home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!wupost!tulane!uflorida!sinkhole.unf.edu!shite
- From: shite@sinkhole.unf.edu (Stephen Hite)
- Newsgroups: comp.unix.sysv386
- Subject: Re: Help: g++ & V.4
- Keywords: g++ c++ V.4 gnu assembler 386
- Message-ID: <1992Jul30.022812.21678@sinkhole.unf.edu>
- Date: 30 Jul 92 02:28:12 GMT
- References: <bilse.712010793@login.dkuug.dk>
- Organization: University of North Florida, Jacksonville
- Lines: 22
-
- In article <bilse.712010793@login.dkuug.dk> bilse@login.dkuug.dk (Per G. Bilse) writes:
- >I have a V.4 system (Consensys, to be specific) and have problems
- >with the assembler output from g++. For constructors and destructors,
- >g++ outputs special sections ...
- >
- > .section .dtors,"a",@progbits .long <address>
- >
- >... (and with .ctors for constructors). However, the assembler distinctly
- >dislikes anything following the ``"a"''. I've tried unsuccessfully
- >to fiddle about with the output from g++, trying to use .rodata etc,
- >but with no success. Now, the question is if the assembler on
- >Consensys is out of line, or if the method of handling C++ information
- >is different on '386 V.4 from that on other V.4s.
- >
-
- This was a problem in an older version of g++ (v2.0 I think). Get
- v2.2.2 and you should be okay. The problem is caused by a missing
- newline in 2 instances of an fprintf in one of the header files (forget
- which one).
-
- Steve Hite
- shite@sinkhole.unf.edu
-