summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2003-12-17_ZLIB_PRIVATE is not used anywhereTodd C. Miller
2003-12-16Add #ifdef SMALL and #ifdef SLOW and include sys/unistd.h instead of unistd.hTodd C. Miller
2003-12-16Update to zlib 1.2.1Todd C. Miller
2003-12-16disabling the gzip header handling that we do ourselves anyway gains usHenning Brauer
1532 more bytes
2003-12-16millert@ found about NOBYFOUR in crc32.c while looking at sys/lib/libz/Henning Brauer
which has teh effect of disabling some new code so that the resulting object file is the exact same size as the one from our local, older copy. get rid of it.
2003-12-16not usedHenning Brauer
2003-12-16do not implement -L (print copyright) with SMALLHenning Brauer
186 bytes more bytes we can waste elsewhere
2003-12-16Remove old definition of crc_table that conflicts with new one whenTodd C. Miller
DYNAMIC_CRC_TABLE is defined. OK henning@
2003-12-16really don't implement unzip #ifdef SMALLHenning Brauer
this was in the tested diffs but slipped through when committing unfortunately, sorry
2003-12-16really don't do -V #ifdef SMALLHenning Brauer
2003-12-16a bit more shrinking. do not bother printing the drive bus interfaces,Theo de Raadt
saves more space.
2003-12-16syncTheo de Raadt
2003-12-16do not include rcsids with -DSMALL and as it is meaningless then do notHenning Brauer
implement -V then. following a discussion with millert@
2003-12-16no more libzHenning Brauer
2003-12-16space saver for the boot floppies:Henning Brauer
on the install media, gzip only need to decompress, but never to compress. so leave the (gz-)compress code out. don't link libz but compile the needed parts in. use an older much smaller crc32.c version and a cut down compress.c tested by nick@ ok deraadt@
2003-12-16when we read the original file size from the gzp header in uncompress modeHenning Brauer
and compare it to total_out to detect I/O errors we need to cast total_out to u_int32_t as the gzip header file is limited to that and total_out is now 64 bits. ok millert@
2003-12-16bump majorHenning Brauer
ok millert@ (previous as well)
2003-12-16total_in and total_out need to be off_t, not unsigned long.Henning Brauer
some bugs return: i fixed the same some months ago when we had this other gzip there. this bug resulted in wrong size stats for > 4GB files, and in the case that the input file was > 4GB and could be compressed to < 4GB gzip not zipping it as it would grow in its eyes.
2003-12-16update to zlib 1.2.1Henning Brauer
ok millert@ deraadt@
2003-12-16Better fix for MALLOC_OPTIONS=J bug. Something had problems withTodd C. Miller
returning a NULL label so now we just the epoch for the timestamp if we cannot stat the file. OK deraadt@
2003-12-16for -DSMALL, do not include rcsid[]sTheo de Raadt
2003-12-16when a bridge filter rule specifies both src and dst mac address, we do notHenning Brauer
want to compare both against the packet's source address. works much better when we compare the dst address to the packet's dst address. ok camield@ canacar@ markus@ jason@
2003-12-16Add SLIST_FOREACH_PREVPTR and SLIST_REMOVE_NEXT. SLIST_FOREACH_PREVPTRTodd C. Miller
is like SLIST_FOREACH but it saves a pointer to the previous element. SLIST_REMOVE_NEXT will remove the element *after* the one passed in. SLIST_FOREACH_PREVPTR is from FreeBSD; SLIST_REMOVE_NEXT was suggested by canacar@; man page additions by yours truly. OK deraadt@ (grudgingly) and man page changes OK jmc@.
2003-12-16reenable sti1 as the issues has been resolvedMichael Shalayeff
2003-12-16return error in ifc_destroy; ok deraadt, itojun, cedric, hshoexerMarkus Friedl
2003-12-16bad free - spotted by vincent@Bob Beck
2003-12-16afs cleanup, realloc strings, abort, etc.Bob Beck
testing by beck, hin, jose, fries. ok deraadt@
2003-12-16uncessary defnsJason Wright
2003-12-16Avoid segmentation fault when empty commands are specified in theOtto Moerbeek
config file; resolves PR 3578. ok deraadt@ millert@
2003-12-16bad printf; leopold-bloom@foni.netTheo de Raadt
2003-12-16Support MANPAGER environment variable plus some KNF I had in my tree.Todd C. Miller
Closes PR 3563
2003-12-16avoid name clash with newer libz; breaks on static archsHenning Brauer
ok millert@
2003-12-16use full sentence for description of cols; from Jared Yanovich;Jason McIntyre
2003-12-16quotes, spaces and bars do not need to be escaped within displays;Jason McIntyre
.Nm does not need an argument; dashes should be escaped;
2003-12-16use <= instead of < in dh_estimate; ok provos/hshoexer; do not return < ↵Markus Friedl
DH_GRP_MIN
2003-12-16application layer keep alive (ServerAliveInterval ServerAliveCountMax)Markus Friedl
for ssh(1), similar to the sshd(8) option; ok beck@; with help from jmc and dtucker@
2003-12-16Don't need linear mappings here eitherJason Wright
2003-12-16don't do a linear mapping (prevents warnings on the leopard)Jason Wright
2003-12-16add ipp (Internet Printing Protocol). ok deraadt@Jakob Schlyter
2003-12-16add daap (Digital Audio Access Protocol). ok deraadt@Jakob Schlyter
2003-12-16add afpovertcp. ok deraadt@Jakob Schlyter
2003-12-16expand documentation of built-ins:Jason McIntyre
group them according to POSIX - arithmetic, string, I/O, general and better document how they work. use section headers for more logical navigation. ok jose@
2003-12-16- s/recieve/receiveJason McIntyre
- kill whitespace at EOL
2003-12-16pfsync.4:Jason McIntyre
- new sentence, new line - kill whitespace at EOL - escape a dot at EOL pflog.4: - subject verb agreement
2003-12-16Compute openings value, out of boot-time count of devices.Niklas Hallqvist
Add reentrancy check of gdt_exec_ccb. The latter change changes timing enough to not trigger a timing bug on really fast machines too often. A real fix is being worked on.
2003-12-16Compute openings value, out of boot-time count of devices.Niklas Hallqvist
Add reentrancy check of gdt_exec_ccb. The latter change changes timing enough to not trigger a timing bug on really fast machines too often. A real fix is being worked on.
2003-12-16Test src.track timeout.Ryan Thomas McBride
2003-12-16BecuaseTheo de Raadt
2003-12-16Don't do all the heavy pfsync processing if there are no bpf listenersRyan Thomas McBride
and no network synchronization is happening.
2003-12-16clean the cpu types and syncMichael Shalayeff