/* Copyright (c) 1997-1998 by Juliusz Chroboczek */
#include <sys/types.h>
#include <unistd.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "types.h"
#include "ttftool.h"
#include "externs.h"
struct TableDirectoryEntry
*readDirectory(int fd, struct OffsetTable *ot)
{
unsigned n;
int i;
struct TableDirectoryEntry *td;
surely_read(fd,ot,sizeof(struct OffsetTable));
FIX_OffsetTable(*ot);
msg(2,"%d tables\n",ot->numTables);
n=sizeof(struct TableDirectoryEntry)*ot->numTables;
td=mymalloc(n);
surely_read(fd,td,n);
for(i=0; i<ot->numTables; i++)
FIX_TableDirectoryEntry(td[i]);
return td;
}
char **
readNamingTable(int fd)
{
TTF_USHORT format;
TTF_USHORT nrecords;
off_t position;
TTF_USHORT offset;
int i,index,maxIndex;
struct NameRecord *records;
char *data;
char **strings;
position=surely_lseek(fd,0,SEEK_CUR);
surely_read(fd,&format,sizeof(TTF_USHORT));
FIX_UH(format);
if(format!=0)
ttf_fail("Bad TTF file\n");
surely_read(fd,&nrecords,sizeof(TTF_USHORT));
FIX_UH(nrecords);
surely_read(fd,&offset,sizeof(TTF_USHORT));
FIX_UH(offset);
records=mymalloc(nrecords*sizeof(struct NameRecord));
surely_read(fd,records,nrecords*sizeof(struct NameRecord));
for(i=0,maxIndex=-1;i<nrecords;i++) {
FIX_NameRecord(records[i]);
index=records[i].offset+records[i].length;
maxIndex=maxIndex>index?maxIndex:index;
}
data=mymalloc(maxIndex);
surely_lseek(fd,position+offset,SEEK_SET);
surely_read(fd,data,maxIndex);
strings=mymalloc(8*sizeof(char*));
for(i=0;i<8;i++)
strings[i]=NULL;
for(i=0; i<nrecords; i++) {
if(records[i].platformID==3 && /* Microsoft */
records[i].encodingID==1 && /* UGL */
records[i].languageID==0x0409 && /* US English */
records[i].nameID<=7) {
strings[records[i].nameID]=mymalloc(records[i].length/2+1);
unistrncpy(strings[records[i].nameID],
data+records[i].offset,
records[i].length);
msg(2,"%d: %s\n",records[i].nameID,
strings[records[i].nameID]);
}
}
free(records);
free(data);
return strings;
}
void
readHeadTable(int fd, struct HeadTable *ht)
{
surely_read(fd,ht,sizeof(struct HeadTable));
FIX_HeadTable(*ht);
msg(2," version %d.%d", ht->version.mantissa,ht->version.fraction);
msg(2," font revision %d.%d",ht->fontRevision.mantissa, ht->fontRevision.fraction);
if(ht->magicNumber!=0x5F0F3CF5)
ttf_fail("Bad magic number in TTF file");
msg(2," %d units per Em\n",ht->unitsPerEm);
}
void read_font(FILE* fontfile)
{
long i, j, k, l, n, m, platform_id, encoding_id;
TTF_FWORD kern_value;
char buf[1024], *p;
dirtab_entry *pd;
kern_entry *pk;
mtx_entry *pm;
long int netabs, length;
cmap_entry *cmap_tab, *e;
seg_entry *seg_tab, *s;
long cmap_offset;
TTF_USHORT *glyphId, format, segCount;
msg(3, "Reading font");
ttf_skip(TTF_FIXED_L_SIZE);
ntabs = get_ushort();
ttf_skip(3*TTF_USHORT_SIZE);
msg(3, "Reading directory table [%d entries]", ntabs);
dir_tab = ttf_alloc(ntabs, dirtab_entry);
for (pd = dir_tab; pd - dir_tab < ntabs; pd++) {
pd->tag[0] = get_char();
