hansb@aie.uucp (Hans Bayle) (03/26/91)
Is there any place from where I can ftp g++, glib++, gcc, and gas source code configured to generate COFF objects for Interactive 386/ix 2.2? I've tried the COFF patches from David Dawes at syphus.physics.su.oz.au, but Interactive's ld cant handle the object files that are generated by gas. I've been looking in the code, and it seems that Interactive's aouthdr.h is different from the a.out.h that came with the Esix COFF patches for gas. Hans Bayle, A.I. Engineering. UUCP: hp4nl.nluug.nl!aie!hansb Internet: bayle@swi.psy.uva.nl
jr@twisted.dkw.com (J.R. Jesson) (03/28/91)
In article <1991Mar25.213918.13974@aie.uucp>, hansb@aie.uucp (Hans Bayle) writes: |> |> Is there any place from where I can ftp g++, glib++, gcc, and gas source code |> configured to generate COFF objects for Interactive 386/ix 2.2? |> |> I've tried the COFF patches from David Dawes at syphus.physics.su.oz.au, |> but Interactive's ld cant handle the object files that are generated by gas. |> I've been looking in the code, and it seems that Interactive's aouthdr.h |> is different from the a.out.h that came with the Esix COFF patches for gas. I would recommend that you blow off "COFF" compatability when compiling gnu under ISC Unix. I have gcc, gdb and g++ running, and got them running simply by specifying the standard 386-sysv entry when configuring. In essence, because gcc and g++ use the stock ISC assembler, you get COFF but not encapsulated COFF. Although you dont get some of the extended COFF stuff (and dbx records), it's probably not much of a loss either, since you would have to duplicate all of the system libraries you like to use (ala robbitussin), and be bound to ranlib (isnt ranlib kind of a pain in the ass?). J.R.