More exact PDF pattern and filesize limit
Also, can now feed image in through stdin (useful for `pv image.iso | ./revocer-pdfs ...` workflow
This commit is contained in:
@@ -4,17 +4,31 @@
|
|||||||
#include <sys/stat.h>
|
#include <sys/stat.h>
|
||||||
|
|
||||||
#define CHUNK 4096
|
#define CHUNK 4096
|
||||||
|
#define MAX_PDF_SIZE (100LL * 1024 * 1024) // 100 MB
|
||||||
|
|
||||||
int match_at(const unsigned char *buf, size_t len, const char *pat) {
|
int valid_pdf_header(const unsigned char *buf, size_t len) {
|
||||||
size_t p = strlen(pat);
|
if (len < 9) return 0;
|
||||||
if (len < p) return 0;
|
|
||||||
return memcmp(buf, pat, p) == 0;
|
if (memcmp(buf, "%PDF-", 5) != 0)
|
||||||
|
return 0;
|
||||||
|
|
||||||
|
if (buf[5] < '0' || buf[5] > '9') return 0;
|
||||||
|
if (buf[6] != '.') return 0;
|
||||||
|
if (buf[7] < '0' || buf[7] > '9') return 0;
|
||||||
|
|
||||||
|
size_t i = 8;
|
||||||
|
|
||||||
|
if (buf[i] >= '0' && buf[i] <= '9') i++;
|
||||||
|
|
||||||
|
if (buf[i] == '\r' && buf[i + 1] == '\n') return 1;
|
||||||
|
if (buf[i] == '\n') return 1;
|
||||||
|
|
||||||
|
return 0;
|
||||||
}
|
}
|
||||||
|
|
||||||
size_t find_in_buffer(const unsigned char *buf, size_t len, const char *pat) {
|
size_t find_eof(const unsigned char *buf, size_t len) {
|
||||||
size_t p = strlen(pat);
|
for (size_t i = 0; i + 5 <= len; i++) {
|
||||||
for (size_t i = 0; i + p <= len; i++) {
|
if (memcmp(buf + i, "%%EOF", 5) == 0)
|
||||||
if (memcmp(buf + i, pat, p) == 0)
|
|
||||||
return i;
|
return i;
|
||||||
}
|
}
|
||||||
return (size_t)-1;
|
return (size_t)-1;
|
||||||
@@ -42,6 +56,9 @@ int main(int argc, char **argv) {
|
|||||||
|
|
||||||
mkdir(outdir, 0755);
|
mkdir(outdir, 0755);
|
||||||
|
|
||||||
|
if (strcmp(infile, "-") == 0) {
|
||||||
|
infile = "/dev/stdin";
|
||||||
|
}
|
||||||
FILE *f = fopen(infile, "rb");
|
FILE *f = fopen(infile, "rb");
|
||||||
if (!f) {
|
if (!f) {
|
||||||
perror("open input");
|
perror("open input");
|
||||||
@@ -50,11 +67,12 @@ int main(int argc, char **argv) {
|
|||||||
|
|
||||||
unsigned char *buf = malloc(CHUNK * 2);
|
unsigned char *buf = malloc(CHUNK * 2);
|
||||||
size_t buf_len = 0;
|
size_t buf_len = 0;
|
||||||
int pdf_count = 0;
|
|
||||||
|
|
||||||
unsigned char *file_data = NULL;
|
unsigned char *pdf_buf = NULL;
|
||||||
size_t file_size = 0;
|
size_t pdf_size = 0;
|
||||||
|
|
||||||
int collecting = 0;
|
int collecting = 0;
|
||||||
|
int pdf_count = 0;
|
||||||
|
|
||||||
while (!feof(f)) {
|
while (!feof(f)) {
|
||||||
size_t r = fread(buf + buf_len, 1, CHUNK, f);
|
size_t r = fread(buf + buf_len, 1, CHUNK, f);
|
||||||
@@ -65,28 +83,40 @@ int main(int argc, char **argv) {
|
|||||||
while (i < buf_len) {
|
while (i < buf_len) {
|
||||||
|
|
||||||
if (!collecting) {
|
if (!collecting) {
|
||||||
if (i + 5 < buf_len && memcmp(buf + i, "%PDF-", 5) == 0) {
|
if (i + 10 < buf_len && valid_pdf_header(buf + i, buf_len - i)) {
|
||||||
collecting = 1;
|
collecting = 1;
|
||||||
file_data = malloc(1024);
|
pdf_buf = malloc(1024);
|
||||||
file_size = 0;
|
pdf_size = 0;
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
if (collecting) {
|
if (collecting) {
|
||||||
if (file_size % 1024 == 0) {
|
|
||||||
file_data = realloc(file_data, file_size + 1024);
|
// HARD SIZE LIMIT CHECK
|
||||||
|
if (pdf_size >= MAX_PDF_SIZE) {
|
||||||
|
free(pdf_buf);
|
||||||
|
pdf_buf = NULL;
|
||||||
|
pdf_size = 0;
|
||||||
|
collecting = 0;
|
||||||
|
continue;
|
||||||
}
|
}
|
||||||
|
|
||||||
file_data[file_size++] = buf[i];
|
if (pdf_size % 1024 == 0) {
|
||||||
|
pdf_buf = realloc(pdf_buf, pdf_size + 1024);
|
||||||
|
}
|
||||||
|
|
||||||
if (file_size > 6) {
|
pdf_buf[pdf_size++] = buf[i];
|
||||||
if (memmem(file_data, file_size, "%%EOF", 5)) {
|
|
||||||
write_pdf(file_data, file_size, pdf_count++, outdir);
|
size_t eof_pos = find_eof(pdf_buf, pdf_size);
|
||||||
free(file_data);
|
if (eof_pos != (size_t)-1) {
|
||||||
file_data = NULL;
|
size_t end = eof_pos + 5;
|
||||||
file_size = 0;
|
|
||||||
collecting = 0;
|
write_pdf(pdf_buf, end, pdf_count++, outdir);
|
||||||
}
|
|
||||||
|
free(pdf_buf);
|
||||||
|
pdf_buf = NULL;
|
||||||
|
pdf_size = 0;
|
||||||
|
collecting = 0;
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|||||||
Reference in New Issue
Block a user