-1; chy=ch=(ch&0x1f)*mark1_y; yy=curY+ch*yRange/baseSize+0.5; ch=myread(); //1 cha=myread(); //2 ch=(ch<<2)+(cha>>2); if(ch&0x20)mark2_x=-1; ch=(ch&0x1f)*mark2_x; xx=curX+(chx+ch)*xRange/baseSize+0.5; ch=myread(); //3 ch=((cha&0x3)<<4)+ch; if(ch&0x20)mark2_y=-1; ch=(ch&0x1f)*mark2_y; yy=curY+(chy+ch)*yRange/baseSize+0.5; b3_3(xx,yy,COLOR); curX=xx;curY=yy; moveto(curX,curY); break; case 13: mark1_x=1; mark1_y=1; mark2_x=1; mark2_y=1; mark3_x=1; mark3_y=1; xx=curX;yy=curY; ch=myread(); if(ch&8)mark1_x=-1; chx=ch=(ch&7)*mark1_x; xx=curX+ch*xRange/baseSize+0.5; ch=myread(); if(ch&8)mark1_y=-1; chy=ch=(ch&7)*mark1_y; yy=curY+ch*yRange/baseSize+0.5; ch=myread(); if((ch)&8)mark2_x=-1; chxx=ch=(ch&7)*mark2_x; xx=curX+(chx+ch)*xRange/baseSize+0.5; ch=myread(); if((ch)&8)mark2_y=-1; chyy=ch=(ch&7)*mark2_y; yy=curY+(chy+ch)*yRange/baseSize+0.5; ch=myread(); if((ch)&8)mark3_x=-1; ch=(ch&7)*mark3_x; xx=curX+(chx+chxx+ch)*xRange/baseSize+0.5; ch=myread(); if((ch)&8)mark3_y=-1; ch=(ch&7)*mark3_y; yy=curY+(chy+chyy+ch)*yRange/baseSize+0.5; b3_4(xx,yy,COLOR); curX=xx;curY=yy; moveto(curX,curY); break; case 14: mark1_x=1; mark1_y=1; mark2_x=1; mark2_y=1; mark3_x=1; mark3_y=1; xx=curX;yy=curY; ch=myread(); //1 cha=myread(); //2 ch=(ch<<2)+(cha>>2); if(ch&0x20)mark1_x=-1; chx=ch=(ch&0x1f)*mark1_x; xx=curX+ch*xRange/baseSize+0.5; ch=myread(); //3 ch=((cha&0x03)<<4)+ch; if(ch&0x20)mark1_y=-1; chy=ch=(ch&0x1f)*mark1_y; yy=curY+ch*yRange/baseSize+0.5; ch=myread(); //1 cha=myread(); //2 ch=(ch<<2)+(cha>>2); if(ch&0x20)mark2_x=-1; chxx=ch=(ch&0x1f)*mark2_x; xx=curX+(chx+ch)*xRange/baseSize+0.5; ch=myread(); //3 ch=((cha&0x03)<<4)+ch; if(ch&0x20)mark2_y=-1; chyy=ch=(ch&0x1f)*mark2_y; yy=curY+(chy+ch)*yRange/baseSize+0.5; ch=myread(); //1 cha=myread(); //2 ch=(ch<<2)+(cha>>2); if(ch&0x20)mark3_x=-1; ch=(ch&0x1f)*mark3_x; xx=curX+(chx+chxx+ch)*xRange/baseSize+0.5; ch=myread(); //3 ch=((cha&0x03)<<4)+ch; if(ch&0x20)mark3_y=-1; ch=(ch&0x1f)*mark3_y; yy=curY+(chy+chyy+ch)*yRange/baseSize+0.5; b3_4(xx,yy,COLOR); curX=xx;curY=yy; moveto(curX,curY); break; case 15: //Only read two bytes, no processing readbyte(); readbyte(); break; } } if((readcount<hz.length))lineto(bx,by); //Connect the last point of the last stroke with the starting point to close the stroke getch(); X0+=xRange*168/baseSize+0.5; //Here 168 is more appropriate, but the reason has not been found if(getmaxx()-X0<xRange*168/baseSize){ X0=0; Y0+=yRange*168/baseSize+0.5; } if(getmaxy()-Y0<yRange*168/baseSize){ Y0=0; X0=0; getch(); cleardevice(); } } setcolor(oldcolor); fp=0; fclose(fpe); fclose(fpfh); fclose(fphz); return; } unsigned char myread(void) //Each time it is called, a four-bit