alan@metasoft.UUCP (Alan Epstein) (04/18/89)
there has been some traffic here about problems with gcc 1.32 and 1.33 with respect to X. most of the talk has surrounded use of the purdue speedups. with gcc 1.34 available, can one safely use this latest version to compile X and X applications? what caveats exist, if any? thanks. ----------------------------- Alan Epstein Meta Software Corp UUCP: ...bbn!metasoft!alan 150 Cambridgepark Dr Internet/ARPA: alan%metasoft@bbn.com Cambridge, MA 02140 USA -----------------------------
rusty@garnet.berkeley.edu (04/18/89)
I've been using gcc 1.34 with x11r3 and the Perdoo fixes with good success. For a Sun 3/50 running SunOS 3.5.2 the ServerDefines in Sun.macros is #define ServerGCCDefines -finline-functions -fcombine-regs #define ServerDefines SunWindowsDefines ServerGCCDefines -DPURDUE -DNO_3_60_CG4 and site.def contains #define OptimizedCDebugFlags -O -traditional -fwritable-strings -fstrength-reduce For a MicroVax running Ultrix 3.0 ServerDefines in Vax.macros is #define ServerDefines -DPURDUE and site.def contains #define OptimizedCDebugFlags -O -traditional -fwritable-strings -fstrength-reduce -Dultrix On the MicroVax, for the qdss display, the files tldma.c and qdcolor.c must be compiled with -fvolatile because they use pointers to i/o registers. There might be similar problems with the qvss but I don't know. The notes for gcc 1.34 are wrong; the -fpcc-struct-return flag doesn't do what it claims; compile oscolor.c with /bin/cc. I vaguely remember trying to use the Purdue patches to dbm.h but had problems and resorted to the path of least resistance (/bin/cc). For the clients I didn't have any major problems. The file Login.c in the xdm directory is missing a semicolon in one of the structs. Likewise for one of the files in the xperfmon directory. I didn't compile everything in the contrib directory so there may be others. -------------------------------------- rusty c. wright rusty@violet.berkeley.edu ucbvax!violet!rusty