add and use flate filter

Sigrid Haflínudóttir committed 6 years ago

Commit a9516691 parent ef6cdd010 files · +400 −37Browse files

Unified · Split

10 files

buffer.c +121 −0
dict.c +1 −1
f_flate.c +38 −0
filter.c +69 −14
main.c +1 −1
mkfile +3 −0
object.c +27 −2
pdf.c +14 −13
pdf.h +33 −6
stream.c +93 −0

@@ -0,0 +1,121 @@
	1	+#include <u.h>
	2	+#include <libc.h>
	3	+#include <bio.h>
	4	+#include "pdf.h"
	5	+
	6	+static int
	7	+bufgrow(Buffer *b, int sz)
	8	+{
	9	+ uchar *r;
	10	+ int maxsz;
	11	+
	12	+ if(b->maxsz < 1){
	13	+ if((b->b = mallocz(128, 1)) == nil)
	14	+ return -1;
	15	+ b->maxsz = 128;
	16	+ }
	17	+ for(maxsz = b->maxsz; b->sz+sz > maxsz; maxsz *= 2);
	18	+ if(maxsz != b->maxsz){
	19	+ if((r = realloc(b->b, maxsz)) == nil)
	20	+ return -1;
	21	+ memset(r+b->maxsz, 0, maxsz-b->maxsz);
	22	+ b->b = r;
	23	+ b->maxsz = maxsz;
	24	+ }
	25	+
	26	+ return 0;
	27	+}
	28	+
	29	+void
	30	+bufinit(Buffer b, uchar d, int sz)
	31	+{
	32	+ memset(b, 0, sizeof(*b));
	33	+ if(d != nil){
	34	+ b->b = d;
	35	+ b->sz = sz;
	36	+ b->ro = 1;
	37	+ }
	38	+}
	39	+
	40	+void
	41	+buffree(Buffer *b)
	42	+{
	43	+ if(b->ro == 0)
	44	+ free(b->b);
	45	+}
	46	+
	47	+int
	48	+bufeof(Buffer *b)
	49	+{
	50	+ return b->off == b->sz;
	51	+}
	52	+
	53	+uchar *
	54	+bufdata(Buffer b, int sz)
	55	+{
	56	+ *sz = b->sz;
	57	+ return b->b;
	58	+}
	59	+
	60	+int
	61	+bufreadn(Buffer b, void bio, int sz)
	62	+{
	63	+ int n, end;
	64	+
	65	+ if(bufgrow(b, sz) != 0)
	66	+ return -1;
	67	+ for(end = b->sz+sz; b->sz < end; b->sz += n){
	68	+ if((n = Bread(bio, b->b+b->sz, sz)) < 1)
	69	+ return -1;
	70	+ sz -= n;
	71	+ }
	72	+ return 0;
	73	+}
	74	+
	75	+int
	76	+bufput(Buffer b, uchar d, int sz)
	77	+{
	78	+ if(b->ro)
	79	+ sysfatal("bufferput on readonly buffer");
	80	+ if(bufgrow(b, sz) != 0)
	81	+ return -1;
	82	+
	83	+ memmove(b->b+b->sz, d, sz);
	84	+ b->sz += sz;
	85	+
	86	+ return sz;
	87	+}
	88	+
	89	+int
	90	+bufget(Buffer b, uchar d, int sz)
	91	+{
	92	+ if(sz == 0)
	93	+ return 0;
	94	+
	95	+ if(b->off > b->sz)
	96	+ sysfatal("buffer: off(%d) > sz(%d)", b->off, b->sz);
	97	+ if(sz > b->sz - b->off)
	98	+ sz = b->sz - b->off;
	99	+ memmove(d, b->b+b->off, sz);
	100	+ b->off += sz;
	101	+ b->eof = sz == 0;
	102	+
	103	+ return sz;
	104	+}
	105	+
	106	+void
	107	+bufdump(Buffer *b)
	108	+{
	109	+ Biobuf bio;
	110	+ int i, j;
	111	+
	112	+ Binit(&bio, 2, OWRITE);
	113	+ Bprint(&bio, "%d bytes:\n", b->sz);
	114	+ for(i = 0; i < b->sz;){
	115	+ Bprint(&bio, "%04x\t", i);
	116	+ for(j = 0; i < b->sz && j < 16; j++, i++)
	117	+ Bprint(&bio, "%02x%s", b->b[i], (j+1)&7 ? " " : " ");
	118	+ Bprint(&bio, "\n");
	119	+ }
	120	+ Bterm(&bio);
	121	+}