pd->tag[1] = get_char();
pd->tag[2] = get_char();
pd->tag[3] = get_char();
ttf_skip(TTF_ULONG_SIZE);
pd->offset = get_ulong();
pd->length = get_ulong();
}
msg(3, "Reading head table");
seek_tab("head", 2*TTF_FIXED_L_SIZE + 2*TTF_ULONG_SIZE + TTF_USHORT_SIZE);
upem = get_ushort();
ttf_skip(16);
FontBBox1 = get_fword();
FontBBox2 = get_fword();
FontBBox3 = get_fword();
FontBBox4 = get_fword();
ttf_skip(TTF_USHORT_SIZE);
ttf_skip(TTF_USHORT_SIZE + TTF_SHORT_SIZE);
loca_format = get_short();
msg(3, "loca table format %d", loca_format);
msg(3,"Reading maxp table");
seek_tab("maxp", TTF_FIXED_L_SIZE);
nglyphs = get_ushort();
msg(3, "%d entries", nglyphs);
mtx_tab = ttf_alloc(nglyphs + 1, mtx_entry); /*need 1 for the loca table*/
/* set initially to 0, so that we can test whether an entry was used
by testing for 0
memset(mtx_tab, 0, nglyphs*sizeof(mtx_entry)); */
for (pm = mtx_tab; pm - mtx_tab < nglyphs; pm++) {
pm->name = 0; /* notdef */
pm->found = 0;
}
msg(3, "Reading hhea table");
seek_tab("hhea", TTF_FIXED_L_SIZE);
Ascender = get_fword();
Descender = get_fword();
ttf_skip(TTF_FWORD_SIZE + TTF_UFWORD_SIZE + 3*TTF_FWORD_SIZE + 8*TTF_SHORT_SIZE);
nhmtx = get_ushort();
msg(3, "Reading hmtx table");
seek_tab("hmtx", 0);
for (pm = mtx_tab; pm - mtx_tab < nhmtx; pm++) {
pm->wx = get_ufword();
ttf_skip(TTF_FWORD_SIZE);
}
i = pm[-1].wx;
for (; pm - mtx_tab < nglyphs; pm++)
pm->wx = i;
msg(3, "Reading post table");
seek_tab("post", 0);
post_format = get_fixed();
msg(3, "post format %d", post_format);
ItalicAngle = get_fixed();
UnderlinePosition = get_fword();
UnderlineThickness = get_fword();
IsFixedPitch = get_ulong();
/*ttf_skip(4*TTF_ULONG_SIZE);*/
minMemType42 = get_ulong();
maxMemType42 = get_ulong();
minMemType1 = get_ulong();
maxMemType1 = get_ulong();
switch (post_format) {
case 0x00010000:
for (pm = mtx_tab; pm - mtx_tab < NMACGLYPHS; pm++)
pm->name = mac_glyph_names[pm - mtx_tab];
break;
case 0x00020000:
l = get_ushort(); /* some fonts have this value different from nglyphs */
msg(3, "post table contains %d entries", l);
for (pm = mtx_tab; pm - mtx_tab < l; pm++)
pm->index = get_ushort();
if ((pd = name_lookup("post")) == 0)
ttf_fail("can't find table `post'");
n = pd->length - (ftell(fontfile) - pd->offset);
ps_glyphs_buf = ttf_alloc(n + 1, char);
for (m = 0, p = ps_glyphs_buf; p - ps_glyphs_buf < n;) {
for (i = get_byte(); i > 0; i--)
*p++ = get_char();
*p++ = 0;
m++;
}
/* now we retrieve the names that are in the font table*/
for (pm = mtx_tab; pm - mtx_tab < l; pm++) {
if (pm->index < NMACGLYPHS) {
pm->name = mac_glyph_names[pm->index];
}
else {
k = pm->index - NMACGLYPHS;
if (k <= m) {
for (p = ps_glyphs_buf; k > 0; k--)
p = (char *)strend(p) + 1;
pm->name = p;
}
else {
warn(1,"index 0x%04x out of range (glyphs: 0x%04x, m=0x%04x)", k,l,m);
pm->name = 0; /* index out of valid range, fix name to notdef */
}
}
}
/*here comes the ucs->uni hack*/
