static void do_glyphs(deark *c, lctx *d)
{
struct de_bitmap_font *font = NULL;
de_byte *font_data = NULL;
de_int64 i;
de_int64 glyph_rowspan;
font = de_create_bitmap_font(c);
font->has_nonunicode_codepoints = 1;
font->nominal_width = (int)d->glyph_width;
font->nominal_height = (int)d->glyph_height;
font->num_chars = d->num_glyphs; // This may increase later
glyph_rowspan = (d->glyph_width+7)/8;
d->num_chars_alloc = d->num_glyphs;
if(d->read_extra_codepoints)
d->num_chars_alloc += MAX_EXTRA_CODEPOINTS;
d->index_of_first_extra_codepoint = d->num_glyphs;
d->num_extra_codepoints = 0;
font->char_array = de_malloc(c, d->num_chars_alloc * sizeof(struct de_bitmap_font_char));
font_data = de_malloc(c, d->font_data_size);
de_read(font_data, d->headersize, d->font_data_size);
for(i=0; i<d->num_chars_alloc; i++) {
font->char_array[i].width = font->nominal_width;
font->char_array[i].height = font->nominal_height;
font->char_array[i].rowspan = glyph_rowspan;
if(i<d->num_glyphs)
font->char_array[i].codepoint_nonunicode = (de_int32)i;
else
font->char_array[i].codepoint_nonunicode = DE_INVALID_CODEPOINT;
font->char_array[i].codepoint_unicode = DE_INVALID_CODEPOINT;
if(i<d->num_glyphs)
font->char_array[i].bitmap = &font_data[i*d->bytes_per_glyph];
}
if(d->has_unicode_table) {
if(d->version==2)
do_psf2_unicode_table(c, d, font);
else
do_psf1_unicode_table(c, d, font);
}
if(d->num_extra_codepoints>0) {
font->num_chars = d->index_of_first_extra_codepoint + d->num_extra_codepoints;
de_dbg(c, "codepoints aliases: %d\n", (int)d->num_extra_codepoints);
de_dbg(c, "total characters: %d\n", (int)font->num_chars);
}
de_font_bitmap_font_to_image(c, font, NULL, 0);
if(font) {
de_free(c, font->char_array);
de_destroy_bitmap_font(c, font);
}
de_free(c, font_data);
}
static int do_decode_rle(deark *c, lctx *d, i64 pos1, dbuf *unc_pixels)
{
u8 b;
i64 count;
i64 k;
u8 buf[8];
i64 pos = pos1;
while(1) {
if(pos >= c->infile->len) break;
if(unc_pixels->len >= d->main_image.img_size_in_bytes) break;
b = de_getbyte(pos);
pos++;
if(b & 0x80) { // RLE block
count = (i64)(b - 0x80) + 1;
de_read(buf, pos, d->bytes_per_pixel);
pos += d->bytes_per_pixel;
for(k=0; k<count; k++) {
dbuf_write(unc_pixels, buf, d->bytes_per_pixel);
}
}
else { // uncompressed block
count = (i64)(b) + 1;
dbuf_copy(c->infile, pos, count * d->bytes_per_pixel, unc_pixels);
pos += count * d->bytes_per_pixel;
}
}
de_dbg(c, "decompressed %d bytes to %d bytes", (int)(pos-pos1), (int)unc_pixels->len);
return 1;
}
static void do_extract_png_or_jp2(deark *c, lctx *d, struct page_ctx *pg)
{
de_byte buf[8];
de_finfo *fi = NULL;
de_dbg(c, "Trying to extract file at %d\n", (int)pg->image_pos);
