rewrite the API, support more object types and actual evaluation

Sigrid Haflínudóttir committed 6 years ago

Commit a080ae81 parent 34238e011 files · +494 −376Browse files

Unified · Split

11 files

array.c +21 −24
dict.c +41 −34
eval.c +18 −7
filter.c +1 −0
main.c +9 −1
misc.c +3 −2
name.c +93 −65
object.c +128 −68
pdf.c +5 −9
pdf.h +12 −38
string.c +163 −128

@@ -1,40 +1,34 @@
1	1	#include <u.h>
2	2	#include <libc.h>
3		−#include <ctype.h>
	3	+#include <bio.h>
4	4	#include "pdf.h"
5	5
	6	+/* 7.3.6 Array Objects */
	7	+
6	8	Object *
7		−pdfarray(char p, char *e, int len)
	9	+pdfarray(Biobuf *b)
8	10	{
9	11	Object o, m;
10	12	Object **a;
11		−
12		− o = nil;
13		− if(len < 3 \|\| p[0] != '['){
14		− werrstr("invalid array");
15		− goto err;
16		− }
17		− p++;
18		− len--;
	13	+ int c, noel;
19	14
20	15	o = calloc(1, sizeof(*o));
21	16	o->type = Oarray;
22		− for(; len >= 1;){
23		− for(; len > 0 && isws(*p); p++, len--);
24		− if(len < 1){
25		− werrstr("too short");
26		− goto err;
27		− }
	17	+ Bgetc(b); /* throw away '[' */
28	18
29		− if(p[0] == ']'){
30		− p++;
	19	+ for(noel = 0;;){
	20	+ if((c = Bgetc(b)) < 0 \|\| c == ']')
31	21	break;
	22	+ if(noel){
	23	+ werrstr("no ']'");
	24	+ goto err;
32	25	}
33	26
34		− if((m = pdfobject(p, e, len)) == nil)
35		− goto err;
36		− len -= *e - p;
37		− p = *e;
	27	+ Bungetc(b);
	28	+ if((m = pdfobject(b)) == nil){
	29	+ noel = 1;
	30	+ continue;
	31	+ }
38	32
39	33	if((a = realloc(o->array.e, (o->array.ne+1)sizeof(Object))) == nil){
40	34	freeobject(m);
@@ -45,11 +39,14 @@ pdfarray(char p, char *e, int len) · 4 unmodified lines
45	39	a[o->array.ne++] = m;
46	40	}
47	41
48		− *e = p;
	42	+ if(c != ']'){
	43	+ werrstr("no ']'");
	44	+ goto err;
	45	+ }
49	46
50	47	return o;
51	48	err:
	49	+ werrstr("array: %r");
52	50	freeobject(o);
53		−
54	51	return nil;
55	52	}

dict.c+41−34

@@ -1,70 +1,77 @@
1	1	#include <u.h>
2	2	#include <libc.h>
3		−#include <ctype.h>
	3	+#include <bio.h>
4	4	#include "pdf.h"
5	5
	6	+/* 7.3.7 Dictionary Objects */
	7	+
6	8	Object *
7		−pdfdict(char p, char *e, int len)
	9	+pdfdict(Biobuf *b)
8	10	{
9	11	Object o, k, *v;
10	12	KeyValue *kv;
	13	+ int c, nokey;
11	14
12		− o = nil;
13		− if(len < 4 \|\| p[0] != '<' \|\| p[1] != '<'){
14		− werrstr("invalid dict");
15		− goto err;
16		− }
17		− p += 2;
18		− len -= 2;
	15	+ /* skip '<<' */
	16	+ Bseek(b, 2, 1);
19	17
	18	+ k = v = nil;
20	19	o = calloc(1, sizeof(*o));
21	20	o->type = Odict;
22		− for(; len >= 2;){
23		− for(; len > 2 && isws(*p); p++, len--);
24		− if(len < 1){
25		− werrstr("too short");
	21	+ for(nokey = 0;;){
	22	+ if((c = Bgetc(b)) < 0)
	23	+ goto err;
	24	+ if(c == '>'){
	25	+ if(Bgetc(b) == '>')
	26	+ break;
	27	+ werrstr("no '>>'");
26	28	goto err;
27	29	}
28		−
29		− if(p[0] == '>' && p[1] == '>'){
30		− p += 2;
31		− break;
	30	+ if(nokey){
	31	+ werrstr("no '>>'");
	32	+ goto err;
32	33	}
33	34
34		− if((k = pdfobject(p, e, len)) == nil)
	35	+ Bungetc(b);
	36	+ if((k = pdfobject(b)) == nil){
	37	+ nokey = 1;
	38	+ continue;
	39	+ }
	40	+ if((v = pdfobject(b)) == nil)
35	41	goto err;
36	42	if(k->type != Oname){
37		− freeobject(k);
38	43	werrstr("expected name as a key");
39	44	goto err;
40	45	}
41		− len -= *e - p;
42		− p = *e;
43	46
44		− if((v = pdfobject(p, e, len)) == nil){
45		− freeobject(k);
	47	+ if((kv = realloc(o->dict.kv, (o->dict.nkv+1)*sizeof(KeyValue))) == nil)
46	48	goto err;
47		− }
48		− len -= *e - p;
49		− p = *e;
50		−
51		− if((kv = realloc(o->dict.kv, (o->dict.nkv+1)*sizeof(KeyValue))) == nil){
52		− freeobject(k);
53		− freeobject(v);
54		− goto err;
55		− }
56	49
57	50	o->dict.kv = kv;
58	51	kv[o->dict.nkv].key = strdup(k->name);
59	52	freeobject(k);
60	53	kv[o->dict.nkv++].value = v;
	54	+ k = v = nil;
61	55	}
62	56
63		− *e = p;
64		−
65	57	return o;
66	58	err:
67	59	freeobject(o);
	60	+ freeobject(k);
	61	+ freeobject(v);
	62	+ werrstr("dict: %r");
68	63
69	64	return nil;
70	65	}
	66	+
	67	+Object *
	68	+pdfdictget(Object o, char name)
	69	+{
	70	+ int i;
	71	+
	72	+ if(o == nil \|\| o->type != Odict \|\| name == nil)
	73	+ return nil;
	74	+ for(i = 0; i < o->dict.nkv && strcmp(name, o->dict.kv[i].key) != 0; i++);
	75	+
	76	+ return i < o->dict.nkv ? o->dict.kv[i].value : nil;
	77	+}