dict.c+1−1

@@ -69,7 +69,7 @@ pdfdictget(Object o, char name) · 68 unmodified lines
69	69	{
70	70	int i;
71	71
72		− if(o == nil \|\| o->type != Odict \|\| name == nil)
	72	+ if(o == nil \|\| (o->type != Ostream && o->type != Odict) \|\| name == nil)
73	73	return nil;
74	74	for(i = 0; i < o->dict.nkv && strcmp(name, o->dict.kv[i].key) != 0; i++);
75	75

f_flate.c+38−0

@@ -0,0 +1,38 @@
	1	+#include <u.h>
	2	+#include <libc.h>
	3	+#include <bio.h>
	4	+#include <flate.h>
	5	+#include "pdf.h"
	6	+
	7	+static int
	8	+bw(void aux, void d, int n)
	9	+{
	10	+ return bufput(aux, d, n);
	11	+}
	12	+
	13	+static int
	14	+bget(void *aux)
	15	+{
	16	+ uchar c;
	17	+
	18	+ return bufget(aux, &c, 1) == 1 ? c : -1;
	19	+}
	20	+
	21	+int
	22	+fFlate(void aux, Buffer bi, Buffer *bo)
	23	+{
	24	+ int r;
	25	+
	26	+ USED(aux);
	27	+
	28	+ do{
	29	+ r = inflatezlib(bo, bw, bi, bget);
	30	+ }while(r == FlateOk && !bufeof(bi));
	31	+
	32	+ if(r != FlateOk){
	33	+ werrstr("%s", flateerr(r));
	34	+ return -1;
	35	+ }
	36	+
	37	+ return 0;
	38	+}

filter.c+69−14

@@ -3,17 +3,72 @@ · 2 unmodified lines
3	3	#include <bio.h>
4	4	#include "pdf.h"
5	5
6		−/*
7		−7.4
8		−
9		−ASCIIHex
10		−ASCII85
11		−LZW
12		−Flate
13		−RunLength
14		−CCITTFax
15		−JBIG2
16		−DCT
17		−JPX
18		−Crypt
19		−*/
	6	+/* 7.4 Filters */
	7	+
	8	+struct Filter {
	9	+ char *name;
	10	+ int (readall)(void aux, Buffer bi, Buffer bo);
	11	+ int (open)(Filter f, Object *o);
	12	+ void (close)(Filter f);
	13	+ void *aux;
	14	+};
	15	+
	16	+int fFlate(void aux, Buffer bi, Buffer *bo);
	17	+
	18	+static Filter filters[] = {
	19	+ {"ASCII85Decode", nil, nil, nil},
	20	+ {"ASCIIHexDecode", nil, nil, nil},
	21	+ {"CCITTFaxDecode", nil, nil, nil},
	22	+ {"CryptDecode", nil, nil, nil},
	23	+ {"DCTDecode", nil, nil, nil},
	24	+ {"FlateDecode", fFlate},
	25	+ {"JBIG2Decode", nil, nil, nil},
	26	+ {"JPXDecode", nil, nil, nil},
	27	+ {"LZWDecode", nil, nil, nil},
	28	+ {"RunLengthDecode", nil, nil, nil},
	29	+};
	30	+
	31	+Filter *
	32	+filteropen(char name, Object o)
	33	+{
	34	+ int i;
	35	+ Filter *f;
	36	+
	37	+ for(i = 0; i < nelem(filters) && strcmp(filters[i].name, name) != 0; i++);
	38	+ if(i >= nelem(filters)){
	39	+ werrstr("no such filter %q", name);
	40	+ return nil;
	41	+ }
	42	+ if(filters[i].readall == nil){
	43	+ werrstr("filter %q not implemented", name);
	44	+ return nil;
	45	+ }
	46	+ if((f = malloc(sizeof(*f))) == nil)
	47	+ return nil;
	48	+ memmove(f, &filters[i], sizeof(*f));
	49	+ if(f->open != nil && f->open(f, o) != 0){
	50	+ free(f);
	51	+ return nil;
	52	+ }
	53	+
	54	+ return f;
	55	+}
	56	+
	57	+int
	58	+filterrun(Filter f, Buffer bi, Buffer *bo)
	59	+{
	60	+ if(f->readall(f->aux, bi, bo) != 0){
	61	+ werrstr("filter[%s]: %r", f->name);
	62	+ return -1;
	63	+ }
	64	+ fprint(2, "filter[%s]: %d → %d %d\n", f->name, bi->sz, bo->sz, bo->off);
	65	+ return 0;
	66	+}
	67	+
	68	+void
	69	+filterclose(Filter *f)
	70	+{
	71	+ if(f->close != nil)
	72	+ f->close(f);
	73	+ free(f);
	74	+}

