dany/pdfio
1
0
Fork 0
mirror of https://github.com/michaelrsweet/pdfio.git synced 2024-12-27 05:48:20 +01:00
Code Issues Packages Projects Releases Wiki Activity

Update token reader to handle balanced parens and to allow arbitrary backslash escapes.

Browse Source 
Update unit test to omit path information from filename.
This commit is contained in:
Michael R Sweet 2021-05-09 21:32:09 -04:00
parent 5f43761537
commit af8c731699
No known key found for this signature in database
GPG Key ID: 999559A027815955
2 changed files with 29 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
pdfio-token.c
View File

@ -148,7 +148,8 @@ _pdfioTokenRead(_pdfio_token_t *tb, // I - Token buffer/stack
char *buffer, // I - String buffer
size_t bufsize) // I - Size of string buffer
{
int ch; // Character
int ch, // Character
parens = 0; // Parenthesis level
char *bufptr, // Pointer into buffer
*bufend, // End of buffer
state = '\0'; // Current state
@ -210,7 +211,7 @@ _pdfioTokenRead(_pdfio_token_t *tb, // I - Token buffer/stack
switch (state)
{
case '(' : // Literal string
while ((ch = get_char(tb)) != EOF && ch != ')')
while ((ch = get_char(tb)) != EOF)
{
if (ch == '\\')
{
@ -267,10 +268,22 @@ _pdfioTokenRead(_pdfio_token_t *tb, // I - Token buffer/stack
break;
default :
_pdfioFileError(tb->pdf, "Unknown escape '\\%c' in literal string.", ch);
return (false);
// Ignore blackslash per PDF spec...
break;
}
}
else if (ch == '(')
{
// Keep track of parenthesis
parens ++;
}
else if (ch == ')')
{
if (parens == 0)
break;
parens --;
}
if (bufptr < bufend)
{

17
testpdfio.c
View File

@ -37,16 +37,23 @@ main(int argc, // I - Number of command-line arguments
{
if ((pdf = pdfioFileOpen(argv[i], NULL, NULL)) != NULL)
{
const char *filename; // Base filename for messages
if ((filename = strrchr(argv[i], '/')) != NULL)
filename ++;
else
filename = argv[i];
num_objs = pdfioFileGetNumObjects(pdf);
num_pages = pdfioFileGetNumPages(pdf);
printf("%s: PDF %s, %d pages, %d objects.\n", argv[i], pdfioFileGetVersion(pdf), (int)num_pages, (int)num_objs);
printf("%s: PDF %s, %d pages, %d objects.\n", filename, pdfioFileGetVersion(pdf), (int)num_pages, (int)num_objs);
for (n = 0; n < num_pages; n ++)
{
if ((obj = pdfioFileGetPage(pdf, n)) == NULL)
{
printf("%s: Unable to get page #%d.\n", argv[i], (int)n + 1);
printf("%s: Unable to get page #%d.\n", filename, (int)n + 1);
}
else
{
@ -64,7 +71,7 @@ main(int argc, // I - Number of command-line arguments
}
}
printf("%s: Page #%d is %gx%g.\n", argv[i], (int)n + 1, media_box.x2, media_box.y2);
printf("%s: Page #%d is %gx%g.\n", filename, (int)n + 1, media_box.x2, media_box.y2);
}
}
@ -72,7 +79,7 @@ main(int argc, // I - Number of command-line arguments
{
if ((obj = pdfioFileGetObject(pdf, n)) == NULL)
{
printf("%s: Unable to get object #%d.\n", argv[i], (int)n);
printf("%s: Unable to get object #%d.\n", filename, (int)n);
}
else
{
@ -81,7 +88,7 @@ main(int argc, // I - Number of command-line arguments
dict = pdfioObjGetDict(obj);
printf("%s: %u %u obj dict=%p(%lu)\n", argv[i], (unsigned)pdfioObjGetNumber(obj), (unsigned)pdfioObjGetGeneration(obj), dict, dict ? (unsigned long)dict->num_pairs : 0UL);
printf("%s: %u %u obj dict=%p(%lu)\n", filename, (unsigned)pdfioObjGetNumber(obj), (unsigned)pdfioObjGetGeneration(obj), dict, dict ? (unsigned long)dict->num_pairs : 0UL);
if (dict)
{
for (np = dict->num_pairs, pair = dict->pairs; np > 0; np --, pair ++)