number of a byte is read, first read the low four bits, then read the high four bits { unsigned char a; if(!next){ //When next is 0, read the low four bits, and when it is 1, read the high four bits a=hzdata&0x0f; next=1; } else{ a=(hzdata&0xf0)>>4; readcount++; next=0; } return a; } int readbyte(void) { int kk=0; unsigned char a,b; a=myread()<<4; b=myread(); kk=(int)(b+a); return kk; } void b3_4(int *x,int *y,unsigned long col) //Cubic B-spline curve, n can specify a constant { int n,*bx,*by,sign=1,j=0,k,i; double a,b,c,d,dt,xx,f0_3,f_0,f_1,f_2; n=xx=sqrt(((double)(x-x))*((double)(x-x))+ ((double)(y-y))*((double)(y-y)))+ sqrt(((double)(x-x))*((double)(x-x))+ ((double)(y-y))*((double)(y-y)))+ sqrt(((double)(x-x))*((double)(x-x))+ ((double)(y-y))*((double)(y-y))); if(xx>n) n++; n*=2; if(n==0) n=20; do { bx=(int *)malloc(n*sizeof(int)); by=(int *)malloc(n*sizeof(int)); if((!bx)||(!by)) { /*printf("Memory Alloction Error!");*/ return; } dt=(double)1/(double)n; a=((-1)*x+3*x-3*x+x); b=3*(x-2*x+x); c=3*(x-x); d=x; f0_3=6*a*dt*dt*dt; f_2=f0_3+2*b*dt*dt; f_1=f0_3/6+b*dt*dt+c*dt; f_0=d; bx=f_0+0.5; for(i=0;i<n;i++) { f_0+=f_1; bx=f_0+0.5; f_1+=f_2; f_2+=f0_3; } a=((-1)*y+3*y-3*y+y); b=3*(y-2*y+y); c=3*(y-y); d=y; f0_3=6*a*dt*dt*dt; f_2=f0_3+2*b*dt*dt; f_1=f0_3/6+b*dt*dt+c*dt; f_0=d; bx=f_0+0.5; for(i=0;i<n;i++) { f_0+=f_1; by=f_0+0.5; f_1+=f_2; f_2+=f0_3; } sign=1;j=0;k=n/100+1; for(i=1;i<n;i++) { if((abs(bx-bx)>1)||(abs(by-by)>1)) j++; if(j>k) { free(bx); free(by); if(n>3) n=n+n/3; else n=6; sign=0; } break; } } while(sign==0); for(i=1;i<n;i++) if((bx!=bx)||(by!=by)) putpixel(bx,by,col); free(bx); free(by); } void b3_3(int *x,int *y,unsigned long col) //Quadratic B-spline curve, n can specify a constant { int n,*bx,*by,sign=1,j=0,k,i; double a,b,c,d,dt,xx,f0_3,f_0,f_1,f_2; n=xx=sqrt(((double)(x-x))*((double)(x-x))+ ((double)(y-y))*((double)(y-y)))+ sqrt(((double)(x-x))*((double)(x-x))+ ((double)(y-y))*((double)(y-y))); if(xx>n) n++; n*=4; if(n==0) n=20; do { bx=(int *)malloc(n*sizeof(int)); by=(int *)malloc(n*sizeof(int)); if((!bx)||(!by)) { /*printf("Memory Alloction Error!");*/ return; } dt=(double)1/(double)n; b=x-2*x+x; c=2*(x-x); d=x; f_2=2*b*dt*dt; f_1=b*dt*dt+c*dt; f_0=d; bx=f_0+0.5; for(i=0;i<n;i++) { f_0+=f_1; bx=f_0+0.5; f_1+=f_2; } b=y-2*y+y; c=2*(y-y); d=y; f_2=2*b*dt*dt; f_1=b*dt*dt+c*dt; f_0=d; by=f_0+0.5; for(i=0;i<n;i++) { f_0+=f_1; by=f_0+0.5; f_1+=f_2; } sign=1;j=0;k=n/100+1; for(i=1;i<n;i++) { if((abs(bx-bx)>1)||(abs(by-by)>1)) j++; if(j>k) { free(bx); free(by); if(n>3) n=n+n/3; else n=6; sign=0; } break; } } while(sign==0); for(i=1;i<n;i++) if((bx!=bx)||(by!=by)) { putpixel(bx,by,col); } free(bx); free(by); }

