Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site sdcc6.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!whuxl!whuxlm!harpo!decvax!ittvax!dcdwest!sdcsvax!sdcc3!sdcc6!ix408
From: ix408@sdcc6.UUCP (Cris Rys)
Newsgroups: net.micro.mac
Subject: Re: MacPaint Format?
Message-ID: <1977@sdcc6.UUCP>
Date: Thu, 21-Mar-85 00:51:39 EST
Article-I.D.: sdcc6.1977
Posted: Thu Mar 21 00:51:39 1985
Date-Received: Sat, 23-Mar-85 00:33:18 EST
References: <638@tty3b.UUCP>
Organization: U.C. San Diego, Academic Computer Center
Lines: 98

Here is the information requested on Macpaint documents.
I got this information from Inside Macintosh.

Macpaint documents use only the data fork of the file system; the
resource fork is not used and may be ignored. The data fork contains
a 512 byte header and then the compressed data representing a single
bitmap of 576 pixels wide by 720 pixels tall. At 72 pixels per inch,
this bitmap occupies the full 8 by 10 inch printable area of the
imagewriter printer page.

Header:

The first 512 bytes of the document form a header with a 4 byte
version number (default = 2), then 38*8 = 304 bytes of patterns,
then 204 unused bytes reserved for future expansion. If the version
number is zero, the rest of the header block is ignored and the
default patterns are used, so programs generating Macpaint documents
can simply write out 512 bytes of zero as the document header. Most
programs which read Macpaint documents can simply skip over the
header when reading.

Bitmap:

Following the header are 720 compressed scanlines of data which form
the 576 wide by 720 tall bitmap. Without compression, this bitmap
would occupy 51840 bytes and chew up disk space pretty fast; typical
Macpaint documents compress to about 10k using PackBits procedure in
the Macintosh ROM to compress runs of equal bytes within each
scanline. The bitmap part of a Macpaint document is simply 720 times
the output of PackBits with 72 bytes input.

READING SAMPLE:
program convert_pic;
{ I am not sure what libraries this program uses.}
{ UnPackBits IS in ROM, I'm not sure about ReadData}
 const
  srcBlocks = 2;  {At least 2, bigger makes it faster}
  srcSize = 1024; {512 *srcBlocks}
 type
  diskBlock = packed array[1..512] of QDByte;
 var
  srcBuf : array[1..srcBlocks] of diskBlock;
  srcptr, dstptr : QDPtr;
begin
 { skip the header }
 ReadData(srcfile, @srcBuf, 512);
 { prime srcBuf }
 ReadData(srcFile, @srcBuf, srcSize);

 { unpack each scanline into dstBits, reading more source as needed}
 srcPtr := @srcBuf;
 dstptr := dstBits.baseAddr;
 for scanline := 1 to 720 do
  begin
   Unpackbits(srcptr, desptr, 72);         { bumps both ptrs }
   { time to read next chunk of packed source ? }
   if ord(srcptr) > ord(@srcbuf) + size - 512 then
    begin
     srcbuf[1] := srcBuf[srcblocks];       { move up last block }
     Readdata(srcfile, @srcbuf[2], srcsize - 512);
     srcptr := pointer(ord(srcptr) - srcsize + 512);
    end;
  end;
end.

WRITING SAMPLE:

To write out a 576 by 720 bitmap which is contained in memory, the
following fragment of code could be used:

program write_pic;
 type
  diskBlock = packed array[1..512] of QDByte;
 var
  dstBuf         : diskBlock;
  srcptr, dstptr : QDPtr;
  dstbytes       : integer;
begin
 {write header, all zeros}
 for i:= 1 to 512 do dstBuf[i]:=0;
 writedata(dstfile, @dstBuf, 512);
 
 {Compress each scanline and write it}
 srcptr := srcBits.baseAddr;
 for scanline := 1 to 720 do
  begin
   dstPtr:=@dstBuf;
   packbits(srcptr, desptr, 72);	 {bumps both ptrs}
   dstBytes:=ORD(dstPtr)-ORD(@dstBuf);   {calc packed size}
   WriteData(dstFile,@dstbuf,dstBytes);  {write packed data}
  end;
end.


Both of the two previous programs are only skeletons, but should
work with minimal additions.

Cris Rys