summaryrefslogtreecommitdiff
path: root/Build/source/texk/xdvik/tfmload.c
diff options
context:
space:
mode:
Diffstat (limited to 'Build/source/texk/xdvik/tfmload.c')
-rw-r--r--Build/source/texk/xdvik/tfmload.c120
1 files changed, 120 insertions, 0 deletions
diff --git a/Build/source/texk/xdvik/tfmload.c b/Build/source/texk/xdvik/tfmload.c
new file mode 100644
index 00000000000..4b904b2576e
--- /dev/null
+++ b/Build/source/texk/xdvik/tfmload.c
@@ -0,0 +1,120 @@
+/* This code was stolen from dvips for xdvi use. The file itself did
+ * not carry any licensing info. According to other filed the code
+ * either public domain, or, in the worst case, GPL.
+ *
+ * Loads a tfm file. It marks the characters as undefined.
+ *
+ * Modified for use by xdvi/t1 by Nicolai Langfeldt */
+
+#include <stdio.h>
+
+#include "xdvi-config.h"
+#include "xdvi.h"
+#include "util.h"
+#include "tfmload.h"
+
+#ifdef T1LIB
+
+#include "kpathsea/tex-file.h"
+
+int fallbacktfm = 0;
+
+Boolean
+tfmload(const char *name, long *design, long *widths, long *fontdimen2)
+{
+ int i;
+ int li;
+ int nw, hd;
+ int bc, ec;
+ int f_param_off, f_param_num;
+ long wtab[256];
+ unsigned short chardat[256];
+ char *filename;
+ FILE *curfnt;
+
+ fallbacktfm = 0;
+
+ filename = kpse_find_tfm(name);
+
+ if (filename == NULL) {
+ filename = kpse_find_tfm("cmr10.tfm");
+ if (fallbacktfm == 0)
+ fallbacktfm = 1;
+ }
+
+ if (filename == NULL)
+ return False;
+
+ curfnt = XFOPEN(filename, "r");
+
+ if (curfnt == NULL)
+ return False;
+
+ /* Next, we read the font data from the tfm file, and store it in
+ * our own array. */
+
+ /* Lengths section, 16 bit unsigned quanities */
+ li = get_bytes(curfnt, 2); /* lf = length of file, in words(4 bytes) */
+ hd = get_bytes(curfnt, 2); /* lh = length of header data, in words */
+ bc = get_bytes(curfnt, 2); /* bc = lowest charcode */
+ ec = get_bytes(curfnt, 2); /* ec = largest charcode */
+ nw = get_bytes(curfnt, 2); /* nw = number of width words */
+ f_param_off = get_bytes(curfnt, 2); /* nh = number of height words */
+ f_param_off += get_bytes(curfnt, 2); /* nd = number of depth words */
+ f_param_off += get_bytes(curfnt, 2); /* ni = number of italic words */
+ f_param_off += get_bytes(curfnt, 2); /* nl = number of words in lig/kern table */
+ f_param_off += get_bytes(curfnt, 2); /* nk = number of words in the kern table */
+ f_param_off += get_bytes(curfnt, 2); /* ne = number of words in the extensible char tab */
+ f_param_num = get_bytes(curfnt, 2); /* np = number of font parameter words */
+
+ /* Header, 32 bit signed quantities */
+ get_bytes(curfnt, 4); /* header[0]: checksum */
+ *design = get_bytes(curfnt, 4); /* header[1]: design size */
+
+ /* Skip the rest: header[2..(hd-1)] */
+ for (i = 2; i < hd; i++)
+ li = get_bytes(curfnt, 4);
+
+ /* Initialize to undef value */
+ for (i = 0; i < 256; i++) {
+ chardat[i] = 256;
+ widths[i] = 0;
+ }
+
+ /* The charinfo array: */
+ for (i = bc; i <= ec; i++) {
+ chardat[i] = get_byte(curfnt); /* 8 bits of width index */
+ get_byte(curfnt); /* skip */
+ get_bytes(curfnt, 2); /* skip, jump */
+ }
+
+ /* The (compressed!) width table */
+ for (i = 0; i < nw; i++)
+ wtab[i] = get_bytes(curfnt, 4);
+
+ /* skip to font parameter words */
+ for (i = 0; i < f_param_off; i++)
+ get_bytes(curfnt, 4);
+
+ /* read fontdimen2 (width of space) */
+ for (i = 0; i < f_param_num; i++) {
+ long l = get_bytes(curfnt, 4);
+ if (i == 1) {
+ *fontdimen2 = l;
+ if (globals.debug & DBG_T1)
+ fprintf(stderr, "fontdimen2 of %s: %ld\n", filename, *fontdimen2);
+ }
+
+ }
+ (void)fclose(curfnt);
+
+ /* Now we have the table and the indexes. Expand into a directly
+ mapped width array */
+ for (i = bc; i <= ec; i++)
+ if (chardat[i] != 256)
+ widths[i] = wtab[chardat[i]];
+
+ return True;
+}
+
+#endif /* T1LIB */