[owc] Weekly src changes ending 2007-07-01

OWC auto at squish.net
Mon Jul 2 07:00:01 BST 2007


OpenBSD src changes summary for 2007-06-24 to 2007-07-01 inclusive
==================================================================

bin/stty                                distrib/sets
games/fortune                           lib/libc
libexec/identd                          libexec/spamd
sbin/badsect                            sbin/disklabel
sbin/fsck_ffs                           sbin/fsdb
sbin/ncheck_ffs                         sbin/newfs
sbin/pfctl                              sbin/quotacheck
share/doc                               share/man
share/termtypes                         sys/arch/amd64/include
sys/arch/amd64/stand/libsa              sys/arch/amd64/stand/mbr
sys/arch/amd64/stand/pxeboot            sys/arch/arm/arm
sys/arch/hppa/dev                       sys/arch/hppa/hppa
sys/arch/hppa/include                   sys/arch/i386/i386
sys/arch/i386/include                   sys/arch/i386/stand/libsa
sys/arch/i386/stand/mbr                 sys/arch/i386/stand/pxeboot
sys/arch/macppc/dev                     sys/arch/macppc/stand
sys/arch/mips64/mips64                  sys/arch/sparc/dev
sys/arch/sparc/sparc                    sys/arch/sparc64/conf
sys/arch/sparc64/dev                    sys/arch/zaurus/stand/zboot
sys/dev                                 sys/dev/eisa
sys/dev/i2c                             sys/dev/ic
sys/dev/isa                             sys/dev/onewire
sys/dev/pci                             sys/dev/pcmcia
sys/dev/tc                              sys/dev/usb
sys/kern                                sys/lib/libsa
sys/net                                 sys/netbt
sys/netinet                             sys/nfs
sys/scsi                                sys/stand
sys/sys                                 usr.bin/awk
usr.bin/checknr                         usr.bin/cvs
usr.bin/lam                             usr.bin/locate
usr.bin/netstat                         usr.bin/rcs
usr.bin/sdiff                           usr.bin/ssh
usr.bin/tcopy                           usr.bin/tip
usr.bin/unifdef                         usr.sbin/config
usr.sbin/pkg_add                        usr.sbin/quot
usr.sbin/tcpdump                        

== bin =============================================================== 01/10 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/bin

stty

  ~ stty.1                                

  > Greg A. Woods: stty(1) should mention tty(4) (jmc@)

== distrib =========================================================== 02/10 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/distrib

sets

  ~ lists/base/mi                         

  > sync (deraadt@)

== games ============================================================= 03/10 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/games

fortune

  ~ fortune/fortune.c                     ~ strfile/strfile.c

  > Set shortest string count to 0 in strfile(8), if there a no strings found
  > and check the number of strings in fortune(6) to prevent empty fortune
  > files to cause division by zero. Fixes PR 5514.
  > OK miod@ ray@ deraadt@ (moritz@)

== lib =============================================================== 04/10 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/lib

libc

  ~ gen/scandir.c                         

  > It seems st_size is not always a good estimator for the number of
  > directory entries.	ZFS, for instance, sets st_size to the number
  > of entries.
  > Use the same algorithm for the initial estimate but keep a minimum
  > to prevent a divide by zero.  When the buffer is filled, instead
  > of re-using the same flawed algorithm, double the buffer size.
  > Also add check for integer overflow.
  > Bug reported by Ed Ravin, ZFS testing environment provided by Bryan
  > Allen.
  > OK millert (ray@)

  ~ stdlib/strtol.3                       ~ stdlib/strtoul.3

  > Document behavior when no conversion could be performed.  Adapted
  > from FreeBSD.  With help from jmc at . (millert@)

  ~ time/Theory                           ~ time/ialloc.c
  ~ time/private.h                        ~ time/tz-art.htm
  ~ time/tz-link.htm                      ~ time/zdump.c
  ~ time/zic.8                            ~ time/zic.c

  > Update to tzcode2007f from elsie.nci.nih.gov (millert@)