// Detect the format
de_read(buf, pg->image_pos, sizeof(buf));
fi = de_finfo_create(c);
de_finfo_set_name_from_sz(c, fi, pg->filename_token, DE_ENCODING_ASCII);
if(buf[4]=='j' && buf[5]=='P') {
dbuf_create_file_from_slice(c->infile, pg->image_pos, pg->image_len, "jp2", fi, 0);
}
else if(buf[0]==0x89 && buf[1]==0x50) {
dbuf_create_file_from_slice(c->infile, pg->image_pos, pg->image_len, "png", fi, 0);
}
else {
de_err(c, "(Image #%d) Unidentified file format\n", pg->image_num);
}
de_finfo_destroy(c, fi);
}
static int do_identify_png_internal(deark *c)
{
u8 buf[8];
de_read(buf, 0, sizeof(buf));
if(!de_memcmp(buf, "\x89\x50\x4e\x47\x0d\x0a\x1a\x0a", 8)) return DE_PNGFMT_PNG;
if(!de_memcmp(buf, "\x8b\x4a\x4e\x47\x0d\x0a\x1a\x0a", 8)) return DE_PNGFMT_JNG;
if(!de_memcmp(buf, "\x8a\x4d\x4e\x47\x0d\x0a\x1a\x0a", 8)) return DE_PNGFMT_MNG;
return 0;
}
static int de_identify_cardfile(deark *c)
{
de_byte buf[4];
de_read(buf, 0, 4);
if(!de_memcmp(buf, "MGC", 3)) return 80;
if(!de_memcmp(buf, "RRG", 3)) return 80;
return 0;
}
static int de_identify_psionapp(deark *c)
{
de_byte b[16];
de_read(b, 0, 16);
if(!de_memcmp(b, "ImageFileType**\0", 16))
return 100;
if(!de_memcmp(b, "OPLObjectFile**\0", 16))
return 100;
return 0;
}
static int de_identify_gzip(deark *c)
{
u8 buf[3];
de_read(buf, 0, 3);
if(buf[0]==0x1f && buf[1]==0x8b) {
if(buf[2]==0x08) return 100;
return 10;
}
return 0;
}
static int identify_fmt(deark *c, de_int64 pos)
{
de_byte buf[3];
de_read(buf, pos, 3);
if(buf[0]!='P') return 0;
if(buf[1]=='7' && buf[2]==0x0a)
return FMT_PAM;
if(buf[1]>='1' && buf[1]<='6')
return buf[1] - '0';
return 0;
}
static void read_compression_type(deark *c, lctx *d, i64 pos)
{
u8 buf[16];
de_dbg(c, "compression type at %d", (int)pos);
de_read(buf, pos, sizeof(buf));
if(!de_memcmp(buf, "lzma\0", 5)) {
d->cmpr_type = DE_RPM_CMPR_LZMA;
}
// Other valid compression types are "gzip", "bzip2", and "xz".
// We'll autodetect most of them, but lzma is hard to detect.
}
static int de_identify_tga(deark *c)
{
u8 b[18];
u8 x;
int has_tga_ext;
if(has_signature(c)) {
return 100;
}
// TGA v1 format has no signature, but there are only a few common types of
// it. We'll at least try to identify anything that we support.
de_read(b, 0, 18);
if(b[1]>1) return 0; // Color map type should be 0 or 1.
// bits/pixel:
if(b[16]!=1 && b[16]!=8 && b[16]!=15 && b[16]!=16 && b[16]!=24 && b[16]!=32) return 0;
if(b[2]!=0 && b[2]!=1 && b[2]!=2 && b[2]!=3 &&
b[2]!=9 && b[2]!=10 && b[2]!=11 && b[2]!=32 && b[2]!=33)
{
return 0; // Unknown image type
}
if(b[12]==0 && b[13]==0) return 0; // Width can't be 0.
if(b[14]==0 && b[15]==0) return 0; // Height can't be 0.
