Re: openssl s_time, speed: use monotime for absolute interval measurement

On Tue, Dec 05 2017, Scott Cheloha <scottchel...@gmail.com> wrote: > Hey, > > Sorry for the delay. > > On Sun, Nov 26, 2017 at 07:00:36PM +0100, Jeremie Courreges-Anglas wrote: >> On Sun, Nov 26 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: >>

Re: iked, don't return NULL in print_host

On Wed, Nov 29 2017, Patrick Wildt <patr...@blueri.se> wrote: > On Wed, Nov 29, 2017 at 04:32:24PM +0100, Claudio Jeker wrote: >> On Wed, Nov 29, 2017 at 02:43:45AM +0100, Jeremie Courreges-Anglas wrote: >> > On Wed, Nov 29 2017, Claudio Jeker <cje...@diehard.n-r-g.com

Re: iked, don't return NULL in print_host

On Wed, Nov 29 2017, Claudio Jeker wrote: > On Wed, Nov 29, 2017 at 01:59:06AM +0100, Claudio Jeker wrote: >> Seen in my log file: >> Nov 28 17:47:22 dramaqueen iked: vfprintf %s NULL in "%s: %s %s from %s to >> %s ms gid %u, %ld bytes%s" >> >> and >> >> Nov 29

Re: TCP/UDP/etc input w/o KERNEL_LOCK()

On Tue, Nov 28 2017, Alexander Bluhm wrote: > On Tue, Nov 28, 2017 at 02:42:58PM +0100, Martin Pieuchot wrote: >> > login: panic: kernel diagnostic assertion "_kernel_lock_held()" failed: >> > file "/usr/src/sys/kern/uipc_socket.c", line 1882 >> > Stopped at

Hide carp pass and sppp authname from userland

On Mon, Nov 27 2017, Stefan Sperling <s...@stsp.name> wrote: > On Mon, Nov 27, 2017 at 01:31:17AM +0100, Stefan Sperling wrote: >> On Sun, Nov 26, 2017 at 06:17:14PM +0100, Jeremie Courreges-Anglas wrote: >> > >> > I don't think anything has been committed regardi

Re: hide wpakey from root by default

On Mon, Nov 27 2017, Stefan Sperling wrote: > On Mon, Nov 27, 2017 at 02:33:59AM +0100, Stefan Sperling wrote: >> Most people I've talked to seem to be OK with never exposing >> these secrets to userland in the first place. Makes sense. > Better diff for the wireless part. > WEP

Re: FUSE: link should return EPERM if the source file is a directory

On Mon, Nov 27 2017, Helg wrote: > fusefs_link returns the wrong error code when attempting to create a > hard link to a directory. It returns EISDIR when it should instead > return EPERM. Discovered while running the ffs test suite on ntfs-3g > and confirmed by

Re: isakmpd: use monotonic clock for event timeouts

On Fri, Nov 24 2017, Scott Cheloha wrote: > Hi, > > These events are meants to fire after an interval has elapsed, > so we should use the monotonic clock to measure. > > The pattern throughout the daemon of loading the current time, > adding a timeout to the structure

Re: ksh(1): kill the "version" function

On Sun, Nov 26 2017, Theo Buehler wrote: > On Sun, Nov 26, 2017 at 08:18:04PM +0100, Anton Lindqvist wrote: >> While at it, another gem: in Vi normal mode typing ^V inserts the >> version string. Does not seem to be documented either. > > Yes, can we please zap it? It's

ksh(1): kill the "version" function

On Tue, Nov 21 2017, Anton Lindqvist wrote: > Hi, > While writing tests for the Emacs editing mode in ksh I discovered some > potential errors in the manual: > > - Sync the order of key bindings in emacs.c with the manual > > - ^W is bound to delete-word-backward and not

Re: openssl s_time, speed: use monotime for absolute interval measurement

On Sun, Nov 26 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: > On Sat, Nov 25 2017, Brent Cook <bust...@gmail.com> wrote: >> Thanks guys. This will make enabling this on the odder platforms in >> portable easier. > > NB: if we want to able to mix app_tm