== libexec =========================================================== 05/10 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/libexec

identd

  ~ identd.c                              

  > Add a SIGCHLD handler for background mode instead of doing waitpid()
  > at a single point in the event loop (which would always leave
  > a single zombie around after the first connection).
  > Also use daemon() instead of doing it by hand.  OK deraadt@ beck@
  > (millert@)

spamd

  ~ spamd.8                               

  > s/is is/it is/.  From Jim Razmus, jim (at) bonetruck (dot) org; thanks.
  > (tom@)

== sbin ============================================================== 06/10 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/sbin

badsect

  ~ badsect.c                             

  > more daddr64_t changes; checked by otto (deraadt@)

disklabel

  ~ disklabel.c                           

  > some trivial delinting (deraadt@)

fsck_ffs

  ~ dir.c                                 ~ extern.h
  ~ fsck.h                                ~ inode.c
  ~ main.c                                ~ pass1.c
  ~ pass1b.c                              ~ pass2.c
  ~ pass4.c                               ~ pass5.c
  ~ utilities.c                           

  > Teach fsck_ffs about 64-bit block addresses. ok (and help) moritz@ pedro@
  > (otto@)

fsdb

  ~ fsdb.h                                

  > Teach fsdb about 64-bit block addresses. ok moritz@ pedro@ (otto@)

ncheck_ffs

  ~ ncheck_ffs.c                          

  > more daddr64_t changes; checked by otto (deraadt@)

newfs

  ~ mkfs.c                                ~ newfs.c

  > Enable creation of larger than 1TB FFS2 filesystems. Note that this
  > won't work yet. There remains sys/ufs work.  ok deraadt@ (otto@)

pfctl

  ~ pfctl.c                               

  > allow states instead of just state for -s and -F for consistency's sake.
  > since only the first letter matters in the end and state still matches
  > states since we do strncmp this does not change xisting behaviour and
  > just allows plural states to be used as well.
  > pt out by Tom Van Looy <tom at ctors.net> (henning@)

  ~ pfctl.8                               

  > states here too not state (henning@)

quotacheck

  ~ quotacheck.c                          

  > more daddr64_t changes; checked by otto (deraadt@)

== share ============================================================= 07/10 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/share

doc

  ~ psd/05.sysman/2.3.t                   

  > kill "is is"; from Charles Longeau (jmc@)

man

  ~ man9/sensor_attach.9                  

  > document changes to the sensor_task api. task_register now returns a
  > pointer to the sensor_task it just created, and task_unregister takes it
  > to remove that task from the system. (dlg@)

  ~ man9/sensor_attach.9                  

  > tweak the description of sensor_task_unregister; (jmc@)

  ~ man4/san.4                            

  > Add a BUGS section, ok jmc@ (canacar@)

  ~ man5/fs.5                             

  > update struct fs; ok otto (jmc@)

  ~ man9/disk.9                           

  > update struct disk; ok deraadt (jmc@)

  ~ man5/types.5                          

  > - add daddr32_t and daddr64_t
  > - sync comment for daddr_t (jmc@)

  ~ man5/pf.conf.5                        

  > checkd -> checked; from Nicholas Marriott (jmc@)

  ~ man4/malo.4                           

  > malo doesn't support hostap mode so remove example which uses
  > hostap; ok claudio@ (stevesk@)

  ~ man4/cas.4                            

  > Add missing gentbi(4) in SYNOPSYS. (kettenis@)

termtypes

  ~ termtypes.master                      

  > kill "is is"; from Charles Longeau (jmc@)

== sys =============================================================== 08/10 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/sys

arch/amd64/include

  ~ vmparam.h                             

  TAGGED OPENBSD_4_1
  > increase VM_PHYSSEG_MAX some more, some systems need it.
  > Diff from todd, ok weingart (ckuethe@)

  ~ vmparam.h                             

  TAGGED OPENBSD_4_0
  > increase VM_PHYSSEG_MAX some more, some systems need it.
  > Diff from todd, ok weingart (ckuethe@)

arch/amd64/stand/libsa

  ~ dev_i386.c                            

  > wt(4) was removed a while ago, so don't support booting off it anoymore.
  > i386/amd64 was ok tom
  > zaurus ok drahn (mk@)

arch/amd64/stand/mbr

  ~ mbr.S                                 

  > "of of" is wrong - it should be "off of" - but "from" is even nicer (tom@)

arch/amd64/stand/pxeboot

  ~ devopen.c                             

  > wt(4) was removed a while ago, so don't support booting off it anoymore.
  > i386/amd64 was ok tom
  > zaurus ok drahn (mk@)

arch/arm/arm

  ~ cpu.c                                 

  > Move more fpe-related stuff in #ifdef ARMFPU blocks. (miod@)

arch/hppa/dev

  ~ com_ssio.c                            

  > Attach as console when appropriate. (kettenis@)

