SUN/IRAF V2.3 Installation, 12 June 1986 Sun-3, MC68881 floating point ------------------------------------------------------------------------------ /usr/iraf/local/notes Began this new notes file. All subsequent pathnames assume "/usr/iraf/" preceding them. (6/12) unix/ Made a tar backup in /tmp2/iraf/tuc_unix.arc. (6/12) local/ Made a tar backup in /tmp2/iraf/tuc_local.arc. (6/12) bin/ lib/ pkg/ sys/ unix/ Ran rmbin to remove all binary files. (6/12) /usr/iraf/ Unpacked new "wtar -o" source-only archive from lyra. (6/12) unix/as/ unix/mc68000 Replaced VAX contents of as/ with those from mc68000; renamed .SUN's to .s's. (6/12) unix/mkpkg.csh Replaced with the one from the archive; contains SUN_kludge mkpkg.csh command. (6/12) unix/hlib/[dir]1mach.f Replaced with those from archive for 68000; using MC68000 versions of machine-independent stuff for nspp software. (6/12) unix/hlib/mach.h Compared with archive and replaced with archive version to pick up BYTE_SWAP[24] = NO, and to use the 68000 fp epsilons. (6/12) unix/hlib/motd Short message about SUN/IRAF system testing. (6/12) unix/hlib/mkpkg.inc Ran diff; replaced from archive to pick up 68000 compiler and linker switches for this SUN-3, which has MC 68881 floating point (XFLAGS -c -O -f68881; LFLAGS -f68881). Set USE_CALCOMP to NO. (6/12) unix/hlib/mkiraf.csh Replaced iraf root definition. (6/12) unix/hlib/libc/iraf.h Changed root directory to /usr/iraf/, added "#define SUN3". (6/12) unix/hlib/libc/kernel.h Commented out DEBUGMEM so that the normal UNIX malloc() is used. This was and probably still is necessary in order to link with SUNCGI. (6/12) unix/hlib/libc/spp.h Ran diff; replaced with archive version to pick up MC68000 fp epsilons. (6/12) unix/boot/mkpkg/host.c Changed SZ_CMD from 512 to 400 to get around SUN-3 f77 command- line bug. (6/12) unix/boot/spp/xc.c Ran diff; replaced with archive version to pick up all the extra Fortran library definitions and command-line args. (6/12) unix/os/zxwhen.c Ran diff; replaced with archive version to pick up Sun floating point exception definitions. (6/12) local/notes Renamed to notes.vax. (6/12) local/login.cl Renamed root directory in "home"; aliased "xc" to "xc -f68881". (6/12) /usr/include/iraf.h -> $hlib/libc/iraf.h Made sure this symbolic link existed. (6/12) unix/ [BOOTSTRAP] Attempted bootstrap ("sh -x mkpkg.csh"). (6/12) unix/boot/bootlib/mkpkg.csh This time the bootstrap failed because of the "unexpected end of file" error and did NOT create libboot.a. Resolved by inserting "#! /bin/csh -f" as the first line. (6/13) unix/ [BOOTSTRAP] Successful bootstrap this time. (6/13) unix/hlib/alloc.e Changed ownership to root in order to allocate devices at runtime. (6/13) /usr/bin/alloc -> $hlib/alloc.e /usr/bin/generic -> $hlib/generic.e /usr/bin/mkpkg -> $hlib/mkpkg.e /usr/bin/rmbin -> $hlib/rmbin.e /usr/bin/rmfiles -> $hlib/rmfiles.e /usr/bin/rtar -> $hlib/rtar.e /usr/bin/wtar -> $hlib/wtar.e /usr/bin/xc -> $hlib/xc.e /usr/bin/cl -> $hlib/cl.e Made sure these symbolic links existed. (6/13) sys/fmtio/gctod.f F77 compile-time sys/fmtio/xtoc.f F77 compile-time sys/vops/ak/abeqkx.f F77 compile-time sys/vops/ak/abeqx.f F77 compile-time sys/vops/ak/abgekx.f F77 compile-time sys/vops/ak/abgtkx.f F77 compile-time sys/vops/ak/ablekx.f F77 compile-time sys/vops/ak/abltkx.f F77 compile-time sys/vops/ak/abnekx.f F77 compile-time sys/vops/ak/abnex.f F77 compile-time sys/vops/ak/advzx.f F77 compile-time sys/vops/lz/allnx.f F77 compile-time sys/vops/lz/alogx.f F77 compile-time sys/vops/lz/arcpx.f F77 compile-time sys/vops/lz/arczx.f F77 compile-time Hand-compiled without "-f68881" switch (but with optimization); otherwise the compiler fails due to complex datatype expressions. This will be the case until the next Sun f77 release. (6/13) sys/fmtio/fprfmt.x F77 run-time Hand-compiled without optimization (but with "-f68881"). Have not yet determined source of problem (symptom was garbage appearing in text files) -- pursue shortly. (6/13) $iraf Started sysgen at 10:22. (6/13) dev/devices Replaced with version from suntest!, since we now have a 1/2 inch tape drive on tucana. Added mtb entry for the 1/4 inch drive. (6/13) noao$onedspec/getnimage.x F77 run-time There is a bug in the optimizer which seems to be caused by a logical comparison to MAX_INT. To get around the problem do the following after a full sysgen. (6/19) touch getnimage.x; mkpkg xflags=-c fluxcal; mkpkg install $iraf Unpacked an incremental archive from lyra to bring tucana up to current as of 09:15 June 20, 1986. Performed an incremental sysgen. (6/20) noao/twodspec/apextract/* The preceding MKPKG found the entire contents of icfit/* missing except for 3 source files. To be safe, I (sr) made a complete archive of lyra!/...apextract/* and unpacked it here, followed by another root sysgen. (6/20) noao/twodspec/apexold/* Because this directory was created by renaming the directory apextract after the full archive of June 12, it had old file dates and the incremental archives this week didn't pick it up. Unpacked a full archive and MKPKG'd. (6/20) sys/gio/cursor/grcwcs.x F77 run-time Apparent f77 optimizer bug; hand-compiling w/o optimization and relinking the CL clears up the problem of IMPLOT's "l" and "c" commands returning bad pixel rows/columns. Before, clgcur was sometimes returning the y coordinate in pixel, rather than data, units -- i.e. in WCS 1 rather than 2. Haven't had time to try to isolate the source of the optimizer bug. (6/21) $iraf Unpacked an incremental archive from lyra! current to 14:40 June 21; performed an incremental sysgen. (6/21) noao/mtlocal/r2df/r2df.com The preceding sysgen died because this file was missing; it had a date earlier than the incremental archive, so was missed. Added it and rebuilt mtlocal. (6/21) $iraf Unpacked an incremental archive from lyra! current to 08:15 June 23; performed an incremental sysgen. (6/23) noao/lib/scr/* noao/twodspec/apextract/* noao/twodspec/apexold/* Unpacked full, not incremental, archives from lyra! to pick up any files missed during the incremental archives as a result of renaming rather than copying files since the June 12 full archive. mkpkg updates in both apextract directories. (6/23) pkg/cl/unop.c Changed call to sscanf (sval, "%f", &rresult) in case OP_REAL to sscanf (sval, "%lf", &rresult) because "rresult" is declared type "double", and on the SUNs, sscanf was only writing into the first four bytes of rresult, giving erroneous answers to REAL coercion, as in "cl> =real(imgets.value)". Should be changed on master system as well. (6/23)