// Bits per palette entry. Supposed to be 0 if there is no palette, but
// in practice it may be 24 instead.
if((b[1]==0 && b[7]==0) || b[7]==15 || b[7]==16 || b[7]==24 || b[7]==32) {
;
}
else {
return 0;
}
has_tga_ext = de_input_file_has_ext(c, "tga");
x = b[17]&0x0f; // Number of attribute bits
if(x!=0 && x!=1 && x!=8 && !has_tga_ext) return 0;
if(has_tga_ext) {
return 100;
}
if(de_input_file_has_ext(c, "vst")) {
return 40;
}
return 8;
}
static int de_identify_wri(deark *c)
{
u8 buf[6];
de_read(buf, 0, 6);
if((buf[0]==0x31 || buf[0]==0x32) &&
!de_memcmp(&buf[1], "\xbe\x00\x00\x00\xab", 5))
{
i64 pnMac;
pnMac = de_getu16le(48*2);
if(pnMac==0) return 0; // Apparently MSWord, not Write
return 100;
}
return 0;
}
static int de_identify_macpaint(deark *c)
{
de_byte buf[8];
de_read(buf, 65, 8);
// Not all MacPaint files can be easily identified, but this will work
// for some of them.
if(!de_memcmp(buf, "PNTGMPNT", 8)) return 80;
if(!de_memcmp(buf, "PNTG", 4)) return 70;
if(de_input_file_has_ext(c, "mac")) return 10;
if(de_input_file_has_ext(c, "macp")) return 15;
if(de_input_file_has_ext(c, "pntg")) return 15;
return 0;
}
static void wri_convert_image_pal4planar(deark *c, i64 fpos,
i64 bytes_per_row_per_plane, de_bitmap *img)
{
const i64 nplanes = 4;
i64 i, j, plane;
i64 rowspan;
u8 *rowbuf = NULL;
static const u32 pal16[16] = {
0x000000,0x800000,0x008000,0x808000,0x000080,0x800080,0x008080,0x808080,
0xc0c0c0,0xff0000,0x00ff00,0xffff00,0x0000ff,0xff00ff,0x00ffff,0xffffff
};
rowspan = bytes_per_row_per_plane * nplanes;
rowbuf = de_malloc(c, rowspan);
// The usual order seems to be
// row0_plane0x1, row0_plane0x2, row0_plane0x4, row0_plane0x8,
// row1_plane0x1, row1_plane0x2, row1_plane0x4, row1_plane0x8,
// ...
// But I have seen another, and I see no way to detect/support it.
for(j=0; j<img->height; j++) {
de_read(rowbuf, fpos+j*rowspan, rowspan);
for(i=0; i<img->width; i++) {
unsigned int palent = 0;
u32 clr;
for(plane=0; plane<nplanes; plane++) {
unsigned int n = 0;
i64 idx;
idx = bytes_per_row_per_plane*plane + i/8;
if(idx<rowspan) n = rowbuf[idx];
if(n & (1<<(7-i%8))) {
palent |= (1<<plane);
}
}
clr = DE_MAKE_OPAQUE(pal16[palent]);
de_bitmap_setpixel_rgb(img, i, j, clr);
}
}
de_free(c, rowbuf);
}
static void handle_embedded_file(deark *c, lctx *d, de_int64 offset, de_int64 len)
{
de_byte buf[16];
const char *ext;
int extract_this_file;
int is_pic;
de_dbg(c, "embedded file at %d, len=%d\n", (int)offset, (int)len);
is_pic = 0;
ext = "bin";
extract_this_file = 0;
if(len>0 && c->extract_level>=2)
extract_this_file = 1;
// As far as I can tell, there's no way to tell the type of an
// embedded file, except by sniffing it.
de_read(buf, offset, 16);
if(len>=8) {
if(!de_memcmp(buf, "PIC\xdc\x30\x30", 6)) {
// Looks like a PIC file
is_pic = 1;
ext = "pic";
extract_this_file = 1;
}
}
if(extract_this_file) {
if(is_pic && d->convert_images) {
// Convert PIC to PNG.