main.c+1−1

@@ -22,7 +22,7 @@ threadmain(int argc, char **argv) · 21 unmodified lines
22	22	Object *v;
23	23
24	24	quotefmtinstall();
25		− deflateinit();
	25	+ inflateinit();
26	26
27	27	ARGBEGIN{
28	28	default:

mkfile+3−0

@@ -5,8 +5,10 @@ TARG=pdfs · 4 unmodified lines
5	5
6	6	OFILES=\
7	7	array.$O\
	8	+ buffer.$O\
8	9	dict.$O\
9	10	eval.$O\
	11	+ f_flate.$O\
10	12	filter.$O\
11	13	main.$O\
12	14	misc.$O\
@@ -14,6 +16,7 @@ OFILES=\ · 1 unmodified line
14	16	object.$O\
15	17	pdf.$O\
16	18	pdfs.$O\
	19	+ stream.$O\
17	20	string.$O\
18	21
19	22	HFILES=\

object.c+27−2

@@ -13,6 +13,31 @@ static Object null = { · 12 unmodified lines
13	13	.type = Onull,
14	14	};
15	15
	16	+static char *otypes[] = {
	17	+ [Obool] = "bool",
	18	+ [Onum] = "num",
	19	+ [Ostr] = "str",
	20	+ [Oname] = "name",
	21	+ [Oarray] = "array",
	22	+ [Odict] = "dict",
	23	+ [Ostream] = "stream",
	24	+ [Onull] = "null",
	25	+ [Oindir] = "indir",
	26	+};
	27	+
	28	+int
	29	+Tfmt(Fmt *f)
	30	+{
	31	+ Object *o;
	32	+
	33	+ o = va_arg(f->args, Object*);
	34	+ if(o == nil \|\| o == &null)
	35	+ return fmtprint(f, "null");
	36	+ if(o->type < 0 \|\| o->type >= nelem(otypes))
	37	+ return fmtprint(f, "????");
	38	+ return fmtprint(f, "%s", otypes[o->type]);
	39	+}
	40	+
16	41	/* General function to parse an object of any type. */
17	42	Object *
18	43	pdfobj(Pdf pdf, void b)
@@ -49,8 +74,8 @@ pdfobj(Pdf pdf, void b) · 30 unmodified lines
49	74	goto err;
50	75	}
51	76	o->type = Ostream;
52		− o->stream.length = m->num;
53		− o->stream.offset = Boffset(b);
	77	+ o->stream.len = m->num;
	78	+ o->stream.off = Boffset(b);
54	79	return o;
55	80	}
56	81	Bseek(b, off, 0);

