shithub: pdffs

Info  •  Files  •  Log

xref reading: workaround for broken PDFs with startxref pointint to garbage (not "xref" table)

--- a/pdf.c

+++ b/pdf.c

@@ -109,9 +109,26 @@

 		}else if(isdigit(tmp[0])){ /* could be 7.5.8 xref stream (since PDF 1.5) */

 			if(Sseek(stream, xreftb, 0) != xreftb)

 				goto badxref;

-			if(xrefreadstream(pdf) != 0)

-				goto err;

-			break;

+			if(xrefreadstream(pdf) != 0){

+lastresort:

+				/* some PDFs have "startxref" pointing at garbage */

+				for(i = 0; i < 4096; i++){

+					if(Sread(stream, tmp, 4) != 4)

+						goto badxref;

+					if(memcmp(tmp, "xref", 4) == 0){

+						Sseek(stream, -4, 1);

+						if(xrefreadold(pdf) != 0)

+							goto badxref;

+						break;

+					}

+					Sseek(stream, -3, 1);

+				}

+				if(i >= 4096)

+					goto err;

+			}else

+				break;

+		}else{

+			goto lastresort;

 		}

 	}