// For consistency, this option shouldn't exist. But I'm not sure that
// PIC files embedded in APP files are really the same as PIC files on
// their own. They might need special handling. Until I'm sure they don't,
// I'll leave this option here.
de_run_module_by_id_on_slice(c, "psionpic", NULL, c->infile, offset, len);
}
else {
// Just extract the file
dbuf_create_file_from_slice(c->infile, offset, len, ext, NULL, is_pic?0:DE_CREATEFLAG_IS_AUX);
}
}
else {
de_dbg(c, "(not extracting this file)\n");
}
}
static void do_identify_bitstream(deark *c, lctx *d, struct stream_info *si, i64 pos, i64 len)
{
u8 idbuf[16];
size_t bytes_to_scan;
size_t k;
bytes_to_scan = (size_t)len;
if(bytes_to_scan > sizeof(idbuf)) {
bytes_to_scan = sizeof(idbuf);
}
de_read(idbuf, pos, bytes_to_scan);
for(k=0; k<DE_ITEMS_IN_ARRAY(stream_type_info_arr); k++) {
if(!de_memcmp(idbuf, stream_type_info_arr[k].magic,
stream_type_info_arr[k].magic_len))
{
si->sti = &stream_type_info_arr[k];
si->stream_type = si->sti->stream_type;
break;
}
}
if(si->stream_type==STREAMTYPE_VORBIS) {
d->found_vorbis = 1;
}
else if(si->stream_type==STREAMTYPE_THEORA) {
d->found_theora = 1;
}
else if(si->stream_type==STREAMTYPE_SKELETON) {
d->found_skeleton = 1;
}
else if(si->sti && (si->sti->flags&0x1)) {
d->found_ogm = 1;
}
if(si->stream_type!=STREAMTYPE_VORBIS && si->stream_type!=STREAMTYPE_THEORA) {
d->has_non_vorbis_non_theora_stream = 1;
}
de_dbg(c, "bitstream type: %s", si->sti?si->sti->name:"unknown");
}
static int read_pam_header_line(deark *c, lctx *d, struct page_ctx *pg, de_int64 pos,
de_int64 *content_len, de_int64 *total_len,
char *linebuf, size_t linebuf_len)
{
int ret;
de_int64 amt_to_read;
linebuf[0]='\0';
ret = dbuf_find_line(c->infile, pos,
content_len, total_len);
if(!ret) return 0;
amt_to_read = *content_len;
if(amt_to_read > (de_int64)(linebuf_len-1)) amt_to_read = (de_int64)(linebuf_len-1);
de_read((de_byte*)linebuf, pos, amt_to_read);
*content_len = amt_to_read;
linebuf[amt_to_read] = '\0';
return 1;
}
static void de_run_rpm(deark *c, de_module_params *mparams)
{
lctx *d = NULL;
i64 pos;
u8 buf[8];
const char *ext;
i64 section_size = 0;
de_finfo *fi = NULL;
char filename[128];
d = de_malloc(c, sizeof(lctx));
if(!do_lead_section(c, d)) {
goto done;
}
pos = 96;
if(!do_header_structure(c, d, 1, pos, §ion_size)) {
goto done;
}
pos += section_size;
// Header structures are 8-byte aligned. The first one always starts at
// offset 96, so we don't have to worry about it. But we need to make
// sure the second one is aligned.
pos = ((pos + 7)/8)*8;
if(!do_header_structure(c, d, 0, pos, §ion_size)) {
goto done;
}
pos += section_size;
de_dbg(c, "data pos: %d", (int)pos);
if(pos > c->infile->len) goto done;
// There is usually a tag that indicates the compression format, but we
// primarily figure out the format by sniffing its magic number, on the
// theory that that's more reliable.
// TODO: I think it's also theoretically possible that it could use an archive
// format other than cpio.
