home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!edcastle!edcogsci!cogsci!rjc
- From: rjc@cogsci.ed.ac.uk (Richard Caley)
- Newsgroups: comp.windows.x
- Subject: Re: FLAME, FLAME ON X!!!
- Message-ID: <RJC.92Nov19154904@daiches.cogsci.ed.ac.uk>
- Date: 19 Nov 92 15:49:04 GMT
- References: <1683@igd.fhg.de> <1992Nov9.235741.25166@dsd.es.com>
- <1992Nov18.183826.2652@lsl.co.uk>
- Sender: rjc@cogsci.ed.ac.uk
- Organization: Human Communication Research Center
- Lines: 28
- In-reply-to: snail@lsl.co.uk's message of 18 Nov 92 17:38:26 GMT
-
- In article <1992Nov18.183826.2652@lsl.co.uk>, snail (s) writes:
-
- s> *** This is not a UNIX feature, this is a C feature, to be found on
- s> PCs, Ataris , UNIX boxes, VAXes etc. This a feature, ie: C is
- s> designed this way. ****
-
- Er, no. C only specifies that you have malloc and free. How those are
- implimented is up to the person writing the library. It's a unix
- problem in as far as many unixes don't allow memory to come back and
- that made mallocs which _could_ return memory non-starters, other
- system inherited from there. As for messydos, the best memory
- reclemation scheme is to wipe the disk and install an operating system
- to give the memory back to.
-
- Lets say it's a feature of poor memory allocators, with a rider that
- memory allocation is one of those problems where just about anything
- general is poor and anything good is so specific it will be awful in
- some common cases. I'd love to be proved wrong here, while you're at
- it, a perfect garbage collector and P=NP would be nice too. :-)
-
- Sorry, not very X related anymore, but it looked like a misconception
- that needed to be nailed before we all have to live with non-returning
- mallocs for ever 'cos people think they have to be that way...
-
- --
- rjc@cogsci.ed.ac.uk _O_
- |<
-
-