Commit Graph

156 Commits

Author SHA1 Message Date
tg af606537e2 if we don't have rlim_t, assume it's long 2007-01-17 23:18:55 +00:00
tg c3de1d7564 <sys/sysmacros.h> contains major/minor/makedev on Linux 2.0 2007-01-17 23:04:19 +00:00
tg 4c82f74d2b new option -nd to force static linking of mksh but not mirtoconf tests 2007-01-17 22:59:25 +00:00
tg da2f916010 remove -x (cross-build) option, use -d instead, it does the same
side effect: TARGET_OS is now honoured without -x too
2007-01-17 22:55:47 +00:00
tg ddd2dac47d * support old environments without libgen.h (ancient GNU/Linux)
and stdbool.h (ancient GNU/Linux; NetBSD® 1.6.1)
* __dead must come after, not before, to accomodate gcc 2.7.2.3
2007-01-17 22:51:47 +00:00
tg 8fea7398d3 we need "$CC -E -" as cpp, not "$CC -E"
found on
Linux ecce 2.0.38 #1 Wed Jul 26 22:05:46 2000 i686 unknown
Reading specs from /usr/lib/gcc-lib/i386-linux/2.7.2.3/specs
gcc version 2.7.2.3

Note: during the entire compile, there are warnings like
| gcc: unrecognized option `-std=gnu99'
but these won't trigger -Werror since they are in the driver,
not in the front-end, so they are harmless, I guess. (I don't
know of a way to make them vanish, either.)
2007-01-17 22:35:09 +00:00
tg ceb5a7dba0 check if __RCSID() can be used multiple times; req'd eg. on Mac 2007-01-17 21:42:23 +00:00
tg 42783fdfd9 only call nroff to see if it's gnroff if -r is not given 2007-01-17 17:57:31 +00:00
tg f077c0ab5a list all "known good" OSes in the $TARGET_OS switch,
echo a warning (but continue) for unknown OSes, with
the desire the porter should send in diffs
2007-01-17 17:46:18 +00:00
tg 8620379fcc * add #ifdef MKSH_NOPWNAM to prevent pulling in getpwnam()
* document possible "functionality omission defines" in Build.sh
* sort and sectionise mirtoconf checks; add big comments
2007-01-17 17:42:22 +00:00
tg c231ab8cc4 * order tests: we want to use correct NOWARN as much as possible
* "$e" and "echo" fixes for if the stuff begins with a -, and
  quoting cleanup (where feasible, feel free to send diffs)
* cosmetics

I hereby name this baby "Mirtoconf".
2007-01-17 17:31:59 +00:00
tg f946d405f8 if both "$CPP" (from parent) and "$CC -E" fail, try "cpp" (from $PATH) and
"/usr/libexec/cpp" and "/lib/cpp" last
2007-01-17 17:14:25 +00:00
tg 8211416e5a make a loop (removes dup code) 2007-01-17 17:12:43 +00:00
tg 478374ba3b autoscan for c præprocessor now prefers $CPP to "$CC -E" 2007-01-17 17:08:34 +00:00
tg da03ad5be2 autoconf for -O2 (if CFLAGS is _empty_), -Wall, -fwrapv, -fno-strict-aliasing
XXX untested
2007-01-17 17:03:59 +00:00
tg d4b4ca5e92 simplify 2007-01-17 16:57:41 +00:00
tg e19a99efaf more creative use of ac_flags() 2007-01-17 16:52:00 +00:00
tg f6dbde5115 thinko 2007-01-17 16:44:19 +00:00
tg bcb0852dcc new shell function for autoconfing compiler flags, reduces redundancy 2007-01-17 16:39:55 +00:00
tg 69584be3bb add two TODOs 2007-01-17 01:59:38 +00:00
tg 263f518dfd * Build.sh: add -fwrapv to standard CFLAGS, just to be on the safe
side (I don't have capacities to scan 71711 files in MirOS for a
  standards-compliance whack)
* copyright: mention CVS changelogs as place of contributors' name
  and credit information, too
* check.t, sh.h: bump
2007-01-17 01:24:29 +00:00
tg a00fbd3eb7 autoconf for -std=gnu99 (not c99 because of brokenness in the field)
from ankon -> solaris 10 with gcc 3.4.6 from sunfreeware
2007-01-14 21:38:25 +00:00
tg 03e8f0ec83 CPP is honoured now as well (but only needed for the siglist) 2007-01-12 03:30:40 +00:00
tg 9c888e103a oops… really unbreak -d 2007-01-12 03:20:07 +00:00
tg 22538dcf0d unbreak -d 2007-01-12 03:17:08 +00:00
tg e9d944e825 user should be able to override if he wants setlocale calls in his
extra small mksh…
2007-01-12 02:46:27 +00:00
tg ca7a9c3eb0 ] scn.c:1: warning: ISO C forbids an empty source file
yeaaah, we #error out anyway, but…
2007-01-12 02:40:18 +00:00
tg e768017c58 gcc 4.1.2 pedantic warning free configure 2007-01-12 02:37:32 +00:00
tg 2c46d96042 * instead of setting LDSTATIC='' for OSes which did not support it
until now, set it to '@'
* if LDSTATIC is '@' make it '-static' if MKSH_SMALL, '' otherwise

