decrement the recursion count.
Fixes the problem of uic dying with a no-next-thread assertion while
building kdeadmin3; reported by Perry Metzger and others on
tech-pkg.
that these calls be thread-safe with respect to the environment, and it
causes serious problems for threaded applications which call vfork() and
exec*() (including indirectly, via popen() or system()).
Acquire and release __environ_lock in the parent in popen() and system() to
play safe and provide the child with a stable environment.
__environ_lock should also have an atfork() handler; still under development.
A dynamically linked program invokes the rtld cleanup routine via
an atexit handler. This rtld cleanup routine invokes _fini() for
shared libraries, which in-turn invoke __cxa_finalize() with their
DSO handle. By luck, this happens to work okay for non-threaded
programs, but for a threaded program, this leads to deadlock (sometimes
manifested as an assertion failure, if the program didn't actually
create any threads).
Fixed by teaching __cxa_finalize() that it can be recursively invoked,
adjusting the handler list manipulation accordingly.
as specified by the C++ ABI for Itanium. Despite the ABI document's name,
many of the items in it are implemented for other architectures.
See the following URL for a reference:
http://www.codesourcery.com/cxx-abi/abi.html#dso-dtor
Fixes lib/18379 (from Joel Baker).
round has been tested on Solaris/x86 and Linux hosts.
* Add host tools cap_mkdb, ctags, m4, uudecode.
* Protect __RCSID() and __COPYRIGHT() better.
* Reduce the number of places that need to include "config.h", to keep
sources closer to their "vanilla" versions.
* Add more compat #defines and autoconf-checked functions.
since some "cleanups" in the master (UCB?) version of this file
were made in 1994 (which assumed the size of long is 4 bytes).
This fixes krb5-enabled logins on the Alpha.
* Don't return if wait()ing for the child is interrupted.
* Fix a race condition for the parent's SIG{INT,QUIT} disposition.
* While we're here, convert from using signal(3) to sigaction(2).
1. use uintptr_t instead of u_long
2. check for overflow in map_pages and malloc_pages
3. bring in fixes from FreeBSD [int -> size_t, and a missing THREAD_UNLOCK]
4. rewrite map_pages to use sbrk() only to grow memory (avoids extra syscall
and elides bug in brk(2) that ross is fixing)
5. restore the break point to its original value if the mmap(2) for the page
directory or the alignment sbrk breaks.
reviewed by: chuq and ross
tested by: make build and reboot
Now memtest nearly works; unfortunately there is no way currently to lower
the break point as we free, so memtest keeps trying to reduce memory when
mlock() fails and that does not work.
to <sys/types.h> and <sys/stdint.h>.
* Add a new C99 <stdint.h> header, which provides integer types of
explicit width, related limits and integer constant macros.
* Extend <inttypes.h> to provide <stdint.h> definitions and format
macros for printf() and scanf().
* Add C99 strtoimax() and strtoumax() functions.
* Use the latter within scanf().
* Add C99 %j, %t and %z printf()/scanf() conversions for
intmax_t, pointer-type and size_t arguments.
shot-term TODO:
- enable multibyte locales. this includes loadable locale modules
(#ifdef DLRUNE).
- populate LC_CTYPE definition files.
- other functions such as iconv()
- make libc functions more friendly with multibyte.
- multilingualize other LC_xx
- torture tests
XXX don't forget to "make depend" when you rebuild libc, multibyte.c is
moved from lib/libc/stdlib to lib/libc/locale.
to 0 and have getopt(3)/getopt_long(3) continue to function properly.
This happens to work with GNU getopt_long(3), so mimmick the behavior.
Fixes pkg/10980 (Manuel Bouyer): rsyncd was relying on this behavior.
When complaining about a long option which requires an argument,
print the actual option, not some pointer-garbage.
(Hint: "%c" and "%s" are different.)
actually does something, this can be expensive. From discussion on
current-users.
Note that the man page is already correct - we imported malloc.c from
FreeBSD before they made the same change but malloc.3 after...
In the function malloc_make_chunks, don't define the local variable ``l'',
which is used to hold the actual size of pginfo, as size_t. Define it as
a signed integer as before. We expect it to become zero or negative value,
and we know that it is much smaller than pagesize and it is small enough to
fit in signed interger.
(as originally declared in System V) to "char * const []", as was done in
1003.2 and subsequently propagated into XPG4.
From Andrew Cagney in PR lib/8930.
descriptors against -1 (as appropriate).
* add actual checks which to detect stuff that would trigger_DIAGASSERT(),
and attempt to return a sane error condition.
* knf some code
* remove some `register' decls.
the first two items result in the addition of code similar to the
following in various functions:
_DIAGASSERT(path != NULL)
#ifdef _DIAGNOSTIC
if (path == NULL) {
errno = EFAULT;
return (-1);
}
#endif
units of storage and returning a null pointer in System V mode; this was
broken by the `fix' in rev. 1.24. Also, as it is stated in ISO C that
such operation does not constitute an allocation failure, do not abort()
even if the `X' option is set.
Amusingly enough the SVID, Fourth Edition, specifies the `unique pointer'
return behaviour for this kind of allocation, so this is kind of mis-named.
(also from freebsd). comment out references to reallocf with .\"XXX"
for now. i also fixed some roff misuse as well.
replace calloc(3), free(3) and realloc(3) with links to malloc(3)