eval.c+18−7

@@ -1,23 +1,34 @@
1	1	#include <u.h>
2	2	#include <libc.h>
	3	+#include <bio.h>
3	4	#include "pdf.h"
4	5
5	6	int
6	7	pdfeval(Pdf pdf, Object o)
7	8	{
	9	+ Object *d;
8	10	Xref *x;
9	11	int i;
10	12
11		− if(o->type != Oindir)
	13	+ if(o == nil \|\| o->type != Oindir)
12	14	return 0;
13	15
14		− for(i = 0; i < pdf->nxref; i++){
15		− x = &pdf->xref[i];
16		− if(x->id == o->indir.id)
17		− return 0;
	16	+ for(i = 0; i < pdf->nxref && pdf->xref[i].id != o->indir.id; i++);
	17	+ if(i >= pdf->nxref){
	18	+ werrstr("no object id %d in xref", o->indir.id);
	19	+ return -1;
18	20	}
	21	+ x = &pdf->xref[i];
19	22
20		− werrstr("no object id %d in xref", o->indir.id);
	23	+ if(Bseek(pdf->bio, x->off, 0) != x->off){
	24	+ werrstr("xref seek failed");
	25	+ return -1;
	26	+ }
	27	+ if((d = pdfobject(pdf->bio)) == nil)
	28	+ return -1;
	29	+ o = d;
	30	+ memset(d, 0, sizeof(*d));
	31	+ freeobject(d);
21	32
22		− return -1;
	33	+ return 0;
23	34	}

filter.c+1−0

@@ -1,5 +1,6 @@
1	1	#include <u.h>
2	2	#include <libc.h>
	3	+#include <bio.h>
3	4	#include "pdf.h"
4	5
5	6	/*

main.c+9−1

@@ -1,8 +1,11 @@
1	1	#include <u.h>
2	2	#include <libc.h>
3	3	#include <thread.h>
	4	+#include <bio.h>
4	5	#include "pdf.h"
5	6
	7	+int mainstacksize = 32768;
	8	+
6	9	static void
7	10	usage(void)
8	11	{
@@ -15,6 +18,7 @@ threadmain(int argc, char **argv) · 6 unmodified lines
15	18	{
16	19	int fd;
17	20	Pdf *pdf;
	21	+ Object *v;
18	22
19	23	quotefmtinstall();
20	24
@@ -40,7 +44,11 @@ threadmain(int argc, char **argv) · 19 unmodified lines
40	44	if((fd = open(argv[0], OREAD)) < 0)
41	45	sysfatal("%r");
42	46	if((pdf = pdfopen(fd)) == nil)
43		− sysfatal("pdfopen: %r");
	47	+ sysfatal("%r");
	48	+ if((v = pdfdictget(pdf->info, "Creator")) != nil)
	49	+ fprint(2, "creator: %s\n", v->str);
	50	+ if((v = pdfdictget(pdf->info, "Producer")) != nil)
	51	+ fprint(2, "producer: %s\n", v->str);
44	52	pdfclose(pdf);
45	53
46	54	threadexitsall(nil);

misc.c+3−2

@@ -1,10 +1,11 @@
1	1	#include <u.h>
2	2	#include <libc.h>
	3	+#include <bio.h>
3	4	#include "pdf.h"
4	5
5	6	/* 7.2.2 whitespace */
6	7	int
7		−isws(char c)
	8	+isws(int c)
8	9	{
9	10	return /* \0 is missing on purpose */
10	11	c == '\t' \|\| c == '\n' \|\| c == '\f' \|\| c == '\r' \|\|
@@ -13,7 +14,7 @@ isws(char c) · 2 unmodified lines
13	14
14	15	/* 7.2.2 delimeters */
15	16	int
16		−isdelim(char c)
	17	+isdelim(int c)
17	18	{
18	19	return
19	20	c == '(' \|\| c == ')' \|\| c == '<' \|\| c == '>' \|\|

