home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.mac.programmer
- Path: sparky!uunet!amgen!nuntius
- From: Michael Brennan <mikeb@sam.amgen.com>
- Subject: Re: HINT on upgrading from THINK C 4.0 to 5.0.2
- Message-ID: <1992Aug14.182109.3182@amgen.com>
- Sender: news@amgen.com
- Nntp-Posting-Host: swalker
- Organization: Amgen
- X-Useragent: Nuntius v1.1
- References: <1992Aug5.180617.16041@Csli.Stanford.EDU>
- <PHILS.92Aug7094944@chaos.cs.brandeis.edu>
- Date: Fri, 14 Aug 1992 18:21:09 GMT
- Lines: 24
-
- In article <1992Aug5.193816.27403@hobbes.kzoo.edu> Jamie R. McCarthy,
- k044477@hobbes.kzoo.edu writes:
- >Does anyone else have any problems (known or rumored) with any of the
- >optimizations?
-
- I encountered quite a few very nasty bugs with optimization in 5.0.
- Mostly with the global
- optimizations. In one instance, it took an expression within a loop
- (whose value was
- incrementing in the loop), and decided to use a temporary value for the
- expression
- whose value was defined before entering the loop. The code in the loop
- therefore failed.
- In another instance, it decided two variables were never used at the same
- time,
- so it assigned them the same memory space. In fact, the two variables
- WERE used
- at the same time in a number of places, so the optimization wreaked havoc
- with
- the code.
-
- I haven't encountered any such problems with 5.0.2 so far, but I am leary
- of the
- optimizations nonetheless.
-