Recently, out of whim, I started looking at UCDOS's rdfnt.com. Although the knowledge of DOS has been neglected for many years, and the information from back then is basically impossible to look up because it has been put aside (it seems I didn't have any UCDOS materials back then, and now I can't find things like the UCDOS system call specifications even after searching Google for a few days), I still spent several days looking at it out of interest. I mainly looked at the dot matrix font library, and only looked at the part where control codes are converted to coordinates in PS. Although I didn't finish reading it, I once again marveled at the delicacy of assembly programs. A program of about 30KB, among which there are more than 10KB of working storage areas, 4-5KB of various font library parameter definitions, and the instruction part is estimated to be only more than 10KB. Such a short program can complete operations such as retrieval, matching, decoding, and transformation of dot matrix, vector, and PS font libraries, which is really admirable.

I happened to see the post of the thread starter and want to talk about a little bit of my own opinion.

The beginning of the article has

If the font library is version 5.0, the following segment of code should be removed
　if(j==1){ //If it is a Chinese character, adjust the address offset value
　 hz.offset-=0x10000000;
　}

According to my understanding, the offset of Chinese character data should not be subtracted by 0x10000000. This offset should actually be only 24 bits, that is, 3 bytes. The 4th byte does not define the offset, but the additional format of this Chinese character data. There is an operation in a process not far from the acquisition of the Chinese character dot matrix function entered from int 7eh into rdfnt:
　mov　　　ax,es:　The lower word of the offset of the just-read 6-byte PS or vector data
　mov　　　dx,es:　The higher word of the offset of the PS or vector data
　mov　　　cx,es:　Length
　jcxz　　　...
　or　 　　 zk_type,dh
　mov　　　dh,0　　　　　Clear the highest byte of the data pointer, leaving 3 bytes
...
　call　　　read_hz　　　　Move the file pointer, dx, ax = offset, read the file, read cx bytes into ds:di
It can be seen that the highest byte of the 4 bytes is not used for addressing, but for defining information. In addition to reading the dot matrix font library and decoding the PS font library, rdfnt can also decode the vector font library. For different PS and vector font libraries, it is possible that this byte of some font libraries is not 0x10 (I haven't checked each Chinese character of each font library one by one, so I can't confirm). If this fixed value is subtracted from all, it may go wrong.

There is an article called "Analysis of UCDOS's Curve Outline Font Library", which also mentions that the offset in ucdos6 should be subtracted by 0x10000000. I wonder if the subtraction of 0x10000000 here by the thread starter is derived from that place?

Thanks to the guidance from the person above,

The value "0x10000000" indeed comes from the article "Analysis of the Curve Outline Font Library of UCDOS", but I checked the 5.0 font library and the 6.0 font library, and it is true that a definition was added in the fourth byte of the 6.0 curve font library, but I don't know its actual use.

Still hope that the experts can give guidance.

Okay, learned it, thank you

unsigned char *hzstring="然字Mm骤啊H骤g9酆8G字,汉，然。!！戆";//戆您原回骤然燃";//"共有汉字个骤";原骤戆啊汉字
What is this?