name.c+93−65

@@ -1,104 +1,132 @@
1	1	#include <u.h>
2	2	#include <libc.h>
	3	+#include <bio.h>
3	4	#include "pdf.h"
4	5
5		−int
6		−pdfname(char p, char *e, int len)
	6	+/* 7.3.5 Name Objects */
	7	+
	8	+Object *
	9	+pdfname(Biobuf *b)
7	10	{
8		− int i, o;
	11	+ Object *o;
	12	+ char s, r, hex[3];
	13	+ int c, sz, maxsz;
9	14
10		− if(len < 2){
11		− werrstr("too short");
12		− goto err;
13		− }
	15	+ Bgetc(b); /* skip '/' */
14	16
15		− if(p[0] != '/'){
16		− werrstr("invalid first char");
	17	+ maxsz = 32;
	18	+ if((s = malloc(maxsz)) == nil)
17	19	goto err;
18		− }
19	20
20		− for(i = 1, o = 0; i < len; i++){
21		− if(p[i] < '!' \|\| p[i] > '~'){
22		− if(isws(p[i]) \|\| isdelim(p[i]))
	21	+ for(sz = 0;;){
	22	+ if((c = Bgetc(b)) < 0){
	23	+ if(c == Beof)
23	24	break;
24		− werrstr("invalid char %02x", p[i]);
25	25	goto err;
26	26	}
27		− if(p[i] == '#'){
28		− i++;
29		− if(i+2 > len){
30		− werrstr("hex too short");
	27	+
	28	+ if(isws(c) \|\| isdelim(c)){
	29	+ Bungetc(b);
	30	+ break;
	31	+ }
	32	+ if(c < '!' \|\| c > '~'){
	33	+ werrstr("invalid char %02x", c);
	34	+ goto err;
	35	+ }
	36	+ if(c == '#'){
	37	+ if((c = Bgetc(b)) < 0)
31	38	goto err;
32		− }
33		− if(dec16((uchar*)p+o, 1, p+i, 2) != 1){
	39	+ hex[0] = c;
	40	+ if((c = Bgetc(b)) < 0)
	41	+ goto err;
	42	+ hex[1] = c;
	43	+ if(dec16((uchar*)hex, 1, hex, 2) != 1){
34	44	werrstr("invalid hex");
35	45	goto err;
36	46	}
37		− o++;
38		− i++;
39		− }else{
40		− p[o++] = p[i];
	47	+ c = hex[0];
41	48	}
	49	+ if(sz+1 >= maxsz){
	50	+ maxsz *= 2;
	51	+ if((r = realloc(s, maxsz)) == nil)
	52	+ goto err;
	53	+ s = r;
	54	+ }
	55	+ s[sz++] = c;
42	56	}
43	57
44		− p[o] = 0;
45		− *e = p + i;
	58	+ if((o = malloc(sizeof(*o) + sz + 1)) != nil){
	59	+ memmove(o->name, s, sz);
	60	+ o->name[sz] = 0;
	61	+ o->type = Oname;
	62	+ free(s);
	63	+ return o;
	64	+ }
46	65
47		− return o;
48	66	err:
49	67	werrstr("name: %r");
50		− return -1;
	68	+ free(s);
	69	+ return nil;
51	70	}
52	71
53	72	#ifdef TEST
54	73	static struct {
55		− char *i;
56		− int len;
57		− char *o;
58		− int r;
59		− int e;
	74	+ char *in;
	75	+ char *out;
60	76	}t[] = {
61		− {"/SimpleName", 11, "SimpleName", 10},
62		− {"WrongName", 9, nil, -1},
63		− {"/.$()", 5, ".$()", 4},
64		− {"/#30", 4, "0", 1},
65		− {"/#3", 3, nil, -1},
66		− {"/#G0", 4, nil, -1},
67		− {"/#", 2, nil, -1},
68		− {"/Wrong Char", 11, nil, -1},
69		− {"/\xff", 2, nil, -1},
	77	+ {"/SimpleName", "SimpleName"},
	78	+ {"/.$()", ".$"},
	79	+ {"/#30", "0"},
	80	+ {"/#3", nil},
	81	+ {"/#G0", nil},
	82	+ {"/#", nil},
	83	+ {"/Space Between", "Space"},
	84	+ {"/Two/Names", "Two"},
	85	+ {"/\xff", nil,},
70	86	};
71	87
	88	+static char *s;
	89	+static int off, n;
	90	+
	91	+static int
	92	+rd(Biobufhdr , void data, long sz)
	93	+{
	94	+ if(sz > n-off)
	95	+ sz = n-off;
	96	+ memmove(data, s+off, sz);
	97	+ off += sz;
	98	+ return sz;
	99	+}
	100	+
72	101	void
73	102	test_pdfname(void)
74	103	{
75		− char *e;
76		− int i, r;
	104	+ Object *o;
	105	+ Biobuf b;
	106	+ int i;
77	107
78	108	fprint(2, "pdfname\n");
79	109	for(i = 0; i < nelem(t); i++){
	110	+ s = t[i].in;
	111	+ n = strlen(s);
	112	+ off = 0;
	113	+ Binit(&b, -1, OREAD);
	114	+ Biofn(&b, rd);
	115	+
80	116	fprint(2, "\t%d: ", i);
81		− r = pdfname(t[i].i, &e, t[i].len);
82		− if(r != t[i].r){
83		− fprint(2, "expected r=%d, got %d", t[i].r, r);
84		− if(r < 0)
85		− fprint(2, " (%r)\n");
86		− else
87		− fprint(2, "\n");
88		− continue;
89		− }else if(r >= 0){
90		− if(t[i].i+t[i].len != e){
91		− fprint(2, "expected e=%p, got %p\n", t[i].i+t[i].len, e);
92		− continue;
93		− }else if(strcmp(t[i].o, t[i].i) != 0){
94		− fprint(2, "expected %q, got %q\n", t[i].o, t[i].i);
95		− continue;
96		− }
97		− }
98		− fprint(2, "OK");
99		− if(r < 0)
100		− fprint(2, " (%r)");
101		− fprint(2, "\n");
	117	+ o = pdfname(&b);
	118	+ if(o == nil && t[i].out != nil)
	119	+ fprint(2, "ERROR: expected %q, got error: %r\n", t[i].out);
	120	+ else if(o != nil && t[i].out == nil)
	121	+ fprint(2, "ERROR: expected error, got %q\n", o->name);
	122	+ else if(o == nil && t[i].out == nil)
	123	+ fprint(2, "OK (%r)\n");
	124	+ else if(strcmp(o->name, t[i].out) != 0)
	125	+ fprint(2, "ERROR: expected %q, got %q\n", t[i].out, o->name);
	126	+ else
	127	+ fprint(2, "OK\n");
	128	+ freeobject(o);
	129	+ Bterm(&b);
102	130	}
103	131	}
104	132	#endif

