Return to BSD News archive
Path: sserve!manuel!munnari.oz.au!spool.mu.edu!caen!zaphod.mps.ohio-state.edu!usc!wupost!darwin.sura.net!jvnc.net!nuscc!ntuix!eoahmad From: eoahmad@ntuix.ntu.ac.sg (Othman Ahmad) Newsgroups: comp.unix.bsd Subject: Re: 386BSD: Anyone using gcc-2.x on 0.1? Message-ID: <1992Jul30.064120.5453@ntuix.ntu.ac.sg> Date: 30 Jul 92 06:41:20 GMT Article-I.D.: ntuix.1992Jul30.064120.5453 References: <1992Jul29.155631.21213@thyme.jpl.nasa.gov> Organization: Nanyang Technological University - Singapore Lines: 41 In article <1992Jul29.155631.21213@thyme.jpl.nasa.gov> kaleb@thyme.jpl.nasa.gov (Kaleb Keithley) writes: : While working my way around the system, I discovered that the version : of gcc shipped in 0.1 is 1.39, at least as reported by 'gcc -v'. : : I'm highly interested in moving to the latest version of gcc-2.x, : (currently 2.2.2) because: : : a) I'd like to take advantage of it's ability to emit 486 specific : code (instruction scheduling, primarily) : : b) Its ability to emit PIC (Position Independent Code) seems to be : a pre-requisite for shared libraries. : : Has anyone else "bootstrapped" up to 2.2.2 yet? Has anyone who I did for 0.0 and it is very easy. Just use stddef.h from 386BSD instead of gstddef.h. Also use tm.h from 0.0 gcc 1.39 but with some modifications. : has, rebuilt the kernel with it yet? Does anyone have any reasons I tried but there are some bugs with my version of gcc 2.2.2. I must have done something wrong because my parser was version 2.1 but it managed to generate executable code because the code-generation parts are from 2.2.2. : why this might not be a good idea? gcc 2.2.2 I suspect is not 100% compatible with gcc-1XX. : : I'll be trying it anyway -- I've got nothing to loose, the worst thing : that could happen is I'd have to reload a few disks :-) It is easy but needs more work. I have no longer keen because I discovered bugs in gcc-2.2.2 when porting it to Dec ultrix with the new libraries. The latest version of gcc is not as stable as gcc-1.40. However you can get 30% speedup but no code size difference for a 386 machine. Good luck. -- Othman bin Ahmad, School of EEE, Nanyang Technological University, Singapore 2263. Internet Email: eoahmad@ntuix.ntu.ac.sg Bitnet Email: eoahmad@ntuvax.bitnet