  ~ ssio.c                                

  > Switch associated pciide(4) into native mode, and route its interrupt
  > appropriately.  Make it use irq 5, and rewire lpt(4) to irq 7.
  > ok deraadt@ (kettenis@)

  ~ apic.c                                

  > Bump number of supported interrupt routing entries. (kettenis@)

  ~ astro.c                               

  > Add a few more registers. (kettenis@)

  ~ astro.c                               

  > Reset PDC devices before enabling iova space if we're not on a serial
  > console.
  > Makes glass console on four-digit B/C/J-class systems work. (kettenis@)

  ~ apic.c                                

  > Correctly count interrupts for devices that share interrupts. (kettenis@)

arch/hppa/hppa

  ~ machdep.c                             

  > Crank initial storage for extents a bit.  We really should allocate extents
  > using malloc once vm is initialised, but this will do for now.
  > ok deraadt@ (kettenis@)

arch/hppa/include

  ~ param.h                               ~ pdc.h

  > Add PDC_IO. (kettenis@)

  ~ param.h                               

  > Oops!  Revert previous commit. (kettenis@)

arch/i386/i386

  ~ trap.c                                

  > Provide a default case so that we don't end up trying to dereference
  > vframe when it hasn't been set.
  > Prompted by a diff from mickey@
  > ok art@ (tom@)

  ~ pmap.c                                

  > According to Intel errata:
  > "AI91 - Update of attribute bits on page directories without immediate
  > tlb shootdown may cause unexpected processor behavior.".
  > When we're allocating kernel page tables, we use the generic page table
  > allocation function that sets PG_u then we immediately remove the PG_u.
  > This might not be enough, because the PDE can get preloaded into the
  > translation cache before we clear the PG_u. So even without the errata,
  > this could cause us horrible trouble.
  > Instead of first entering the PDE with PG_u and then removing it for
  > kernel page tables, just enter it with the right bits every time.
  > tom@ ok (art@)

arch/i386/include

  ~ vmparam.h                             

  TAGGED OPENBSD_4_1
  > increase VM_PHYSSEG_MAX some more, some systems need it.
  > Diff from todd, ok weingart (ckuethe@)

  ~ vmparam.h                             

  TAGGED OPENBSD_4_0
  > increase VM_PHYSSEG_MAX some more, some systems need it.
  > Diff from todd, ok weingart (ckuethe@)

arch/i386/stand/libsa

  ~ dev_i386.c                            

  > wt(4) was removed a while ago, so don't support booting off it anoymore.
  > i386/amd64 was ok tom
  > zaurus ok drahn (mk@)

arch/i386/stand/mbr

  ~ mbr.S                                 

  > "of of" is wrong - it should be "off of" - but "from" is even nicer (tom@)

arch/i386/stand/pxeboot

  ~ devopen.c                             

  > wt(4) was removed a while ago, so don't support booting off it anoymore.
  > i386/amd64 was ok tom
  > zaurus ok drahn (mk@)

arch/macppc/dev

  ~ asms.c                                

  > rework sensor tasks to use the kernels generic workq rather than a special
  > kernel thread of its own. the api has changed (which will be fixed in the
  > manpage shortly) so all the users of sensor tasks that i can find have
  > been fixed too.
  > noone tested, so its going in to force people to run with it.
  > "put it in" deraadt@ (dlg@)

arch/macppc/stand

  ~ main.c                                

  > When bootfile was made global so that macppc could modify it, it
  > clashed with a global variable already used by the bootblocks.
  > Therefore rename bootfile here to kernelfile.
  > ok drahn@ (tom@)

arch/mips64/mips64

  ~ context.S                             

  > In cpu_switch(), store the saved cpl in the pcb as a 64 bit store, so that
  > the high order bits are not undefined when invoking hw_setintrmask() on
  > return. (miod@)

arch/sparc/dev

  ~ xd.c                                  ~ xy.c

  > Fix xd and xy disklabel handling. (miod@)