de_read(buf, pos, 8);
if(buf[0]==0x1f && buf[1]==0x8b) {
ext = "cpio.gz";
}
else if(buf[0]==0x42 && buf[1]==0x5a && buf[2]==0x68) {
ext = "cpio.bz2";
}
else if(buf[0]==0xfd && buf[1]==0x37 && buf[2]==0x7a) {
ext = "cpio.xz";
}
else if(d->cmpr_type==DE_RPM_CMPR_LZMA || buf[0]==0x5d) {
ext = "cpio.lzma";
}
else {
de_warn(c, "Unidentified compression or archive format");
ext = "cpio.bin";
}
if(d->name_srd && c->filenames_from_file) {
const char *version2 = "x";
const char *release2 = "x";
if(d->version_srd) version2 = d->version_srd->sz;
if(d->release_srd) release2 = d->release_srd->sz;
fi = de_finfo_create(c);
de_snprintf(filename, sizeof(filename), "%s-%s.%s",
d->name_srd->sz, version2, release2);
de_finfo_set_name_from_sz(c, fi, filename, 0, DE_ENCODING_ASCII);
}
dbuf_create_file_from_slice(c->infile, pos, c->infile->len - pos, ext, fi, 0);
done:
de_finfo_destroy(c, fi);
if(d) {
de_destroy_stringreaderdata(c, d->name_srd);
de_destroy_stringreaderdata(c, d->release_srd);
de_destroy_stringreaderdata(c, d->version_srd);
de_free(c, d);
}
}
// Note that a header *structure* is distinct from the header *section*.
// Both the signature section and the header section use a header structure.
static int do_header_structure(deark *c, lctx *d, int is_sig, i64 pos1,
i64 *section_size)
{
i64 pos;
i64 indexcount;
i64 storesize;
u8 buf[4];
u8 header_ver;
i64 i;
i64 tag_id, tag_type, tag_offset, tag_count;
i64 data_store_pos;
const char *hdrname;
int retval = 0;
hdrname = is_sig?"sig":"hdr";
pos = pos1;
de_dbg(c, "%s section at %d", hdrname, (int)pos1);
de_dbg_indent(c, 1);
de_read(buf, pos, 4);
if(buf[0]!=0x8e || buf[1]!=0xad || buf[2]!=0xe8) {
de_err(c, "Bad header signature at %d", (int)pos);
goto done;
}
header_ver = buf[3];
if(header_ver != 1) {
de_err(c, "Unsupported header version");
goto done;
}
pos += 8;
indexcount = de_getu32be(pos);
storesize = de_getu32be(pos+4);
de_dbg(c, "%s: pos=%d indexcount=%d storesize=%d", hdrname,
(int)pos, (int)indexcount, (int)storesize);
pos += 8;
if(indexcount>1000) goto done;
data_store_pos = pos + 16*indexcount;
de_dbg(c, "%s: tag table at %d", hdrname, (int)pos);
de_dbg_indent(c, 1);
for(i=0; i<indexcount; i++) {
tag_id = de_getu32be(pos);
tag_type = de_getu32be(pos+4);
tag_offset = de_getu32be(pos+8);
tag_count = de_getu32be(pos+12);
de_dbg2(c, "tag #%d type=%d offset=%d count=%d", (int)tag_id,
(int)tag_type, (int)tag_offset, (int)tag_count);
if(is_sig==0 && tag_id==DE_RPMTAG_PAYLOADCOMPRESSOR && tag_type==DE_RPM_STRING_TYPE) {
read_compression_type(c, d, data_store_pos+tag_offset);
}
else if(is_sig==0 && tag_id==DE_RPMTAG_NAME && tag_type==DE_RPM_STRING_TYPE) {
if(!d->name_srd) {
d->name_srd = dbuf_read_string(c->infile, data_store_pos+tag_offset,
DE_DBG_MAX_STRLEN, DE_DBG_MAX_STRLEN,
DE_CONVFLAG_STOP_AT_NUL, DE_ENCODING_ASCII);
de_dbg(c, "name: \"%s\"", ucstring_getpsz(d->name_srd->str));
}
}
else if(is_sig==0 && tag_id==DE_RPMTAG_VERSION && tag_type==DE_RPM_STRING_TYPE) {
if(!d->version_srd) {
d->version_srd = dbuf_read_string(c->infile, data_store_pos+tag_offset,
DE_DBG_MAX_STRLEN, DE_DBG_MAX_STRLEN,
DE_CONVFLAG_STOP_AT_NUL, DE_ENCODING_ASCII);
de_dbg(c, "version: \"%s\"", ucstring_getpsz(d->version_srd->str));
}
}
else if(is_sig==0 && tag_id==DE_RPMTAG_RELEASE && tag_type==DE_RPM_STRING_TYPE) {
if(!d->release_srd) {
d->release_srd = dbuf_read_string(c->infile, data_store_pos+tag_offset,
DE_DBG_MAX_STRLEN, DE_DBG_MAX_STRLEN,
DE_CONVFLAG_STOP_AT_NUL, DE_ENCODING_ASCII);
de_dbg(c, "release: \"%s\"", ucstring_getpsz(d->release_srd->str));
}
}
pos += 16;
}
de_dbg_indent(c, -1);
pos = data_store_pos;
de_dbg(c, "%s: data store at %d", hdrname, (int)pos);
pos += storesize;
*section_size = pos - pos1;
retval = 1;
done:
de_dbg_indent(c, -1);
return retval;
}
// pos1 points to the ole_id field (should be 0x00000501).
