Skip to content
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
54 changes: 48 additions & 6 deletions PdfPageCount.pas
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,7 @@
*******************************************************************************)

////////////////////////////////////////////////////////////////////////////////
// Summary of steps taken to parse a PDF doc for its page count :-
// Summary of steps taken to parse a PDF doc for its page count :-  
////////////////////////////////////////////////////////////////////////////////
//1. See if there's a 'Linearization dictionary' for easy parsing.
// Mostly there isn't so ...
Expand All @@ -31,14 +31,26 @@
interface

uses
Windows, SysUtils, Classes, AnsiStrings, ZLib, Md5;
{$IFNDEF LINUX}
Windows,
{$ENDIF}
SysUtils,
Classes,
{$IFNDEF FPC}
AnsiStrings,
ZLib,
{$ELSE}
PasZLib,
{$ENDIF}
MD5;

const
PDF_NO_ERROR = 0;
PDF_ERROR_UNDEFINED = -1;
PDF_ERROR_FILE_OPEN = -2;
PDF_ERROR_FILE_FORMAT = -3;
PDF_ERROR_ENCRYPTED_STRM = -4;
PDF_ERROR_NO_INDEX = -5;

(*******************************************************************************
* GetPageCount *
Expand All @@ -51,6 +63,9 @@ function GetPageCount(const filename: string): integer; overload;

implementation

const
MAX_buffer_Size = 64*1024*1024;

type

PPdfObj = ^TPdfObj;
Expand Down Expand Up @@ -116,8 +131,11 @@ TPdfPageCounter = class
// Miscellaneous functions
//------------------------------------------------------------------------------

procedure QuickSortList(SortList: TPointerList;
L, R: Integer; sortFunc: TSortFunc);
{$IFDEF FPC}
procedure QuickSortList(SortList: PPointerList; L, R: Integer; sortFunc: TSortFunc);
{$ELSE}
procedure QuickSortList(SortList: TPointerList; L, R: Integer; sortFunc: TSortFunc);
{$ENDIF}
var
I, J: Integer;
P, T: Pointer;
Expand Down Expand Up @@ -777,12 +795,18 @@ function TPdfPageCounter.DecompressObjIntoBuffer(objNum, genNum: integer): boole

try
//decompress the stream ...
{$IFDEF FPC}
bufferSize := MAX_buffer_Size;
ReallocMem(buffer, bufferSize);
uncompress(pointer(buffer), cardinal(bufferSize), PChar(p), cardinal(len));
{$ELSE}
//nb: I'm not sure in which Delphi version these functions were renamed.
{$IFDEF UNICODE}
zlib.ZDecompress(p, len, pointer(buffer), Integer(bufferSize));
{$ELSE}
zlib.DecompressBuf(p, len, len*3, pointer(buffer), Integer(bufferSize));
{$ENDIF}
{$ENDIF}
except
ErrorFlag := PDF_ERROR_ENCRYPTED_STRM;
DisposeBuffer;
Expand Down Expand Up @@ -811,7 +835,12 @@ function TPdfPageCounter.FindStartXRef: boolean;
'f': dec(p, 8); 'e': dec(p, 7); 'x': dec(p, 5);
'r': dec(p, 3); 'a': dec(p, 2); 't': dec(p, 1);
's':

{$IFDEF FPC}
if StrLComp(p, 'startxref', 9) = 0 then
{$ELSE}
if AnsiStrings.StrLComp(p, 'startxref', 9) = 0 then
{$ENDIF}
begin
result := true;
inc(p, 9);
Expand All @@ -833,7 +862,11 @@ function TPdfPageCounter.GetLinearizedPageNum(out pageNum: integer): boolean;
result := false;
pStop := p + 32;
while (p < pStop) and (p^ <> 'o') do inc(p);
{$IFDEF FPC}
if StrLComp( p, 'obj', 3) <> 0 then exit;
{$ELSE}
if AnsiStrings.StrLComp( p, 'obj', 3) <> 0 then exit;
{$ENDIF}
pStart := p;
if not FindStrInDict('/Linearized') then exit;
p := pStart;
Expand Down Expand Up @@ -902,6 +935,7 @@ function TPdfPageCounter.GetPageNumUsingCrossRefStream: integer;
//if the Index array is empty then use the default values ...
if length(indexArray) = 0 then
begin
Result := PDF_ERROR_NO_INDEX;
setLength(indexArray, 2);
indexArray[0] := 0;
indexArray[1] := bufferSize div (w1 + w2 + w3);
Expand Down Expand Up @@ -950,7 +984,8 @@ function TPdfPageCounter.GetPageNumUsingCrossRefStream: integer;

DisposeBuffer;

QuickSortList(PdfObjList.List, 0, PdfObjList.Count -1, ListSort);
if PdfObjList.Count > 1 then
QuickSortList(PdfObjList.List, 0, PdfObjList.Count -1, ListSort);
if not GotoObject(rootNum) then exit;
if not FindStrInDict('/Pages') then exit;
//get the Pages' object number, go to it and get the page count ...
Expand Down Expand Up @@ -1002,7 +1037,11 @@ function TPdfPageCounter.GetPdfPageCount(stream: TStream): integer;
(k >= ms.size) then exit;

p := PAnsiChar(ms.Memory) + k;
{$IFDEF FPC}
if StrLComp(p, 'xref', 4) <> 0 then
{$ELSE}
if AnsiStrings.StrLComp(p, 'xref', 4) <> 0 then
{$ENDIF}
begin
//'xref' table not found so assume the doc contains
//a cross-reference stream instead (ie PDF doc ver 1.5+)
Expand Down Expand Up @@ -1057,7 +1096,8 @@ function TPdfPageCounter.GetPdfPageCount(stream: TStream): integer;
//Make sure we've got Root's object number ...
if rootNum = $FFFFFFFF then exit;

QuickSortList(PdfObjList.List, 0, PdfObjList.Count -1, ListSort);
if PdfObjList.Count > 1 then
QuickSortList(PdfObjList.List, 0, PdfObjList.Count -1, ListSort);
if not GotoObject(rootNum) then exit;

if not FindStrInDict('/Pages') then exit;
Expand Down Expand Up @@ -1122,3 +1162,5 @@ function GetPageCount(const filename: string): integer;
//------------------------------------------------------------------------------

end.