object.c+128−68

@@ -1,105 +1,165 @@
1	1	#include <u.h>
2	2	#include <libc.h>
3	3	#include <ctype.h>
	4	+#include <bio.h>
4	5	#include "pdf.h"
5	6
	7	+Object pdfstring(Biobuf b);
	8	+Object pdfname(Biobuf b);
	9	+Object pdfarray(Biobuf b);
	10	+Object pdfdict(Biobuf b);
	11	+
	12	+static Object null = {
	13	+ .type = Onull,
	14	+};
	15	+
	16	+/* General function to parse an object of any type. */
6	17	Object *
7		−pdfobject(char p, char *e, int len)
	18	+pdfobject(void *b)
8	19	{
9		− Object *o;
10		− char s, p0;
11		− int sz, gen;
12		−
13		− o = nil;
14		− for(; len > 0 && isws(*p); p++, len--);
15		− if(len < 1){
16		− werrstr("too short");
17		− goto err;
18		− }
	20	+ Object o, o2;
	21	+ vlong off;
	22	+ int c, tf;
19	23
20		− if(*p < 1){
21		− werrstr("unexpected non-ascii char");
	24	+ o = o2 = nil;
	25	+ do; while(isws(c = Bgetc(b)));
	26	+ if(c < 0)
22	27	goto err;
23		− }
24	28
25		− switch(*p){
	29	+ switch(c){
26	30	case '<': /* dictionary or a string */
27		− if(len < 2){
28		− werrstr("too short");
29		− goto err;
30		− }
31		− if(p[1] == '<'){ /* dictionary */
32		− o = pdfdict(p, e, len);
33		− break;
	31	+ c = Bgetc(b);
	32	+ if(c == '<'){
	33	+ Bseek(b, -2, 1);
	34	+ return pdfdict(b);
34	35	}
	36	+ Bungetc(b);
35	37	/* fall through */
36	38
37		− case '(': /* string */
38		− if((sz = pdfstring(p, e, len)) < 0)
39		− goto err;
40		− if((o = malloc(sizeof(*o)+sz+1)) != nil){
41		− o->type = Ostr;
42		− o->str = (char*)(o+1);
43		− strcpy(o->str, p);
44		− }
45		− break;
	39	+ case '(':
	40	+ Bungetc(b);
	41	+ return pdfstring(b);
46	42
47	43	case '/':
48		− if((sz = pdfname(p, e, len)) < 0)
49		− goto err;
50		− if((o = malloc(sizeof(*o)+sz+1)) != nil){
51		− o->type = Oname;
52		− o->str = (char*)(o+1);
53		− strcpy(o->str, p);
54		− }
55		− break;
	44	+ Bungetc(b);
	45	+ return pdfname(b);
56	46
57	47	case '[':
58		− o = pdfarray(p, e, len);
59		− break;
	48	+ Bungetc(b);
	49	+ return pdfarray(b);
	50	+
	51	+ case 'n':
	52	+ off = Boffset(b);
	53	+ if(Bgetc(b) == 'u' && Bgetc(b) == 'l' && Bgetc(b) == 'l' && (isws(c = Bgetc(b)) \|\| isdelim(c))){
	54	+ Bungetc(b);
	55	+ return &null;
	56	+ }
	57	+ Bseek(b, off, 0);
	58	+ c = 'f';
	59	+ goto unexpected;
	60	+
	61	+ case 't':
	62	+ off = Boffset(b);
	63	+ tf = 1;
	64	+ if(Bgetc(b) == 'r' && Bgetc(b) == 'u' && Bgetc(b) == 'e' && (isws(c = Bgetc(b)) \|\| isdelim(c)))
	65	+ goto bool;
	66	+ Bseek(b, off, 0);
	67	+ c = 't';
	68	+ goto unexpected;
	69	+
	70	+ case 'f':
	71	+ off = Boffset(b);
	72	+ tf = 0;
	73	+ if(Bgetc(b) == 'a' && Bgetc(b) == 'l' && Bgetc(b) == 's' && Bgetc(b) == 'e' && (isws(c = Bgetc(b)) \|\| isdelim(c)))
	74	+ goto bool;
	75	+ Bseek(b, off, 0);
	76	+ c = 'f';
	77	+ goto unexpected;
	78	+bool:
	79	+ Bungetc(b);
	80	+ if((o = malloc(sizeof(*o))) == nil)
	81	+ goto err;
	82	+ o->type = Obool;
	83	+ o->bool = tf;
	84	+ return o;
60	85
61	86	default:
62		− if(isdigit(*p)){
63		− if((o = malloc(sizeof(*o))) != nil){
64		− o->type = Onum;
65		− o->num = strtod(p, e);
66		− sz = len - (*e - p);
67		− p0 = *e;
68		− s = p0;
69		− for(; sz > 0 && isws(*s); s++, sz--);
70		− if(sz > 0 && isdigit(*s)){
71		− gen = strtod(s, &p);
72		− sz -= (p - s);
73		− s = p;
74		− for(; sz > 0 && isws(*s); s++, sz--);
75		− if(sz > 0 && s == 'R'){ / indirect object */
76		− s++;
77		− o->type = Oindir;
78		− o->indir.id = o->num;
79		− o->indir.gen = gen;
80		− p0 = s;
81		− }
82		− }
83		− *e = p0;
	87	+ if((o = malloc(sizeof(*o))) == nil)
	88	+ goto err;
	89	+ if(!isdigit(c)){
	90	+unexpected:
	91	+ Bungetc(b);
	92	+ werrstr("unexpected char '%c'", c);
	93	+ goto err;
	94	+ }
	95	+ /* it could be a number or an indirect object */
	96	+ Bungetc(b);
	97	+ Bgetd(b, &o->num); /* get the first number */
	98	+ off = Boffset(b); /* seek here if not an indirect object later */
	99	+
	100	+ if((o2 = pdfobject(b)) != nil && o2->type == Onum){ /* second object is number too */
	101	+ do; while(isws(c = Bgetc(b)));
	102	+ if(c < 0)
	103	+ goto err;
	104	+ if(c == 'R'){ /* indirect object */
	105	+ o->type = Oindir;
	106	+ o->indir.id = o->num;
	107	+ o->indir.gen = o2->num;
	108	+ freeobject(o2);
	109	+ return o;
	110	+ }
	111	+ if(c == 'o' && Bgetc(b) == 'b' && Bgetc(b) == 'j'){ /* object */
	112	+ freeobject(o2);
	113	+ /* FIXME put into a map */
	114	+ return pdfobject(b);
84	115	}
85		− break;
86	116	}
87		− werrstr("unexpected char %c", *p);
88		− goto err;
89		− }
90	117
91		− if(o != nil)
	118	+ /* just a number, go back and return it */
	119	+ o->type = Onum;
	120	+ if(Bseek(b, off, 0) != off){
	121	+ werrstr("seek failed");
	122	+ goto err;
	123	+ }
92	124	return o;
	125	+ }
	126	+
93	127	err:
94	128	werrstr("object: %r");
95	129	freeobject(o);
	130	+ freeobject(o2);
96	131	return nil;
97	132	}
98	133
99	134	void
100	135	freeobject(Object *o)
101	136	{
	137	+ int i;
	138	+
102	139	if(o == nil)
103	140	return;
	141	+
	142	+ switch(o->type){
	143	+ case Onull:
	144	+ return;
	145	+
	146	+ case Obool:
	147	+ case Onum:
	148	+ case Ostr:
	149	+ case Oname:
	150	+ break;
	151	+
	152	+ case Oarray:
	153	+ for(i = 0; i < o->array.ne; i++)
	154	+ freeobject(o->array.e[i]);
	155	+ free(o->array.e);
	156	+ break;
	157	+
	158	+ case Odict:
	159	+ case Ostream:
	160	+ case Oindir:
	161	+ break;
	162	+ }
	163	+
104	164	free(o);
105	165	}