pdf.c+14−13

@@ -75,9 +75,7 @@ err: · 74 unmodified lines
75	75	static int
76	76	trailerread(Pdf *pdf)
77	77	{
78		− int i;
79	78	Object *o;
80		− KeyValue *kv;
81	79
82	80	if((o = pdfobj(pdf, pdf->bio)) == nil)
83	81	goto err;
@@ -87,12 +85,8 @@ trailerread(Pdf *pdf) · 3 unmodified lines
87	85	goto err;
88	86	}
89	87
90		− for(i = 0, kv = o->dict.kv; i < o->dict.nkv; i++, kv++){
91		− if(strcmp(kv->key, "Root") == 0)
92		− pdf->root = kv->value;
93		− else if(strcmp(kv->key, "Info") == 0)
94		− pdf->info = kv->value;
95		− }
	88	+ pdf->root = pdfdictget(o, "Root");
	89	+ pdf->info = pdfdictget(o, "Info");
96	90	pdfobjfree(o);
97	91	o = nil;
98	92
@@ -101,8 +95,6 @@ trailerread(Pdf *pdf) · 2 unmodified lines
101	95	werrstr("no root");
102	96	goto err;
103	97	}
104		− if(pdfeval(pdf, pdf->root) != 0 \|\| pdfeval(pdf, pdf->info) != 0)
105		− goto err;
106	98
107	99	return 0;
108	100	err:
@@ -121,6 +113,9 @@ pdfopen(int fd) · 12 unmodified lines
121	113	int nxref; /* 7.5.4 xref subsection number of objects */
122	114	int xreftb; /* 7.5.4 xref table offset from the beginning of the file */
123	115	int i, n, off;
	116	+ Stream *stream;
	117	+
	118	+ fmtinstall('T', Tfmt);
124	119
125	120	b = nil;
126	121	o = nil;
@@ -196,10 +191,16 @@ badxref: · 69 unmodified lines
196	191	}
197	192	}else if(isdigit(tmp[0])){ /* could be 7.5.8 xref stream (since PDF 1.5) */
198	193	Bseek(b, xreftb, 0);
199		− if((o = pdfobj(pdf, b)) == nil \|\| pdfeval(pdf, o) != 0)
200		− goto err;
201		−
	194	+ if((o = pdfobj(pdf, b)) == nil \|\| (stream = streamopen(pdf, o)) == nil)
	195	+ goto badxref;
	196	+ streamclose(stream);
	197	+ pdf->root = pdfdictget(o, "Root");
	198	+ pdf->info = pdfdictget(o, "Info");
202	199	}
	200	+ if(pdfeval(pdf, pdf->root) != 0 \|\| pdfeval(pdf, pdf->info) != 0)
	201	+ goto err;
	202	+ fprint(2, "root %T\n", pdf->root);
	203	+ fprint(2, "info %T\n", pdf->info);
203	204
204	205	return pdf;
205	206	err:

pdf.h+33−6

@@ -10,11 +10,23 @@ enum { · 9 unmodified lines
10	10	Oindir, /* 7.3.10 */
11	11	};
12	12
	13	+typedef struct Buffer Buffer;
	14	+typedef struct Filter Filter;
13	15	typedef struct KeyValue KeyValue;
14	16	typedef struct Object Object;
15	17	typedef struct Pdf Pdf;
16		−typedef struct Xref Xref;
17	18	typedef struct Stream Stream;
	19	+typedef struct Xref Xref;
	20	+#pragma incomplete Filter
	21	+
	22	+struct Buffer {
	23	+ uchar *b;
	24	+ int ro;
	25	+ int maxsz;
	26	+ int sz;
	27	+ int off;
	28	+ int eof;
	29	+};
18	30
19	31	struct Object {
20	32	int type;
@@ -45,8 +57,8 @@ struct Object { · 24 unmodified lines
45	57	struct {
46	58	KeyValue *kv;
47	59	int nkv;
48		− u32int length; /* packed */
49		− u32int offset;
	60	+ u32int len; /* packed */
	61	+ u32int off;
50	62	}stream;
51	63	};
52	64	};
@@ -72,9 +84,8 @@ struct Xref { · 19 unmodified lines
72	84	};
73	85
74	86	struct Stream {
75		− Biobuf;
76		− Object *o;
77		− u32int offset;
	87	+ Buffer buf;
	88	+ void *bio;
78	89	};
79	90
80	91	Pdf *pdfopen(int fd);
@@ -96,3 +107,19 @@ Object pdfdictget(Object o, char *name); · 15 unmodified lines
96	107
97	108	Stream streamopen(Pdf pdf, Object *o);
98	109	void streamclose(Stream *s);
	110	+
	111	+Filter filteropen(char name, Object *o);
	112	+int filterrun(Filter f, Buffer bi, Buffer *bo);
	113	+void filterclose(Filter *f);
	114	+
	115	+void bufinit(Buffer b, uchar d, int sz);
	116	+void buffree(Buffer *b);
	117	+int bufeof(Buffer *b);
	118	+uchar bufdata(Buffer b, int *sz);
	119	+int bufreadn(Buffer b, void bio, int sz);
	120	+int bufput(Buffer b, uchar d, int sz);
	121	+int bufget(Buffer b, uchar d, int sz);
	122	+void bufdump(Buffer *b);
	123	+
	124	+#pragma varargck type "T" Object *
	125	+int Tfmt(Fmt *f);