msg(3, "Reading cmap table");
seek_tab("cmap", TTF_USHORT_SIZE); /* skip the table vesrion number (0) */
netabs = get_ushort();
msg(3, "%d entries", netabs);
cmap_offset = ftell(fontfile) - 2*TTF_USHORT_SIZE;
cmap_tab = ttf_alloc(netabs, cmap_entry);
for (e = cmap_tab; e - cmap_tab < netabs; e++) {
e->platform_id = get_ushort();
e->encoding_id = get_ushort();
e->offset = get_ulong();
}
for (e = cmap_tab; e - cmap_tab < netabs; e++) {
seek_off("cmap", cmap_offset + e->offset);
format = get_ushort();
if (format != 4) {
msg(4, "found table in unsupported format %d, skipping.", format);
continue;
}
msg(3, "format=%d (should be 4)", format);
/*uni_to_glyph = 1;*/
length = get_ushort(); /* length of subtable in bytes */
get_ushort(); /* skip the version number */
segCount = get_ushort();
segCount /= 2;
msg(3, "segment count %d, length %d", segCount, length);
get_ushort(); /* skip searchRange */
get_ushort(); /* skip entrySelector */
get_ushort(); /* skip rangeShift */
seg_tab = ttf_alloc(segCount, seg_entry);
for (s = seg_tab; s - seg_tab < segCount; s++)
{
s->endCode = get_ushort();
msg(4, "segment end 0x%04x", s->endCode);
}
get_ushort(); /* skip reversedPad */
for (s = seg_tab; s - seg_tab < segCount; s++)
{
s->startCode = get_ushort();
msg(4, "segment start 0x%04x", s->startCode);
}
for (s = seg_tab; s - seg_tab < segCount; s++)
{
s->idDelta = get_ushort();
msg(4, "segment delta 0x%04x", s->idDelta);
}
for (s = seg_tab; s - seg_tab < segCount; s++)
{
s->idRangeOffset = get_ushort();
msg(4, "segment range offset 0x%04x", s->idRangeOffset);
}
length -= 8*TTF_USHORT_SIZE + 4*segCount*TTF_USHORT_SIZE;
length /= TTF_USHORT_SIZE; /* conver byte length to number if items */
msg(3, "should contain %d glyphs", length);
glyphId = ttf_alloc(length, TTF_USHORT);
for (i = 0; i < length; i++)
{
msg(5, "i=%d", i);
glyphId[i] = get_ushort();
}
msg(3, "retrieved glyphs");
/* set uni_map to 0, so that we can test for duplicate usage */
for (i = 0; i <= MAX_CHAR_CODE; i++)
{
uni_map[i] = 0;
uni_to_glyph[i] = 0;
}
/* the following loop translates the unicode value i into the glyph index k*/
for (i = 0; i <= MAX_CHAR_CODE; i++)
{
for (s = seg_tab; s - seg_tab < segCount; s++)
if (s->endCode >= i)
break;
if (s - seg_tab < segCount && s->startCode <= i) {
if (s->idRangeOffset != 0) {
/*
we should not need to do this, but there seem to be
some broken fonts around that can produce n outside
of the bounds of the glyphId table, making us to
SIGSEGV
*/
n = (i-s->startCode) + s->idRangeOffset/2 - (segCount-(s-seg_tab));
if(n < length)
k = glyphId[n];
else
k = 0;
if (k != 0)
k = (k + s->idDelta) % 0xFFFF;
}
else
k = (s->idDelta + i) % 0xFFFF;
}
else
k = 0; /*not found*/
if(uni_map[k] == 0 || k == 0)
{
uni_map[k] = i;
}
else
warn(1,"warning: duplicate ucs; index %d, ucs 0x%04x, prev. ucs 0x%04x\n", k, i, uni_map[k]);
uni_to_glyph[i] = k;
}
for (i = 0; i < nglyphs; i++)
mtx_tab[i].uni = uni_map[i];