  ~ si.c                                  

  > Do not define DEBUG unconditionnaly. (miod@)

  ~ xd.c                                  ~ xy.c

  > Perform minimal disklabel setup without needing to do the
  > disklabel_to_sun()
  > dance, so that we do not depend on its behaviour. No functional change.
  > (miod@)

  ~ obio.c                                ~ power.c

  > Use mapiodev() instead of mapdev() when applicable. This is just syntactic
  > sugar. (miod@)

arch/sparc/sparc

  ~ disksubr.c                            

  > Fix xd and xy disklabel handling. (miod@)

  ~ disksubr.c                            

  > Perform minimal disklabel setup without needing to do the
  > disklabel_to_sun()
  > dance, so that we do not depend on its behaviour. No functional change.
  > (miod@)

  ~ auxreg.c                              

  > Use mapiodev() instead of mapdev() when applicable. This is just syntactic
  > sugar. (miod@)

arch/sparc64/conf

  ~ GENERIC                               

  > Add gentbi(4). (kettenis@)

arch/sparc64/dev

  ~ pcf8591_ofw.c                         

  > rework sensor tasks to use the kernels generic workq rather than a special
  > kernel thread of its own. the api has changed (which will be fixed in the
  > manpage shortly) so all the users of sensor tasks that i can find have
  > been fixed too.
  > noone tested, so its going in to force people to run with it.
  > "put it in" deraadt@ (dlg@)

  ~ pcons.c                               

  > sick of extra static messing up tracebacks; ok miod (deraadt@)

arch/zaurus/stand/zboot

  ~ devopen.c                             

  > wt(4) was removed a while ago, so don't support booting off it anoymore.
  > i386/amd64 was ok tom
  > zaurus ok drahn (mk@)

dev

  ~ softraid.c                            

  > rework sensor tasks to use the kernels generic workq rather than a special
  > kernel thread of its own. the api has changed (which will be fixed in the
  > manpage shortly) so all the users of sensor tasks that i can find have
  > been fixed too.
  > noone tested, so its going in to force people to run with it.
  > "put it in" deraadt@ (dlg@)

dev/eisa

  ~ eisa.c                                

  > Fix a couple of "of of"s in comments. (tom@)

dev/i2c

  ~ ad741x.c                              ~ adm1021.c
  ~ adm1024.c                             ~ adm1025.c
  ~ adm1026.c                             ~ adm1030.c
  ~ adm1031.c                             ~ adt7460.c
  ~ asb100.c                              ~ ds1631.c
  ~ fcu.c                                 ~ fintek.c
  ~ gl518sm.c                             ~ lm87.c
  ~ maxim6690.c                           ~ pca9554.c
  ~ tsl2560.c                             ~ w83l784r.c

  > rework sensor tasks to use the kernels generic workq rather than a special
  > kernel thread of its own. the api has changed (which will be fixed in the
  > manpage shortly) so all the users of sensor tasks that i can find have
  > been fixed too.
  > noone tested, so its going in to force people to run with it.
  > "put it in" deraadt@ (dlg@)

dev/ic

  ~ ami.c                                 ~ ciss.c
  ~ lm78.c                                ~ lm78var.h
  ~ mfi.c                                 

  > rework sensor tasks to use the kernels generic workq rather than a special
  > kernel thread of its own. the api has changed (which will be fixed in the
  > manpage shortly) so all the users of sensor tasks that i can find have
  > been fixed too.
  > noone tested, so its going in to force people to run with it.
  > "put it in" deraadt@ (dlg@)

  ~ lm78.c                                ~ lm78var.h

  > support early W83627EHF-A; tested by Sam Fourman Jr; discussed with Gong
  > Jun; ok kettenis@ (cnst@)

