home *** CD-ROM | disk | FTP | other *** search
/ ETO Development Tools 4 / ETO Development Tools 4.iso / Essentials / MacApp Documentation / MacApp.TECH$ Archives / 1991 / Jan 91 / MacApp.Tech$ 1⁄18⁄91 / 2700-Re Re[m] Pascal '9x-Jan91 < prev    next >
Encoding:
Text File  |  1991-03-06  |  2.1 KB  |  54 lines  |  [TEXT/GEOL]
  1. Item    3805834                         16-Jan-91        08:47
  2.  
  3. From:   MALCOLM@APPLE.COM@INTERNET#     Gateway to Internet/BITNET/UUCP
  4.  
  5. To:     MACAPP.TECH$                    MacApp Technical
  6.  
  7. Item forwarded by       SPA.DTS      to SPA0144 
  8.  
  9. INTERNET# Document Id: <48155@apple.Apple.COM>
  10.  
  11. ------------------------------------------------------------------------------
  12.  
  13. Sub:    Re: Re[m]: Pascal '9x
  14.  
  15. If you use AppleLink 6.0, you win! The Reply button works for gatewayed E-mail.
  16. Otherwise, copy & paste this: malcolm@Apple.COM@INTERNET#
  17.  
  18. From: malcolm@Apple.COM (Malcolm Slaney)
  19. References: <6342133@AppleLink.Apple.COM>
  20. Lines: 25
  21. Organization: Apple Computer Inc., Cupertino, CA
  22. Newsgroups: apple.mac.app
  23. Path: apple!malcolm
  24. To: macapp.tech$@applelink.apple.com
  25.  
  26.  
  27. KRIPALU@AppleLink.Apple.COM (Kripalu Ctr, Michael A Latta,PRT) writes:
  28. >    I strongly agree.  After several years of Smalltalk I find static owners
  29. hip
  30. >a real limitation.  I often find myself copying data to support static
  31. >ownership that would be unneeded in a garbage collection environment.  I also
  32. >acknowledge the difficulty in retrofitting a conventional language with garb
  33. age
  34. >collection because tracing all the data strcutures of a conventional language
  35. >requires a lot of meta data and more complex coding than the Smalltalk
  36. >equivelent.
  37.  
  38. There was some work done at Xerox a few years ago that showed that adding
  39. Garbage Collection (GC) to a conventional language like C wasn't that hard.
  40. They basically scanned memory looking for anything that might be a pointer.
  41. Odd numbers in memory, for example, can't possibly be pointers.  Also, most
  42. small integers are probably not addresses.  Anyway, they went through all
  43. of memory, marking any block that had a pointer to it.   If you don't have
  44. a pointer to the block then you can assume it is free.
  45.  
  46. It worked very well.  They described finding several memory leaks in
  47. conventional UNIX code.
  48.  
  49. This of course doesn't work if you disguise your pointers.  This is before
  50. my time but I guess Mac programmers were good at adding a few extra flags to
  51. a 24 bit address.
  52.  
  53.                                                                 Malcolm
  54.