1 /*
2  * conmakehash.c
3  *
4  * Create arrays for initializing the kernel folded tables (using a hash
5  * table turned out to be to limiting...)  Unfortunately we can't simply
6  * preinitialize the tables at compile time since kfree() cannot accept
7  * memory not allocated by kmalloc(), and doing our own memory management
8  * just for this seems like massive overkill.
9  *
10  * Copyright (C) 1995-1997 H. Peter Anvin
11  *
12  * This program is a part of the Linux kernel, and may be freely
13  * copied under the terms of the GNU General Public License (GPL),
14  * version 2, or at your option any later version.
15  */
16 
17 #include <stdio.h>
18 #include <stdlib.h>
19 #include <sysexits.h>
20 #include <string.h>
21 #include <ctype.h>
22 
23 #define MAX_FONTLEN 256
24 
25 typedef unsigned short unicode;
26 
usage(char * argv0)27 void usage(char *argv0)
28 {
29   fprintf(stderr, "Usage: \n"
30          "        %s chartable [hashsize] [hashstep] [maxhashlevel]\n", argv0);
31   exit(EX_USAGE);
32 }
33 
getunicode(char ** p0)34 int getunicode(char **p0)
35 {
36   char *p = *p0;
37 
38   while (*p == ' ' || *p == '\t')
39     p++;
40   if (*p != 'U' || p[1] != '+' ||
41       !isxdigit(p[2]) || !isxdigit(p[3]) || !isxdigit(p[4]) ||
42       !isxdigit(p[5]) || isxdigit(p[6]))
43     return -1;
44   *p0 = p+6;
45   return strtol(p+2,0,16);
46 }
47 
48 unicode unitable[MAX_FONTLEN][255];
49 				/* Massive overkill, but who cares? */
50 int unicount[MAX_FONTLEN];
51 
addpair(int fp,int un)52 void addpair(int fp, int un)
53 {
54   int i;
55 
56   if ( un <= 0xfffe )
57     {
58       /* Check it isn't a duplicate */
59 
60       for ( i = 0 ; i < unicount[fp] ; i++ )
61 	if ( unitable[fp][i] == un )
62 	  return;
63 
64       /* Add to list */
65 
66       if ( unicount[fp] > 254 )
67 	{
68 	  fprintf(stderr, "ERROR: Only 255 unicodes/glyph permitted!\n");
69 	  exit(EX_DATAERR);
70 	}
71 
72       unitable[fp][unicount[fp]] = un;
73       unicount[fp]++;
74     }
75 
76   /* otherwise: ignore */
77 }
78 
main(int argc,char * argv[])79 int main(int argc, char *argv[])
80 {
81   FILE *ctbl;
82   char *tblname;
83   char buffer[65536];
84   int fontlen;
85   int i, nuni, nent;
86   int fp0, fp1, un0, un1;
87   char *p, *p1;
88 
89   if ( argc < 2 || argc > 5 )
90     usage(argv[0]);
91 
92   if ( !strcmp(argv[1],"-") )
93     {
94       ctbl = stdin;
95       tblname = "stdin";
96     }
97   else
98     {
99       ctbl = fopen(tblname = argv[1], "r");
100       if ( !ctbl )
101 	{
102 	  perror(tblname);
103 	  exit(EX_NOINPUT);
104 	}
105     }
106 
107   /* For now we assume the default font is always 256 characters. */
108   fontlen = 256;
109 
110   /* Initialize table */
111 
112   for ( i = 0 ; i < fontlen ; i++ )
113     unicount[i] = 0;
114 
115   /* Now we come to the tricky part.  Parse the input table. */
116 
117   while ( fgets(buffer, sizeof(buffer), ctbl) != NULL )
118     {
119       if ( (p = strchr(buffer, '\n')) != NULL )
120 	*p = '\0';
121       else
122 	fprintf(stderr, "%s: Warning: line too long\n", tblname);
123 
124       p = buffer;
125 
126 /*
127  * Syntax accepted:
128  *	<fontpos>	<unicode> <unicode> ...