Yep, this might break, e.g. Darwin or Cygwin. But let's test that.
2007-01-12 02:34:46 +00:00
tg 21e2397c1b disallow the user to specify (overload) SRCS, if there is special
target platform specific crap, it belongs into LIBS instead
2007-01-12 02:27:40 +00:00
tg 6e5437f86b interesting way to break things… oeps ☺ 2007-01-12 02:12:16 +00:00
tg 552aebd24f more verbose 2007-01-12 02:09:10 +00:00
tg 80393db1b6 fix configuration on solaris 8 with whitespaces in both curdir and
srcdir, mostly due to 'eval' issues and a picky /bin/sh (yeaaah…)
2007-01-12 02:01:48 +00:00
tg 5891915f1b * Scan for __attribute__((...)) in general (the earliest was 2.5,
where we had 'noreturn' etc. but no '__noreturn__')
* Scan for __attribute__((bounded)) and __attribute__((used))
  if we have __attribute__((noreturn))
* To be able to scan if certain attributes give warnings,
  scan for -Werror with a simple programme which hopefully triggers none
* Convert __attribute__((unused)) to __unused, noreturn -> __dead
* Unify other attributes
* Clean up typography a little more
2007-01-12 01:49:29 +00:00
tg 70eb6d19c2 solaris /bin/sh:
| v() { ... }
| v=1
undefines v()
2007-01-12 01:44:32 +00:00
tg 636ef70a24 underline actual autoconf results (test names are in bold)
FWIW, works ATM, even with spaces in directory name
2007-01-12 01:32:27 +00:00
tg 6fb4af0a9d * typo
* include path
2007-01-12 01:27:28 +00:00
tg 5309e33974 implement colouring of "autoconf" output. yay!
NB: only if stdout isatty
2007-01-12 01:17:10 +00:00
tg 2f15a11c55 Clean up the signal mess, saves 172 Bytes:
* 'sigseen' in Build.sh goes away
* Signal name existence is checked in this order:
  have our own¹ -> sys_signame[] -> _sys_signame[] -> build our own²
* Signal description existence is checked in this order:
  sys_siglist[] -> _sys_siglist[] -> strsignal() -> NULL³
¹ Predefined list of items, for operating systems where we
  cannot build² them, i.e. Plan 9 and Minix 3 (e.g. no $CPP -dD)
² The usual cpp(1) stuff
³ Changed later, see below
* Make $CPP test dependent on $NEED_MKSH_SIGNAME (others can
  be added here, this is not absolute)
* Make signal name list generation² dependent on $NEED_MKSH_SIGNAME
* Fix check if the generation worked
* Guarantee that sigtraps[*].name and sigtraps[*].mess are valid
  C strings; this makes the code shorter *and* removes a few pos-
  sible nil pointer dereferences
* Embed autoconf'd usages of sys_sig* / strsignal / mksh_sigpairs
  into inittraps()
* Check for each signal 0<=i<=NSIG that
  name is not NULL or "" -> replace with ("%d", i)
  mess is not NULL or "" -> replace with ("Signal %d", i)
  name does not start (case-insensitive) with "SIG" -> name += 3
* In gettrap(), fix check if signal name starts, case-sensitive
  or case-insensitive, depending on need, with "SIG" (bug from millert@)

Other changes:
* Build.sh: ac_test[n]() are documented
* Build.sh: ac_test[n]() can have negative prereqs as well now
* Build.sh: use <<-'EOF' consistently
* bump patchlevel to today
2007-01-12 00:25:40 +00:00
tg 3b63b986c3 * pasto (patcho?)
* improve output readability

I wonder if I should use ANSI escapes to make the results
from the configuration bold… but then, this'd look worse
in e.g. mc or less. Suggestions?
2007-01-11 03:03:34 +00:00
tg 3b6b37c671 make use of the new $CPP variable (and enable the cpp check if needed)
@Benny:	this is why I don't use GNU autoconf: writing configure.in
	files is, supposedly, easy - but nobody teaches you which
	changes you have to apply to your source files. Here I know.
2007-01-11 02:41:53 +00:00
tg 4d2624fd1d add test how to run the C preprocessor and fix the $NOWARN check and docs 2007-01-11 02:37:40 +00:00
tg 546e97621a small mkshs also shouldn't pull in the entire locale package,
so don't even call setlocale, lest alone nl_langinfo (already
disabled before)
2007-01-11 02:11:46 +00:00
tg e269c4ed29 if building a small mksh, don't check for persistent history,
since it's not supported anyway; makes regression tests pass there
2007-01-11 02:08:50 +00:00
tg f3168d3e9d introduce TARGET_OS, for crossbuilds, contains uname -s output 2007-01-11 00:57:56 +00:00
tg 8bfc693178 this double-backtick in double-backtick (SunOS /bin/sh) stuff
looked waay too weird to me, use a temporary variable instead
and make the two occurencies of the code look more alike
2007-01-11 00:47:40 +00:00
tg 8dc1a064fa use the same matching for NSIG as for the others - cut off
everything which doesn't look like a dec/hex/oct number
2007-01-11 00:44:08 +00:00
tg c822cbedd8 oops, match hex numbers; seen while looking at ankon's (FreeBSD) problem 2007-01-11 00:38:37 +00:00
tg a080fc385d add a TODO here 2006-12-11 19:58:37 +00:00
tg 2122e253e7 fix bitching of: gcc version 3.3.5 (Debian 1:3.3.5-13) 2006-11-21 21:45:24 +00:00