pdf.c+5−9

@@ -33,8 +33,8 @@ xrefread(Pdf *pdf, int xref0, int nxref) · 32 unmodified lines
33	33	/* store non-free objects only */
34	34	newnxref = pdf->nxref;
35	35	for(e = s, i = 0; i < nxref; i++, e += 20){
36		− if(e[10] != ' ' \|\| e[18] != '\r' \|\| e[19] != '\n'){
37		− werrstr("invalid xref line");
	36	+ if(!isspace(e[10]) \|\| !isspace(e[18]) \|\| !isspace(e[19])){
	37	+ werrstr("invalid xref line (%d/%d)", i, nxref);
38	38	goto err;
39	39	}
40	40	xref.id = xref0 + i;
@@ -76,15 +76,11 @@ static int · 35 unmodified lines
76	76	trailerread(Pdf *pdf)
77	77	{
78	78	int i;
79		− char s, e;
80	79	Object *o;
81	80	KeyValue *kv;
82	81
83		− o = nil;
84		− if((s = Brdstr(pdf->bio, 0, 1)) == nil \|\| (o = pdfobject(s, &e, Blinelen(pdf->bio))) == nil)
	82	+ if((o = pdfobject(pdf->bio)) == nil)
85	83	goto err;
86		− free(s);
87		− s = nil;
88	84
89	85	if(o->type != Odict){
90	86	werrstr("isn't a dictionary");
@@ -92,7 +88,6 @@ trailerread(Pdf *pdf) · 1 unmodified line
92	88	}
93	89
94	90	for(i = 0, kv = o->dict.kv; i < o->dict.nkv; i++, kv++){
95		− fprint(2, "# %s %p\n", kv->key, kv->value);
96	91	if(strcmp(kv->key, "Root") == 0)
97	92	pdf->root = kv->value;
98	93	else if(strcmp(kv->key, "Info") == 0)
@@ -112,7 +107,6 @@ trailerread(Pdf *pdf) · 13 unmodified lines
112	107	return 0;
113	108	err:
114	109	freeobject(o);
115		− free(s);
116	110	return -1;
117	111	}
118	112
@@ -127,6 +121,7 @@ pdfopen(int fd) · 8 unmodified lines
127	121	int xreftb; /* 7.5.4 xref table offset from the beginning of the file */
128	122	int i, n, off;
129	123
	124	+ b = nil;
130	125	if((pdf = calloc(1, sizeof(*pdf))) == nil \|\| (b = Bfdopen(fd, OREAD)) == nil)
131	126	goto err;
132	127	pdf->bio = b;
@@ -204,6 +199,7 @@ badxref: · 71 unmodified lines
204	199
205	200	return pdf;
206	201	err:
	202	+ werrstr("pdfopen: %r [at %p]", (void*)Boffset(b));
207	203	pdfclose(pdf);
208	204	return nil;
209	205	}

