home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cs.utexas.edu!ut-emx!ibmchs!auschs!awdprime.austin.ibm.com!scott
- From: scott@austin.ibm.com
- Newsgroups: comp.unix.aix
- Subject: Re: Buggy malloc workaround
- Message-ID: <1992Sep3.175326.18097@awdprime.austin.ibm.com>
- Date: 3 Sep 92 17:53:26 GMT
- References: <Btx211.2DD@watdragon.uwaterloo.ca>
- Sender: news@awdprime.austin.ibm.com (USENET News)
- Organization: IBM Austin
- Lines: 20
- Originator: scott@porter.austin.ibm.com
-
-
- In article <Btx211.2DD@watdragon.uwaterloo.ca>, mskucher@watdragon.uwaterloo.ca (Murray S. Kucherawy [MFCF]) writes:
- > My problems with putuserattr() calls are, according to IBM, caused by a bug
- > in malloc implemented in AIX V3.2. According to the man page for malloc,
- > you can set the environment variable MALLOCTYPE to "3.1" and it will
- > use the older version of malloc.
- >
- > It's not clear whether this is to be done at runtime or compile time.
- > Anyone know? My program is to be run by a shell script remotely, and it will
- > be tough to ensure that the environment variable is set.
-
- I'm sure what they meant to say was that putuserattr() has a bug in its use
- of malloc which is now evident when using the AIXv3.2 malloc. The AIXv3.2
- malloc does a better job of reusing storage and so programs that don't
- initialize storage correctly may break when they worked OK on AIXv3.1.
-
- The MALLOCTYPE environment variable is used at runtime.
-
- Scott L. Porter IBM PSP Austin / AIX Kernel Development
- Internet: scott@glasnost.austin.ibm.com
-