// Caller must have looked ahead to check the type.
static int do_picture_ole_embedded_rendition(deark *c, lctx *d, struct para_info *pinfo,
int rendition_idx, i64 pos1, i64 *bytes_consumed)
{
i64 pos = pos1;
i64 stringlen;
i64 data_len;
u8 buf[16];
struct de_stringreaderdata *srd_typename = NULL;
struct de_stringreaderdata *srd_filename = NULL;
struct de_stringreaderdata *srd_params = NULL;
pos += 4; // 0x00000501
pos += 4; // "type" (probably already read by caller)
stringlen = de_getu32le_p(&pos);
srd_typename = dbuf_read_string(c->infile, pos, stringlen, 260, DE_CONVFLAG_STOP_AT_NUL,
DE_ENCODING_ASCII);
de_dbg(c, "typename: \"%s\"", ucstring_getpsz(srd_typename->str));
pos += stringlen;
stringlen = de_getu32le_p(&pos);
srd_filename = dbuf_read_string(c->infile, pos, stringlen, 260, DE_CONVFLAG_STOP_AT_NUL,
DE_ENCODING_ASCII);
de_dbg(c, "filename: \"%s\"", ucstring_getpsz(srd_filename->str));
pos += stringlen;
stringlen = de_getu32le_p(&pos);
srd_params = dbuf_read_string(c->infile, pos, stringlen, 260, DE_CONVFLAG_STOP_AT_NUL,
DE_ENCODING_ASCII);
de_dbg(c, "params: \"%s\"", ucstring_getpsz(srd_params->str));
pos += stringlen;
data_len = de_getu32le_p(&pos);
de_dbg(c, "embedded ole rendition data: pos=%d, len=%d", (int)pos, (int)data_len);
// TODO: I don't know the extent to which it's better to sniff the data, or
// rely on the typename.
de_read(buf, pos, sizeof(buf));
if(!de_strcmp(srd_typename->sz, "CDraw") &&
!de_memcmp(&buf[0], (const void*)"RIFF", 4) &&
!de_memcmp(&buf[8], (const void*)"CDR", 3) )
{
// Looks like CorelDRAW
dbuf_create_file_from_slice(c->infile, pos, data_len, "cdr", NULL, 0);
}
else if(buf[0]=='B' && buf[1]=='M') {
// TODO: Detect true length of data
dbuf_create_file_from_slice(c->infile, pos, data_len, "bmp", NULL, 0);
}
else {
if(d->extract_ole) {
extract_unknown_ole_obj(c, d, pos, data_len, srd_typename);
}
else {
de_warn(c, "Unknown/unsupported type of OLE object (\"%s\") at %d",
ucstring_getpsz(srd_typename->str), (int)pos1);
}
}
pos += data_len;
*bytes_consumed = pos - pos1;
de_destroy_stringreaderdata(c, srd_typename);
de_destroy_stringreaderdata(c, srd_filename);
de_destroy_stringreaderdata(c, srd_params);
return 1;
}
// Caller must initialize *repeat_count.