  ~ ar5xxx.c                              

  > Implement bit swap for a byte in O(log(N)) rather than O(N), prompted
  > by an interview question I was asked.
  > ok reyk@ (tom@)

dev/isa

  ~ it.c                                  

  > rework sensor tasks to use the kernels generic workq rather than a special
  > kernel thread of its own. the api has changed (which will be fixed in the
  > manpage shortly) so all the users of sensor tasks that i can find have
  > been fixed too.
  > noone tested, so its going in to force people to run with it.
  > "put it in" deraadt@ (dlg@)

  ~ elink.c                               ~ seagate.c

  > ansify/de-register
  > no binary change (jasper@)

  ~ lm78_isa.c                            

  > fix potential crash due to wrong ca_devsize; whilst here, also fix iobase
  > type; ok grange, kettenis (cnst@)

dev/onewire

  ~ owsbm.c                               ~ owtemp.c

  > rework sensor tasks to use the kernels generic workq rather than a special
  > kernel thread of its own. the api has changed (which will be fixed in the
  > manpage shortly) so all the users of sensor tasks that i can find have
  > been fixed too.
  > noone tested, so its going in to force people to run with it.
  > "put it in" deraadt@ (dlg@)

dev/pci

  ~ arc.c                                 

  > rework sensor tasks to use the kernels generic workq rather than a special
  > kernel thread of its own. the api has changed (which will be fixed in the
  > manpage shortly) so all the users of sensor tasks that i can find have
  > been fixed too.
  > noone tested, so its going in to force people to run with it.
  > "put it in" deraadt@ (dlg@)

  ~ pciide_natsemi_reg.h                  

  > Fix typo in (unused) #define. (kettenis@)

  ~ ixgb_ee.c                             

  > Fix "the the"s.  ok miod at . (tom@)

  ~ pciide.c                              

  > Make natsemi_pci_intr() use pciide_intr_flag().  Needed on the PC87415 part
  > integrated on the PC87560 Legacy IO chip found on several hppa workstation
  > models.
  > ok jsg@ (kettenis@)

  ~ if_msk.c                              

  > Shrink code by not using __inline without static; like we did for sk(4).
  > From brad at comstyle dot com. (kettenis@)

  ~ ahci.c                                

  > define away some of the magic on jmicron controllers (dlg@)

  ~ ahci.c                                

  > demystify a little more of the jmicron pci control registers (dlg@)

  ~ ahci.c                                

  > when we attach to the jmicron controllers, whack their configuration around
  > a bit. this guarantees that the controller has its ahci stuff enabled for
  > us to use (which was previously done with magic numbers). this also configs
  > the controller to enable a second function for pciide to attach to. (dlg@)

  ~ pci_quirks.c                          

  > force the pci bus probe routines to look for extra functions on jmicron
  > controllers. pciide on function 1 isnt enabled until ahci gets to touch it,
  > which is too late for the probe routines to discover that there are high
  > functions to look at. (dlg@)

  ~ pcidevs                               

  > Add a few devices found in submitted dmesgs (jsg@)

  ~ pcidevs.h                             ~ pcidevs_data.h

  > regen (jsg@)

dev/pcmcia

  ~ if_malo.c                             ~ if_maloreg.h
  ~ if_malovar.h                          

  > Add 802.11d (domain info) command. (mglocker@)

dev/tc

  ~ tc.c                                  

  > Fix a couple of "of of"s in comments. (tom@)

dev/usb

  ~ if_kuereg.h                           

  > Alignment fixes from mickey. (jsg@)

  ~ ugen.c                                

  > respect timeouts. prompted by freebsd pr110122, but also proplerly
  > calculate the timeout. help & ok jsg (henning@)

kern

  ~ kern_sensors.c                        

  > rework sensor tasks to use the kernels generic workq rather than a special
  > kernel thread of its own. the api has changed (which will be fixed in the
  > manpage shortly) so all the users of sensor tasks that i can find have
  > been fixed too.
  > noone tested, so its going in to force people to run with it.
  > "put it in" deraadt@ (dlg@)

  ~ subr_disk.c                           

  > use daddr64_t for size in bounds_check_with_label(); prompted by miod@;
  > ok deraadt@ (otto@)

lib/libsa

  ~ loadfile.c                            

  > Fix the description of loadfile() semantics in comment (and kill "and
  > and" at the same time). (tom@)

net

  ~ if_pfsync.c                           ~ pf.c
  ~ pf_ioctl.c                            ~ pfvar.h