pdf.h+12−38

@@ -20,8 +20,11 @@ struct Object { · 19 unmodified lines
20	20	union {
21	21	int bool;
22	22	double num;
23		− char *str;
24		− char *name;
	23	+ struct {
	24	+ int len;
	25	+ char str[1];
	26	+ };
	27	+ char name[1];
25	28
26	29	struct {
27	30	u32int id;
@@ -50,8 +53,8 @@ struct Pdf { · 22 unmodified lines
50	53	Xref *xref;
51	54	int nxref; /* 7.5.4 xref subsection number of objects */
52	55
53		− Object root; / 7.5.5 root object */
54		− Object info; / 7.5.5 info dictionary */
	56	+ Object root; / 7.7.2 root object */
	57	+ Object info; / 14.3.3 info dictionary */
55	58	};
56	59
57	60	struct Xref {
@@ -63,45 +66,16 @@ struct Xref { · 5 unmodified lines
63	66	Pdf *pdfopen(int fd);
64	67	void pdfclose(Pdf *pdf);
65	68
66		−/*
67		− * General function to parse an object of any type.
68		− */
69		−Object pdfobject(char p, char **e, int len);
70		−
	69	+Object pdfobject(void b);
71	70	void freeobject(Object *o);
72	71
73		−/*
74		− * 7.3.4 String Objects
75		− *
76		− * Rewrites the string in place with null termination and returns the
77		− * length in bytes, without the null terminator.
78		− * Returns < 0 if parsing failed.
79		− * (*e) is advanced to the position after the string pointed by (p).
80		− */
81		−int pdfstring(char p, char *e, int len);
82		−
83		−/*
84		− * 7.3.5 Name Objects
85		− *
86		− * Works the same way as pdfstring, but for name objects.
87		− */
88		−int pdfname(char p, char *e, int len);
89		−
90		−/*
91		− * 7.3.6 Array Objects
92		− */
93		−Object pdfarray(char p, char **e, int len);
94		−
95		−/*
96		− * 7.3.7 Dictionary Objects
97		− */
98		−Object pdfdict(char p, char **e, int len);
99		−
100	72	/*
101	73	* If the object is indirect, resolve it. Operation is not recursive, ie
102	74	* values of a dictionary won't be resolved automatically.
103	75	*/
104	76	int pdfeval(Pdf pdf, Object o);
105	77
106		−int isws(char c);
107		−int isdelim(char c);
	78	+int isws(int c);
	79	+int isdelim(int c);
	80	+
	81	+Object pdfdictget(Object o, char *name);