static void uncompress_line(deark *c, lctx *d, dbuf *unc_line,
i64 pos1, i64 rownum,
i64 *bytes_consumed, i64 *repeat_count)
{
i64 pos;
u8 b0, b1;
u8 val;
i64 count;
i64 k;
i64 tmp_repeat_count;
i64 unc_line_len_orig;
*bytes_consumed = 0;
pos = pos1;
unc_line_len_orig = unc_line->len;
while(1) {
if(pos >= c->infile->len) break;
if(unc_line->len - unc_line_len_orig >= d->rowspan_per_plane) break;
b0 = de_getbyte(pos++);
if(b0==0) { // Pattern run or scanline run
b1 = de_getbyte(pos++);
if(b1>0) { // pattern run
de_read(d->pattern_buf, pos, d->patlen);
pos += d->patlen;
count = (i64)b1;
for(k=0; k<count; k++) {
dbuf_write(unc_line, d->pattern_buf, d->patlen);
}
}
else { // (b1==0) scanline run
u8 flagbyte;
flagbyte = de_getbyte(pos);
if(flagbyte==0xff) {
pos++;
tmp_repeat_count = (i64)de_getbyte(pos++);
if(tmp_repeat_count == 0) {
de_dbg(c, "row %d: bad repeat count", (int)rownum);
}
else {
*repeat_count = tmp_repeat_count;
}
}
else {
de_dbg(c, "row %d: bad scanline run marker: 0x%02x",
(int)rownum, (unsigned int)flagbyte);
}
}
}
else if(b0==0x80) { // "Uncompressed bit string"
count = (i64)de_getbyte(pos++);
dbuf_copy(c->infile, pos, count, unc_line);
pos += count;
}
else { // "solid run"
val = (b0&0x80) ? 0xff : 0x00;
count = (i64)(b0 & 0x7f);
dbuf_write_run(unc_line, val, count);
}
}
*bytes_consumed = pos - pos1;
}
static void do_psf2_unicode_table(deark *c, lctx *d, struct de_bitmap_font *font)
{
de_int64 cur_idx;
de_int64 pos;
int ret;
de_int64 foundpos;
de_int64 char_data_len;
de_byte char_data_buf[200];
de_int32 ch;
de_int64 utf8len;
de_dbg(c, "Unicode table at %d\n", (int)d->unicode_table_pos);
de_dbg_indent(c, 1);
pos = d->unicode_table_pos;
cur_idx = 0;
while(1) {
de_int64 pos_in_char_data;
de_int64 cp_idx;
if(cur_idx >= d->num_glyphs) break;
if(pos >= c->infile->len) break;
// Figure out the size of the data for this glyph
ret = dbuf_search_byte(c->infile, 0xff, pos,
c->infile->len - pos, &foundpos);
if(!ret) break;
char_data_len = foundpos - pos;
if(char_data_len<0) char_data_len=0;
else if(char_data_len>(de_int64)sizeof(char_data_buf)) char_data_len=(de_int64)sizeof(char_data_buf);
// Read all the data for this glyph
de_read(char_data_buf, pos, char_data_len);
// Read the codepoints for this glyph
cp_idx = 0;
pos_in_char_data = 0;
while(1) {
if(pos_in_char_data >= char_data_len) break;
ret = de_utf8_to_uchar(&char_data_buf[pos_in_char_data], char_data_len-pos_in_char_data,
&ch, &utf8len);
if(!ret) {
// If there are any multi-codepoint aliases for this glyph, we
// expect de_utf8_to_uchar() to fail when it hits the 0xfe byte.
// So, this is not necessarily an error.
break;
}
if(cp_idx==0) {
// This is the primary Unicode codepoint for this glyph
de_dbg2(c, "char[%d] = U+%04x\n", (int)cur_idx, (unsigned int)ch);
font->char_array[cur_idx].codepoint_unicode = ch;
}
else {
do_extra_codepoint(c, d, font, cur_idx, ch);
}
cp_idx++;
pos_in_char_data += utf8len;
}
if(cp_idx==0) {
de_warn(c, "Missing codepoint for char #%d\n", (int)cur_idx);
}
// Advance to the next glyph
pos = foundpos+1;
cur_idx++;
}
font->has_unicode_codepoints = 1;
font->prefer_unicode = 1;
de_dbg_indent(c, -1);
}