129  *	<range>		idem
130  *	<range>		<unicode range>
131  *
132  * where <range> ::= <fontpos>-<fontpos>
133  * and <unicode> ::= U+<h><h><h><h>
134  * and <h> ::= <hexadecimal digit>
135  */
136 
137       while (*p == ' ' || *p == '\t')
138 	p++;
139       if (!*p || *p == '#')
140 	continue;	/* skip comment or blank line */
141 
142       fp0 = strtol(p, &p1, 0);
143       if (p1 == p)
144 	{
145 	  fprintf(stderr, "Bad input line: %s\n", buffer);
146 	  exit(EX_DATAERR);
147         }
148       p = p1;
149 
150       while (*p == ' ' || *p == '\t')
151 	p++;
152       if (*p == '-')
153 	{
154 	  p++;
155 	  fp1 = strtol(p, &p1, 0);
156 	  if (p1 == p)
157 	    {
158 	      fprintf(stderr, "Bad input line: %s\n", buffer);
159 	      exit(EX_DATAERR);
160 	    }
161 	  p = p1;
162         }
163       else
164 	fp1 = 0;
165 
166       if ( fp0 < 0 || fp0 >= fontlen )
167 	{
168 	    fprintf(stderr,
169 		    "%s: Glyph number (0x%x) larger than font length\n",
170 		    tblname, fp0);
171 	    exit(EX_DATAERR);
172 	}
173       if ( fp1 && (fp1 < fp0 || fp1 >= fontlen) )
174 	{
175 	    fprintf(stderr,
176 		    "%s: Bad end of range (0x%x)\n",
177 		    tblname, fp1);
178 	    exit(EX_DATAERR);
179 	}
180 
181       if (fp1)
182 	{
183 	  /* we have a range; expect the word "idem" or a Unicode range of the
184 	     same length */
185 	  while (*p == ' ' || *p == '\t')
186 	    p++;
187 	  if (!strncmp(p, "idem", 4))
188 	    {
189 	      for (i=fp0; i<=fp1; i++)
190 		addpair(i,i);
191 	      p += 4;
192 	    }
193 	  else
194 	    {
195 	      un0 = getunicode(&p);
196 	      while (*p == ' ' || *p == '\t')
197 		p++;
198 	      if (*p != '-')
199 		{
200 		  fprintf(stderr,
201 "%s: Corresponding to a range of font positions, there should be a Unicode range\n",
202 			  tblname);
203 		  exit(EX_DATAERR);
204 	        }
205 	      p++;
206 	      un1 = getunicode(&p);
207 	      if (un0 < 0 || un1 < 0)
208 		{
209 		  fprintf(stderr,
210 "%s: Bad Unicode range corresponding to font position range 0x%x-0x%x\n",
211 			  tblname, fp0, fp1);
212 		  exit(EX_DATAERR);
213 	        }
214 	      if (un1 - un0 != fp1 - fp0)
215 		{
216 		  fprintf(stderr,
217 "%s: Unicode range U+%x-U+%x not of the same length as font position range 0x%x-0x%x\n",
218 			  tblname, un0, un1, fp0, fp1);
219 		  exit(EX_DATAERR);
220 	        }
221 	      for(i=fp0; i<=fp1; i++)
222 		addpair(i,un0-fp0+i);
223 	    }
224         }
225       else
226 	{
227 	    /* no range; expect a list of unicode values for a single font position */
228 
229 	    while ( (un0 = getunicode(&p)) >= 0 )
230 	      addpair(fp0, un0);
231 	}
232       while (*p == ' ' || *p == '\t')
233 	p++;
234       if (*p && *p != '#')
235 	fprintf(stderr, "%s: trailing junk (%s) ignored\n", tblname, p);
236     }
237 
238   /* Okay, we hit EOF, now output hash table */
239 
240   fclose(ctbl);
241 
242 
243   /* Compute total size of Unicode list */
244   nuni = 0;
245   for ( i = 0 ; i < fontlen ; i++ )
246     nuni += unicount[i];
247 
248   printf("\
249 /*\n\
250  * Do not edit this file; it was automatically generated by\n\
251  *\n\
252  * conmakehash %s > [this file]\n\
253  *\n\
254  */\n\
255 \n\
256 #include <linux/types.h>\n\
257 \n\
258 u8 dfont_unicount[%d] = \n\
259 {\n\t", argv[1], fontlen);
260 
261   for ( i = 0 ; i < fontlen ; i++ )
262     {
263       printf("%3d", unicount[i]);
264       if ( i == fontlen-1 )
265         printf("\n};\n");
266       else if ( i % 8 == 7 )
267         printf(",\n\t");
268       else
269         printf(", ");
270     }
271 
272   printf("\nu16 dfont_unitable[%d] = \n{\n\t", nuni);
273 
274   fp0 = 0;
275   nent = 0;
276   for ( i = 0 ; i < nuni ; i++ )
277     {
278       while ( nent >= unicount[fp0] )
279 	{
280 	  fp0++;
281 	  nent = 0;
282 	}
283       printf("0x%04x", unitable[fp0][nent++]);
284       if ( i == nuni-1 )
285          printf("\n};\n");
286        else if ( i % 8 == 7 )
287          printf(",\n\t");
288        else
289          printf(", ");
290     }
291 
292   exit(EX_OK);
293 }
294