string.c+163−128

@@ -1,8 +1,11 @@
1	1	#include <u.h>
2	2	#include <libc.h>
	3	+#include <bio.h>
3	4	#include "pdf.h"
4	5
5		−static int esc[] = {
	6	+/* 7.3.4 String Objects */
	7	+
	8	+static char esc[] = {
6	9	['n'] = '\n',
7	10	['r'] = '\r',
8	11	['t'] = '\t',
@@ -14,175 +17,207 @@ static int esc[] = { · 5 unmodified lines
14	17	['\n'] = -1,
15	18	};
16	19
17		−static int
18		−pdfstringhex(char p, char *e, int len)
	20	+static Object *
	21	+stringhex(Biobuf *b)
19	22	{
20		− int i;
	23	+ char *s;
	24	+ Object *o;
	25	+ int len, n;
21	26
22		− for(i = 1; i < len; i += 1){
23		− if(p[i] == '>')
24		− break;
	27	+ if((s = Brdstr(b, '>', 0)) == nil)
	28	+ return nil;
	29	+ len = Blinelen(b) - 1;
	30	+ if(s[len] != '>'){
	31	+ werrstr("no '>'");
	32	+ free(s);
	33	+ return nil;
25	34	}
26		− if(i >= len){
27		− werrstr("hex not closed");
28		− return -1;
29		− }
30		− p[i] = '0'; /* the final zero may be missing */
31		− *e = p+i+1;
32		− i = dec16((uchar*)p, i/2, p+1, i) == i/2 ? i/2 : -1;
33		− if(i < 0)
	35	+ s[len] = '0'; /* the final zero may be missing */
	36	+ n = len/2;
	37	+ o = nil;
	38	+ if(dec16((uchar*)s, n, s+1, len) != n){
34	39	werrstr("invalid hex");
35		− p[i] = 0;
36		− return i;
	40	+ }else if((o = malloc(sizeof(*o) + n + 1)) != nil){
	41	+ memmove(o->str, s, n);
	42	+ o->str[n] = 0;
	43	+ o->len = n;
	44	+ o->type = Ostr;
	45	+ }
	46	+
	47	+ free(s);
	48	+ return o;
37	49	}
38	50
39		−int
40		−pdfstring(char p, char *e, int len)
	51	+Object *
	52	+pdfstring(Biobuf *b)
41	53	{
42		− Rune r;
43		− int c, j, i, o, n, paren;
	54	+ Object *o;
	55	+ char s, r;
44	56	char oct[4];
	57	+ int i, c, paren, sz, maxsz;
45	58
46		− if(len < 2){
47		− werrstr("too short");
48		− goto err;
49		− }
	59	+ maxsz = 64;
	60	+ if((s = malloc(maxsz)) == nil)
	61	+ return nil;
50	62
51		− paren = 0;
52		− for(i = o = 0; i < len;){
53		− if((n = chartorune(&r, p+i)) == 1 && r == Runeerror){
54		− werrstr("rune error at byte %d", n);
55		− return -1;
56		− }
	63	+ for(paren = sz = 0;;){
	64	+ if((c = Bgetc(b)) <= 0)
	65	+ break;
57	66
58		− if(i == 0){
59		− if(r == '('){
60		− paren = 1;
61		− i++;
62		− continue;
63		− }
64		− if(r == '<'){
65		− len = pdfstringhex(p, e, len);
66		− if(len < 0)
67		− goto err;
68		− return len;
	67	+ switch(c){
	68	+ case '<':
	69	+ if(sz == 0){
	70	+ Bungetc(b);
	71	+ return stringhex(b);
69	72	}
70		− werrstr("invalid first char");
71		− goto err;
72		− }
	73	+ break;
73	74
74		− if(r == '\\'){
75		− if(++i >= len){
76		− werrstr("escaped char out of string len");
77		− goto err;
78		− }
79		− if((n = chartorune(&r, p+i)) == 1 && r == Runeerror){
80		− werrstr("rune error at byte %d", i);
81		− goto err;
82		− }
83		− if(r >= '0' && r <= '9'){ /* octal */
84		− n = 0;
85		− for(j = 0; j < 3 && i < len && p[i] >= '0' && p[i] <= '9'; j++, i++)
86		− oct[j] = p[i];
87		− oct[j] = 0;
88		− c = strtol(oct, nil, 8);
89		− }else if(r >= nelem(esc) \|\| (c = esc[r]) == 0){
90		− werrstr("unknown escape char at byte %d", i);
91		− goto err;
92		− }
93		− r = c;
94		− i += n;
95		− if(c < 0)
96		− continue;
97		− }else if(r == '('){
	75	+ case '(':
98	76	paren++;
99		− i++;
100	77	continue;
101		− }else if(r == ')'){
	78	+
	79	+ case ')':
102	80	paren--;
103		− i++;
104		− if(paren == 0)
	81	+ if(paren < 1){
	82	+ c = 0;
105	83	break;
	84	+ }
106	85	continue;
107		− }else{
108		− i += n;
	86	+
	87	+ case '\\':
	88	+ if((c = Bgetc(b)) <= 0)
	89	+ break;
	90	+ if(c >= '0' && c <= '7'){ /* octal */
	91	+ oct[0] = c;
	92	+ for(i = 1; i < 3 && (c = Bgetc(b)) >= '0' && c <= '7'; i++)
	93	+ oct[i] = c;
	94	+ if(c <= 0)
	95	+ break;
	96	+ if(c < '0' \|\| c > '7')
	97	+ Bungetc(b);
	98	+ oct[i] = 0;
	99	+ c = strtol(oct, nil, 8);
	100	+ }else if(c >= nelem(esc) \|\| (c = esc[c]) == 0){
	101	+ werrstr("unknown escape char %c", c);
	102	+ goto err;
	103	+ }else if(c < 0){
	104	+ continue;
	105	+ }
	106	+ break;
	107	+
	108	+ default:
	109	+ if(paren < 1){
	110	+ werrstr("unexpected char '%c'", c);
	111	+ goto err;
	112	+ }
	113	+ break;
109	114	}
110	115
111		− o += runetochar(p+o, &r);