}
break;
case 0x00030000:
if (print_index == 0) {
warn(1,"no names available in `post' table, printing by index forced");
print_index = 2;
}
break;
default:
ttf_fail("unsupported format (%.8X) of `post' table", post_format);
}
msg(3, "Reading loca table");
seek_tab("loca", 0);
for (pm = mtx_tab; pm - mtx_tab <= nglyphs; pm++) /*loca table contains nglyphs+1 entries*/
pm->offset = (loca_format == 1 ? get_ulong() : get_ushort() << 1);
msg(3, "Reading glyf table");
if ((pd = name_lookup("glyf")) == 0)
ttf_fail("can't find table `glyf'");
for (n = pd->offset, pm = mtx_tab; pm - mtx_tab < nglyphs; pm++)
{
msg(5, "indx %d of %d [off 0x%04x, next off 0x%04x]", pm - mtx_tab, nglyphs-1,n + pm->offset, n + (pm+1)->offset);
if (pm->offset != (pm + 1)->offset) {
seek_off("glyf", n + pm->offset);
ttf_skip(TTF_SHORT_SIZE);
pm->bbox[0] = get_fword();
msg(6, "got bbox[0]");
pm->bbox[1] = get_fword();
msg(6, "got bbox[1]");
pm->bbox[2] = get_fword();
msg(6, "got bbox[2]");
pm->bbox[3] = get_fword();
msg(6, "got bbox[3]");
}
else { /* get the BBox from .notdef */
pm->bbox[0] = mtx_tab[0].bbox[0];
pm->bbox[1] = mtx_tab[0].bbox[1];
pm->bbox[2] = mtx_tab[0].bbox[2];
pm->bbox[3] = mtx_tab[0].bbox[3];
}
}
msg(3, "Reading name table");
seek_tab("name", TTF_USHORT_SIZE);
i = ftell(fontfile);
n = get_ushort();
j = get_ushort() + i - TTF_USHORT_SIZE;
i += 2*TTF_USHORT_SIZE;
while (n-- > 0) {
seek_off("name", i);
platform_id = get_ushort();
encoding_id = get_ushort();
get_ushort(); /* skip language_id */
k = get_ushort();
l = get_ushort();
if ((platform_id == 1 && encoding_id == 0) &&
(k == 0 || k == 4 || k == 6)) {
seek_off("name", j + get_ushort());
for (p = buf; l-- > 0; p++)
*p = get_char();
*p++ = 0;
p = (char *) mymalloc(strlen(buf)+1);
p = strcpy(p,buf);
switch (k) {
case 0: Notice = p; break;
case 4: FullName = p; break;
case 6: FontName = p; break;
}
if (Notice != 0 && FullName != 0 && FontName != 0)
break;
}
i += 6*TTF_USHORT_SIZE;
}
msg(3, "Reading PCLT table");
if ((pd = name_lookup("PCLT")) != 0) {
seek_off("PCLT", pd->offset + TTF_FIXED_L_SIZE + TTF_ULONG_SIZE + TTF_USHORT_SIZE);
XHeight = get_ushort();
ttf_skip(2*TTF_USHORT_SIZE);
CapHeight = get_ushort();
}
msg(3, "Reading kern table");
if ((pd = name_lookup("kern")) == 0)
return;
kern_tab = ttf_alloc(nglyphs, kern_entry);
for (pk = kern_tab; pk - kern_tab < nglyphs; pk++) {
pk->next = 0;
pk->value = 0;
}
seek_off("kern", pd->offset + TTF_USHORT_SIZE);
for (n = get_ushort(); n > 0; n--) {
ttf_skip(2*TTF_USHORT_SIZE);
k = get_ushort();
if (!(k & 1) || (k & 2) || (k & 4))
return;
if (k >> 8 != 0) {
warn(1, "warning: only format 0 supported of `kern' \
subtables, others are ignored\n");
continue;
}
k = get_ushort();
ttf_skip(3*TTF_USHORT_SIZE);
while (k-- > 0) {
i = get_ushort();
j = get_ushort();
kern_value = get_fword();
if (kern_value != 0) {
store_kern_value(i, j, kern_value);
nkernpairs++;
}
}
}
}