  > Save some bytes and make code more readable by removing junk union and
  > unused ifname (this information is in struct pf_state_sync now).
  > Also a bit of KNF on the pf_state struct.
  > ok mpf@ henning@ (mcbride@)

  ~ if_pfsync.c                           ~ pf.c
  ~ pfvar.h                               

  > pretty mechanical change: now that the state tables use seperate state
  > keys that can map to multiple states (last not least for ifbound) we don't
  > need state tables hanging off each struct kif representing an interface
  > any more. use two globals for the two tables. ok markus ryan (henning@)

  ~ if.h                                  

  > crank ifq_maxlen from 50 to 256, so it is not smaller than most interfaces
  > rx rings any more. forwarding boxes with many fast interfaces can still use
  > some more, but this is a saner default.
  > ok deraadt markus henric (henning@)

  ~ if_pfsync.c                           

  > Fix a race condition during ruleset reload; make sure we don't walk off
  > the end of the array of rule pointers when attaching a pfsync'd state
  > to a rule. Reported in PR5508 by mayer at netlab.nec.de.
  > ok henning@ (mcbride@)

netbt

  ~ bt_input.c                            ~ bt_proto.c
  ~ bt_var.h                              ~ hci.h
  ~ hci_unit.c                            

  > Remove some unused leftover code; original request from jasper@ (uwe@)

  ~ rfcomm_upper.c                        

  > Fix "the the"s.  ok miod at . (tom@)

netinet

  ~ tcp_subr.c                            ~ tcp_usrreq.c
  ~ tcp_var.h                             

  > merge tcp_set_iss() and tcp_set_tsm(); ok mcbride, djm (on earlier version)
  > (markus@)

nfs

  ~ nfs_socket.c                          ~ nfs_syscalls.c

  > replace magic number with a define.
  > (2 -> SHUT_RDWR in calls to soshutdown()).
  > From Iruata Souza.
  > ok deraadt@ (thib@)

scsi

  ~ safte.c                               ~ ses.c

  > rework sensor tasks to use the kernels generic workq rather than a special
  > kernel thread of its own. the api has changed (which will be fixed in the
  > manpage shortly) so all the users of sensor tasks that i can find have
  > been fixed too.
  > noone tested, so its going in to force people to run with it.
  > "put it in" deraadt@ (dlg@)

stand

  ~ boot/boot.c                           

  > When bootfile was made global so that macppc could modify it, it
  > clashed with a global variable already used by the bootblocks.
  > Therefore rename bootfile here to kernelfile.
  > ok drahn@ (tom@)

  ~ boot/boot.c                           

  > loadfile() returns an open file descriptor on successful exit, so
  > close it before going off to execute the newly-loaded kernel.
  > ok deraadt@ (tom@)

sys

  ~ sensors.h                             

  > rework sensor tasks to use the kernels generic workq rather than a special
  > kernel thread of its own. the api has changed (which will be fixed in the
  > manpage shortly) so all the users of sensor tasks that i can find have
  > been fixed too.
  > noone tested, so its going in to force people to run with it.
  > "put it in" deraadt@ (dlg@)

  ~ cdefs.h                               

  > Fix "the the"s.  ok miod at . (tom@)

== usr.bin =========================================================== 09/10 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin

awk

  ~ awk.1                                 

  > printf(1) is the correct reference here, not printf(3);
  > from otto; ok millert (jmc@)

checknr

  ~ checknr.c                             

  > From Tobias Stoeckmann:
  > - Fix a file descriptor leak.
  > - Print usage if -a and -c flags are passed a zero-length string.
  > OK millert and moritz (ray@)

cvs

  ~ client.c                              ~ cmd.c
  ~ file.c                                ~ getlog.c
  ~ log.c                                 ~ log.h
  ~ remote.h                              ~ server.c

  > add support for cvs rlog.
  > from Mickey.
  > ok joris@ ray@ (niallo@)

  ~ Makefile                              ~ cmd.c
  ~ cvs.h                                 + release.c