	116	+ if(c <= 0)
	117	+ break;
	118	+ if(sz+1 > maxsz){
	119	+ maxsz *= 2;
	120	+ if((r = realloc(s, maxsz)) == nil)
	121	+ goto err;
	122	+ s = r;
	123	+ }
	124	+ s[sz++] = c;
112	125	}
113	126
114		− if(paren > 0){
115		− werrstr("non-closed paren");
	127	+ if(paren != 0){
	128	+ werrstr("bad paren");
	129	+ goto err;
	130	+ }
	131	+ if(c < 0){
	132	+ werrstr("short");
116	133	goto err;
117	134	}
118	135
119		− p[o] = 0;
120		− *e = p + i;
	136	+ if(c >= 0 && (o = malloc(sizeof(*o) + sz + 1)) != nil){
	137	+ memmove(o->str, s, sz);
	138	+ o->str[sz] = 0;
	139	+ o->len = sz;
	140	+ o->type = Ostr;
	141	+ free(s);
	142	+ return o;
	143	+ }
121	144
122		− return o;
123	145	err:
	146	+ free(s);
124	147	werrstr("string: %r");
125		− return -1;
	148	+ return nil;
126	149	}
127	150
128	151	#ifdef TEST
129	152	static struct {
130		− char *i;
131		− int len;
132		− char *o;
133		− int r;
134		− int e;
	153	+ char *in;
	154	+ char *out;
135	155	}t[] = {
136		− {"(simple string)", 15, "simple string", 13},
137		− {"(non-closed paren", 17, nil, -1},
138		− {"wrong first char", 16, nil, -1},
139		− {"(parens((()((())))()))", 22, "parens", 6},
140		− {"(\\0053)", 7, "\x053", 2},
141		− {"(\\053)", 6, "+", 1},
142		− {"(\\53)", 5, "+", 1},
143		− {"()", 2, "", 0},
144		− {")", 1, nil, -1},
145		− {"(\\)\\()", 6, ")(", 2},
146		− {"(\\\\)", 4, "\\", 1},
147		− {"a", 1, nil, -1},
148		− {"(1\\\n2)", 6, "12", 2},
149		− {"<323130>", 8, "210", 3},
150		− {"<32313>", 7, "210", 3},
151		− {"<>", 2, "", 0},
152		− {"<", 1, nil, -1},
153		− {"<zz>", 4, nil, -1},
	156	+ {"", nil},
	157	+ {"(test, success)", "test, success"},
	158	+ {"(simple string)", "simple string"},
	159	+ {"(non-closed paren", nil},
	160	+ {"wrong first char", nil},
	161	+ {"(parens((()((())))()))", "parens"},
	162	+ {"(\\0053)", "\x053"},
	163	+ {"(\\053)", "+"},
	164	+ {"(\\53)", "+"},
	165	+ {"()", ""},
	166	+ {")", nil},
	167	+ {"(\\)\\()", ")("},
	168	+ {"(\\\\)", "\\"},
	169	+ {"a", nil},
	170	+ {"(1\\\n2)", "12"},
	171	+ {"<323130>", "210"},
	172	+ {"<32313>", "210"},
	173	+ {"<>", ""},
	174	+ {"<", nil},
	175	+ {"<zz>", nil},
	176	+ {">", nil},
154	177	};
155	178
	179	+static char *s;
	180	+static int off, n;
	181	+
	182	+static int
	183	+rd(Biobufhdr , void data, long sz)
	184	+{
	185	+ if(sz > n-off)
	186	+ sz = n-off;
	187	+ memmove(data, s+off, sz);
	188	+ off += sz;
	189	+ return sz;
	190	+}
	191	+
156	192	void
157	193	test_pdfstring(void)
158	194	{
159		− char *e;
160		− int i, r;
	195	+ Object *o;
	196	+ Biobuf b;
	197	+ int i;
161	198
162	199	fprint(2, "pdfstring\n");
163	200	for(i = 0; i < nelem(t); i++){
	201	+ s = t[i].in;
	202	+ n = strlen(s);
	203	+ off = 0;
	204	+ Binit(&b, -1, OREAD);
	205	+ Biofn(&b, rd);
	206	+
164	207	fprint(2, "\t%d: ", i);
165		− r = pdfstring(t[i].i, &e, t[i].len);
166		− if(r != t[i].r){
167		− fprint(2, "expected r=%d, got %d", t[i].r, r);
168		− if(r < 0)
169		− fprint(2, " (%r)\n");
170		− else
171		− fprint(2, "\n");
172		− continue;
173		− }else if(r >= 0){
174		− if(t[i].i+t[i].len != e){
175		− fprint(2, "expected e=%p, got %p\n", t[i].i+t[i].len, e);
176		− continue;
177		− }else if(strcmp(t[i].o, t[i].i) != 0){
178		− fprint(2, "expected %q, got %q\n", t[i].o, t[i].i);
179		− continue;
180		− }
181		− }
182		− fprint(2, "OK");
183		− if(r < 0)
184		− fprint(2, " (%r)");
185		− fprint(2, "\n");
	208	+ o = pdfstring(&b);
	209	+ if(o == nil && t[i].out != nil)
	210	+ fprint(2, "ERROR: expected %q, got error: %r\n", t[i].out);
	211	+ else if(o != nil && t[i].out == nil)
	212	+ fprint(2, "ERROR: expected error, got %q\n", o->str);
	213	+ else if(o == nil && t[i].out == nil)
	214	+ fprint(2, "OK (%r)\n");
	215	+ else if(strcmp(o->str, t[i].out) != 0)
	216	+ fprint(2, "ERROR: expected %q, got %q\n", t[i].out, o->str);
	217	+ else
	218	+ fprint(2, "OK\n");
	219	+ freeobject(o);
	220	+ Bterm(&b);
186	221	}
187	222	}
188	223	#endif

rewrite the API, support more object types and actual evaluation

Keyboard shortcuts

General

Issues and pull requests

Issues, pull requests and Inbox

Repository code