| Commit | Line | Data |
|---|
| ea8b5122 |
1 | # This file has been put together by Anno Siegel <siegel@zrz.TU-Berlin.DE> |
| 2 | # and Andreas Koenig <k@franz.ww.TU-Berlin.DE>. Comments, questions, and |
| 3 | # improvements welcome! |
| 4 | # |
| 5 | |
| 6 | # These hints are intended for NeXT 3.3. If you're running the 3.3 |
| 7 | # "user" version of the NeXT OS, you should not change the malloc |
| 8 | # related hints (USE_PERL_SBRK, HIDEMYMALLOC, usemymalloc). If you're |
| 9 | # running the 3.3 "dev" version of the OS, I do not know what to |
| 10 | # recommend (I have no 3.3 dev). |
| 11 | |
| 12 | # From about perl5.002beta1h perl became unstable on the |
| 13 | # NeXT. Intermittent coredumps were frequent on 3.2 OS. There were |
| 14 | # reports, that the developer version of 3.3 didn't have problems, so it |
| 15 | # seemed pretty obvious that we had to work around an malloc bug in 3.2. |
| 16 | # This hints file reflects a patch to perl5.002_01 that introduces a |
| 17 | # home made sbrk routine (remember, NeXT's sbrk _never_ worked). This |
| 18 | # sbrk makes it possible to run perl with its own malloc. Thanks to |
| 19 | # Ilya who showed me the way to his sbrk for OS/2!! |
| 20 | # andreas koenig, 1996-06-16 |
| 21 | |
| 22 | ccflags='-DUSE_NEXT_CTYPE -DUSE_PERL_SBRK -DHIDEMYMALLOC' |
| 23 | POSIX_cflags='ccflags="-posix $ccflags"' |
| 24 | ldflags='-u libsys_s' |
| 25 | libswanted='dbm gdbm db' |
| 26 | |
| 27 | lddlflags='-r' |
| 28 | # Give cccdlflags an empty value since Configure will detect we are |
| 29 | # using GNU cc and try to specify -fpic for cccdlflags. |
| 30 | cccdlflags=' ' |
| 31 | |
| 32 | i_utime='undef' |
| 33 | groupstype='int' |
| 34 | direntrytype='struct direct' |
| 35 | d_strcoll='undef' |
| 36 | |
| 37 | # the simple program `for ($i=1;$i<38771;$i++){$t{$i}=123}' fails |
| 38 | # with Larry's malloc on NS 3.2 due to broken sbrk() |
| 39 | ###################################################################### |
| 40 | # above comment should stay here, but is not longer of importance # |
| 41 | # with -DUSE_PERL_SBRK and -DHIDEMYMALLOC we can now say 'yes' to # |
| 42 | # usemymalloc. We call this hintsfile next_3_2.sh, so folks with 3.3 # |
| 43 | # can decide what they prefer. Actually folks with 3.3 "user" version# |
| 44 | # will also need this hintsfile, but how can I discern which 3.3 it # |
| 45 | # is? # |
| 46 | ###################################################################### |
| 47 | usemymalloc='y' |
| 48 | |
| 49 | d_uname='define' |
| 50 | d_setpgid='define' |
| 51 | d_setsid='define' |
| 52 | d_tcgetpgrp='define' |
| 53 | d_tcsetpgrp='define' |
| 54 | |
| 55 | # |
| 56 | # On some NeXT machines, the timestamp put by ranlib is not correct, and |
| 57 | # this may cause useless recompiles. Fix that by adding a sleep before |
| 58 | # running ranlib. The '5' is an empirical number that's "long enough." |
| 59 | # |
| 60 | ranlib='sleep 5; /bin/ranlib' |
| 61 | |
| 62 | |
| 63 | # |
| 64 | # There where reports that the compiler on HPPA machines |
| 65 | # fails with the -O flag on pp.c. |
| 66 | # |
| 67 | if [ `arch` = "hppa" ]; then |
| 68 | pp_cflags='optimize="-g"' |
| 69 | fi |
projects / p5sagit/p5-mst-13.2.git / blame