  > Bring back the release command. More work to be done though.
  > OK joris@ niallo at . (xsa@)

  ~ checkout.c                            ~ client.c
  ~ cvs.h                                 ~ update.c

  > We were missing the needed clue/magic to have our remote setup
  > handle files that have been removed from the repository
  > correctly, which resulted in them still being in the working copy
  > after doing an update.
  > noticed and diff tested by ckuethe at . (joris@)

  ~ diff.c                                

  > fix diff output in a remote setup so our client does not freak out.
  > noticed and "yes, that's much shinier" by ckuethe. (joris@)

  ~ cmd.c                                 ~ cvs.h

  > Remove last bits about login and logout commands, which will never be
  > implemented anyway. OK ray@ niallo at . (xsa@)

  ~ client.c                              ~ remote.h
  ~ server.c                              

  > have the server be more strict about the requests the client
  > sends that need a 'Directory' request first, so we do not
  > end up with very weird and possibly insecure behaviour.
  > more soon!
  > ok niallo@ ray@ xsa@ (joris@)

  ~ history.c                             

  > when logging to CVSROOT/history as a server fill in the
  > directory as <remote> instead of /tmp/cvs-serv<pid> (joris@)

  ~ update.c                              

  > do not call cvs_checkout_file() for file removals in local mode,
  > only required in server mode due to the fact we need to let the client
  > now its been removed, found this the hard way. (joris@)

  ~ commit.c                              

  > When commiting to a server in a remote setup classify
  > the files based upon their entry in CVS/Entries instead of
  > using the local classifying code and trying to locate the RCSfile
  > on our local disks, which ofc does not work correctly.
  > OK niallo@ (joris@)

  ~ add.c                                 ~ admin.c
  ~ checkout.c                            ~ client.c
  ~ cvs.h                                 ~ diff.c
  ~ diff3.c                               ~ edit.c
  ~ file.c                                ~ getlog.c
  ~ history.c                             ~ import.c
  ~ rcs.c                                 ~ remove.c
  ~ status.c                              ~ tag.c
  ~ update.c                              

  > Sync revisions and time buffers size to be consistent with each others.
  > Simplifies further size tweaks if needed.
  > OK niallo@ ray at . (xsa@)

  ~ diff_internals.c                      

  > Sync cvs/rcs diff code.
  > OK joris and xsa. (ray@)

  ~ client.c                              ~ remote.h
  ~ server.c                              

  > Server-side bits for release command. (xsa@)

  ~ edit.c                                

  > Sync time buffer size; missed in previous commit. (xsa@)

lam

  ~ lam.c                                 

  > Fix buffer overflow reported by Tobias Stoeckmann by increasing the
  > number of file descriptors to the correct limit, NOFILE_MAX, and
  > improve calculation of open file count.
  > OK moritz (ray@)

locate

  ~ locate/locate.1                       

  > some wording improvements for -d and -m; from Jacek Masiulaniec (jmc@)

netstat

  ~ main.c                                

  > Bring the index defines back in sync with struct nlist after removal of IPX
  > symbols from the table. This fixes the netstat issues some people reported.
  > OK henning@ (claudio@)

rcs

  ~ ci.c                                  

  > allow ci -N<tag name> even when no changes.  mimics GNU behaviour.
  > based on a diff from Tero Koskinen <tero.koskinen at iki.fi>
  > ok joris@ (niallo@)

  ~ diff.c                                

  > Some syncs with diff(1):
  > - prepare() only returns 0, so change it to void and remove code
  > that checks its return value.
  > - Rename diff_len[] to len[].
  > - Remove casts that made the code a no-op:
  > (size_t)filesize <= SIZE_MAX
  > OK xsa (ray@)

  ~ diff.c                                

  > diff(1) sync, mostly spacing and moving crap around, except:
  > - stone() and newcand() cannot return < 0, so don't check it.
  > - \r, \v, \f are also whitespace
  > OK joris (ray@)

  ~ diff.c                                

  > Sync cvs/rcs diff code.
  > OK joris and xsa. (ray@)

  ~ ci.c                                  ~ co.c
  ~ diff3.c                               ~ rcs.c
  ~ rcsclean.c                            ~ rcsdiff.c
  ~ rcsmerge.c                            ~ rcsprog.c
  ~ rcsprog.h                             ~ rlog.c

