summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
1999-05-26compile i386 ramdisk with -Os (conserve size)Marc Espie
1999-05-26Slight tweaks to account for the new cpp driver.Marc Espie
1999-05-26New egcs cpp driver does always define gnu C.Marc Espie
1999-05-26Link in egcs, kill gcc and libg++Marc Espie
1999-05-26Wrappers + lib versionsMarc Espie
1999-05-26Add x-openbsd fragment and reconfigure accordingly.Marc Espie
1999-05-26Make overriding libiberty's location possible.Marc Espie
1999-05-26Fragment for sysv4.hMarc Espie
1999-05-26Configuration fragments for arches that have not been committed to egcsMarc Espie
main repository yet (rs6000 to come)
1999-05-26Remove bison/yacc output files and change Makefiles rules to regenerateMarc Espie
them.
1999-05-26Work-around: egcs ICE on m68kMarc Espie
1999-05-26Work-around: a few files trigger Internal Compiler Errors if compiled withMarc Espie
-O2 on m68k.
1999-05-26Tweak for egcs -fwrite-strings oddityMarc Espie
1999-05-26egcs projects compiler systemMarc Espie
Exact copy of the snapshot, except for the removal of texinfo/ gcc/ch/ libchill/
1999-05-26Configure raid devices at boot (from NetBSD, ok by niklas@).Jakob Schlyter
1999-05-26need string.hTheo de Raadt
1999-05-26need string.hTheo de Raadt
1999-05-26back out sigpending hideDavid Leonard
1999-05-26forgot to hide sigpendingDavid Leonard
1999-05-26arguments to -perm option are `[-]mode', not `[-mode]'; from NetBSDAaron Campbell
1999-05-26Xr st(4); from NetBSDAaron Campbell
1999-05-26mention that parameter name1 (the file to link to) need not exist; from NetBSDAaron Campbell
1999-05-26typo; from NetBSDAaron Campbell
1999-05-26new 3Com NIC man pages (aka, work no one else wants to do) :-)Aaron Campbell
1999-05-26sync with FreeBSDDavid Leonard
1999-05-26document two new states (deadlock and spinblock)David Leonard
1999-05-26add mb (memory barrier) instruction. from FreeBSDDavid Leonard
1999-05-26document double-unlock check and how it relates to the std. (POSIX says in ↵David Leonard
one paragraph that the behaviour is undefined, then in another says that EPERM should be returned
1999-05-26bye bye gethostbyname_r and getservbyname_r testsDavid Leonard
1999-05-26count and chec the number of times the self-cancelling thread tries to ↵David Leonard
cancel itself
1999-05-26show a bit more information when checks failDavid Leonard
1999-05-25explain the third argument more thoroughly; some cmds take it as an int,Aaron Campbell
others a pointer to a struct flock (record locking), and the rest ignore it
1999-05-25Export kn_keycompare() in the API, update the spec and man pages.Angelos D. Keromytis
Note that the *.1 manpages will all be rolled into a single one soon.
1999-05-25kill some old macrosAaron Campbell
1999-05-25Stub out src/ ports makefile fragments, so that everything is underMarc Espie
ports/ (simpler to keep synchronized).
1999-05-25syncTheo de Raadt
1999-05-25for getopt, check rval != -1, not > 0Todd C. Miller
1999-05-25punctuationAaron Campbell
1999-05-25Move hilsoftinit() call to before interrupts are enabled, from NetBSD.Jason Downs
1999-05-25also explain about makeoptions PROF="-pg".Niels Provos
1999-05-25explain about option GPROF.Niels Provos
1999-05-25all functions need a return type and at least one arg, even if it's voidAaron Campbell
1999-05-25add .Ft macros and move the argument type into each function definitionAaron Campbell
1999-05-25whoops, tzset() needs a void argument; deraadt@Aaron Campbell
1999-05-25out with old macros, in with new mdoc; kwesterback@home.comAaron Campbell
1999-05-24new mdoc ipftest(1) man page; kwesterback@home.comAaron Campbell
1999-05-24This patch lets sparc ld handle pic/PIC relocations mix, by simply forcingMarc Espie
everything to fit into pic mode. More extensive changes (like tagging relocs with pic/PIC, sorting, and putting pic nearest the beginning of the GOT) would be needed for full handling pic relocs with a sizeable number of PIC relocs.
1999-05-24Only do basic work in the ethernet interrupt context, and queue packets toJason Wright
be bridged. Do the real work in a scheduled netisr.
1999-05-24set the close-on-exec flag for newly opened filesDavid Leonard
Reason: vi uses flock to help you notice when you are already editing a file. But, when using :shell or :! the fdesc for the locked file is inherited by the subprocess. This means that if you (say) do a :!xterm& and then quit vi, the fdesc is kept open by the xterm and the lock is never released. future vi's on that file will then complain. I sent this off to bostic.com a long time ago but have as yet heard nothing.
1999-05-24show '(not a tty)' instead of '(null)' when wall is run with stderr not a ↵David Leonard
tty. (e.g. when run from cron)