stream.c+93−0

@@ -0,0 +1,93 @@
	1	+#include <u.h>
	2	+#include <libc.h>
	3	+#include <bio.h>
	4	+#include "pdf.h"
	5	+
	6	+static int
	7	+bufiof(Biobufhdr b, void data, long n)
	8	+{
	9	+ Stream *s;
	10	+
	11	+ s = (void)((char)b - sizeof(*s));
	12	+
	13	+ return bufget(&s->buf, data, n);
	14	+}
	15	+
	16	+Stream *
	17	+streamopen(Pdf pdf, Object o)
	18	+{
	19	+ Stream *s;
	20	+ Buffer b, x;
	21	+ Object of, *flts;
	22	+ Filter *f;
	23	+ int i, nflts;
	24	+
	25	+ s = nil;
	26	+ if(pdfeval(pdf, o) != 0 \|\| o == nil \|\| o->type != Ostream) /* FIXME open a string object as a stream as well? */
	27	+ return nil;
	28	+
	29	+ bufinit(&b, nil, 0);
	30	+ if(Bseek(pdf->bio, o->stream.off, 0) != o->stream.off)
	31	+ return nil;
	32	+ if(bufreadn(&b, pdf->bio, o->stream.len) < 0)
	33	+ goto err;
	34	+ bufdump(&b);
	35	+
	36	+ /* see if there are any filters */
	37	+ if((of = pdfdictget(o, "Filter")) != nil){
	38	+ if(pdfeval(pdf, of) != 0)
	39	+ goto err;
	40	+ if(of->type == Oname){ /* one filter */
	41	+ flts = &of;
	42	+ nflts = 1;
	43	+ }else if(of->type == Oarray){ /* array of filters */
	44	+ flts = of->array.e;
	45	+ nflts = of->array.ne;
	46	+ }else{
	47	+ werrstr("filters type invalid (%T)", of);
	48	+ goto err;
	49	+ }
	50	+
	51	+ for(i = nflts-1; i >= 0; i--){
	52	+ if(flts[i]->type != Oname){
	53	+ werrstr("filter type invalid (%T)", flts[i]);
	54	+ goto err;
	55	+ }
	56	+ if((f = filteropen(flts[i]->name, o)) == nil)
	57	+ goto err;
	58	+ bufinit(&x, nil, 0);
	59	+ if(filterrun(f, &b, &x) != 0){
	60	+ buffree(&x);
	61	+ goto err;
	62	+ }
	63	+ buffree(&b);
	64	+ b = x;
	65	+ }
	66	+ }
	67	+
	68	+ if((s = calloc(1, sizeof(*s)+sizeof(Biobuf))) == nil){
	69	+ buffree(&b);
	70	+ return nil;
	71	+ }
	72	+ s->bio = (uchar*)(s+1);
	73	+ s->buf = b;
	74	+ Binit(s->bio, Bfildes(pdf->bio), OREAD);
	75	+ Biofn(s->bio, bufiof);
	76	+
	77	+ bufdump(&s->buf);
	78	+
	79	+ return s;
	80	+err:
	81	+ werrstr("stream: %r");
	82	+ buffree(&b);
	83	+ free(s);
	84	+ return nil;
	85	+}
	86	+
	87	+void
	88	+streamclose(Stream *s)
	89	+{
	90	+ buffree(&s->buf);
	91	+ Bterm(s->bio);
	92	+ free(s);
	93	+}

add and use flate filter

Keyboard shortcuts

General

Issues and pull requests

Issues, pull requests and Inbox

Repository code