  > Sync revisions and time buffers size to be consistent with each others.
  > Simplifies further size tweaks if needed.
  > OK ray at . (xsa@)

sdiff

  ~ sdiff.c                               

  > Add missing undocumented --ignore-case flag, to improve compatibility
  > with GNU sdiff.
  > From Steven Kreuzer <skreuzer () exit2shell ! com> (ray@)

  ~ sdiff.1                               ~ sdiff.c

  > Alias `1' and `2' to `l' and `r', making interactive merges easier
  > on the fingers.  This was recently added to GNU sdiff:
  > http://www.nabble.com/sdiff-UI-enhancement-t3877253.html
  > Initial diff from Steven Kreuzer <skreuzer (at) exit2shell dot com>.
  > OK jmc, millert, sobrado, deraadt. (ray@)

  ~ sdiff.1                               

  > groff understands `|', so no need to duplicate macros. (jmc@)

  ~ sdiff.1                               

  > *long* options are undocumented; ok ray (jmc@)

ssh

  ~ channels.c                            

  > Correct test for window updates every three packets; prevents sending
  > window updates for every single packet.  ok markus@ (dtucker@)

  ~ atomicio.c                            

  > Include <poll.h> like the man page says rather than <sys/poll.h>.  ok djm@
  > (dtucker@)

tcopy

  ~ tcopy.c                               

  > mt_count is just an int (deraadt@)

tip

  ~ tip.h                                 

  > kill "is is"; from Charles Longeau (jmc@)

unifdef

  ~ unifdef.c                             

  > kill "is is"; from Charles Longeau (jmc@)

== usr.sbin ========================================================== 10/10 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/usr.sbin

config

  ~ ukc.c                                 

  > Add check for nlist failure to prevent a segfault.
  > Remove ret variable, which is only set but not never checked.
  > return (1) instead of exit(1) for consistency.
  > Original diff from Rainer Giedat.
  > OK millert, maja, and otto. (ray@)

pkg_add

  ~ OpenBSD/Add.pm                        ~ OpenBSD/Delete.pm

  > typo (espie@)

  ~ Makefile                              + OpenBSD/UpdateSet.pm

  > placeholder to split off the code that handles UpdateSets, as it's
  > currently scattered all over the place. (espie@)

  ~ pkg_add                               ~ OpenBSD/Dependencies.pm
  ~ OpenBSD/Vstat.pm                      

  > the big gap one: create UpdateSet for all packages being installed, and
  > consider only UpdateSet internally. This involves moving stuff around a
  > bit to keep pkg_add -n working, adding creators from_location for
  > UpdateSet,
  > and tweaking dependencies heavily.
  > More changes later, like sanitizing PackageLocation state engine, or
  > creating UpdateSet earlier for actual updates. (espie@)

  ~ pkg_add                               

  > missed one mark_as_installed. (espie@)

  ~ pkg_info                              

  > fix pkg_info's find_by_path behavior, which I accidentally broke two
  > versions back... (espie@)

  ~ OpenBSD/Delete.pm                     

  > fix saving of old partial packages... problem noticed by Paul Irofti
  > (espie@)

  ~ pkg_add                               ~ pkg_delete
  ~ OpenBSD/UpdateSet.pm                  ~ OpenBSD/Vstat.pm

  > move the UpdateSet stuff into its separate file, leaving Vstat much
  > cleaner. (espie@)

quot

  ~ quot.c                                

  > daddr64_t changes to match other parts of tree; ok otto (deraadt@)

tcpdump

  ~ print-ip.c                            ~ print-ip6.c
  ~ print-atalk.c                         

  > When aligning buffers correctly handle the case where the
  > buffers overlap, which happens on 64 bit archs, when
  > handling encapsulated packets. Reported and tested by Jurjen Oskam
  > additional testing by Stuart Henderson and todd@, ok henning@ (canacar@)

  ~ nfs.h                                 ~ print-nfs.c

  > make it easy to spot wich nfs version the traffic is.
  > input & OK canacar@ (thib@)

===============================================================================


More information about the owc mailing list