home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.lang.modula2
- Path: sparky!uunet!newsgate.watson.ibm.com!yktnews!admin!nordruth!seurer
- From: seurer@rchland.vnet.ibm.com (Bill Seurer)
- Subject: Re: Generic stack implementation
- Sender: news@rchland.ibm.com
- Message-ID: <1992Dec17.150809.15780@rchland.ibm.com>
- Date: Thu, 17 Dec 1992 15:08:09 GMT
- Reply-To: BillSeurer@vnet.ibm.com
- Disclaimer: This posting represents the poster's views, not necessarily those of IBM
- References: <9212091316.AA14856@elg> <911288c.32.724187845@ace.acadiau.ca> <1992Dec15.150410.21677@informatik.uni-ulm.de> <1992Dec17.114912.1@wombat.newcastle.edu.au>
- Nntp-Posting-Host: nordruth.rchland.ibm.com
- Organization: IBM Rochester
- Lines: 18
-
- I have ported lots of Modula-2 and C code between machines with VASTLY
- different architectures, especially when it comes to addresses. The
- Modula-2 code was generally easier to port but both it and the C code
- suffered from one big problem.
-
- On most Unix machines the size of pointers and integers is the same.
- Vast amounts of C code is written that exploits this fact and makes it
- extremely difficult to port to machines where this isn't the case. Even
- code that is supposed to be relatively machine independent usually has
- this flaw. Often it is very subtle, easy to miss, and worst of all,
- compiles OK.
-
- The "ADDRESS is compatible with CARDINAL" part of Modula-2 is a mistake,
- pure and simple.
- --
-
- - Bill Seurer Language and Compiler Development IBM Rochester, MN
- Internet: BillSeurer@vnet.ibm.com America On-Line: BillSeurer@aol.com
-