Re: openssl s_time, speed: use monotime for absolute interval measurement

y named app_timer_real() and app_timer_user(). The Windows version of app_timer_real() could start as a simple copy of app_timer_user(), marked with XXX... > On Fri, Nov 24, 2017 at 7:03 AM, Scott Cheloha <scottchel...@gmail.com> > wrote: > >> > On Nov 24, 2017, at 6:58 AM, Jerem

Re: hide wpakey from root by default

I don't think anything has been committed regarding this issue, right? On Sat, Nov 18 2017, Stuart Henderson wrote: > On 2017/11/17 18:28, Ted Unangst wrote: >> Stefan Sperling wrote: >> > Or is modifying ifconfig sufficient? >> > We are more concerned about textual

Re: Could not savecore after reproducible crash - solved

On Sun, Nov 26 2017, Christopher Zimmermann wrote: > This 'crash' can be prevented by turning ddb.console off. So there's still a bug. ddb.console=1 shouldn't crash the host. :) > Sorry for the noise. > I'm still wondering why I could not recover the dump. (No idea

Re: openssl s_time, speed: use monotime for absolute interval measurement

On Wed, Nov 22 2017, Scott Cheloha wrote: > Whoops, ignore that last patch, it lacked the > static changes in apps_posix.c This looks good to me. I'm tempted to commit the apps_posix.c part first: it seems to me that app_tminterval() could be reused in s_time.c, leading

Re: openssl s_time, speed: use monotime for absolute interval measurement

On Sat, Nov 11 2017, Scott Cheloha wrote: > Hi, Hi, > times(3) is okay for user CPU measurement but is inappropriate > for absolute interval measurement as its output is subject to > changes by both adjtime(2) and settimeofday(2). > > The attached diff replaces it with

Re: __warn_references: drop redundant "warning: " prefix

On Sun, Nov 12 2017, Scott Cheloha wrote: > Hi, > > GNU ld has prefixed the contents of .gnu.warning.SYMBOL sections > with "warning: " since 2003, so the messages themselves need not > contain the prefix anymore. > > If LLVM ld ever acknowledges .gnu.warning sections I

Re: pppd: explicit_bzero sensitive buffers

On Sat, Nov 11 2017, Scott Cheloha wrote: > Hi, > > You want explicit_bzero(3) for these buffers. > > Zeroing a buffer is compiler- and system-dependent, so I added a > new macro. I have committed the fixed version, with the macro. > I'll send a pull request upstream if

Re: hide wpakey from root by default

On Fri, Nov 17 2017, Stefan Sperling wrote: > There have been several instances of people mailing out WPA keys as > part of ifconfig output, e.g. in bug reports. This happens when you > run ifconfig as root and copy/paste without thinking. > > I see no real need to ever show the

mandoc warning tweak

I hit this while reviewing a recent diff: ritchie /usr/src/usr.sbin/ospf6d$ mandoc -Tlint ospf6d.conf.5 mandoc: ospf6d.conf.5:211:5: STYLE: typo in section name: Sh AREAS instead of CAVEATS which is a false positive. The distance between "AREAS" and "CAVEATS" is 3, maybe it's a bit much

Re: [PATCH] amd64/bsd.rd: add growfs(8)

On Mon, Nov 06 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: > On Mon, Nov 06 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: >> On Sun, Nov 05 2017, Job Snijders <j...@openbsd.org> wrote: >>> Goodmorning everyone, >>> >>>

Re: [PATCH] amd64/bsd.rd: add growfs(8)

On Mon, Nov 06 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: > On Sun, Nov 05 2017, Job Snijders <j...@openbsd.org> wrote: >> Goodmorning everyone, >> >> While quite some resizing scenarios can be done from within single user >> mode, resizing t

Re: [PATCH] amd64/bsd.rd: add growfs(8)

On Sun, Nov 05 2017, Job Snijders wrote: > Goodmorning everyone, > > While quite some resizing scenarios can be done from within single user > mode, resizing the root partition requires you to bring your own > growfs(8) binary into the ramdisk environment. The below patch adds >

Re: document how ospfd interacts with carp

On Mon, Nov 06 2017, Remi Locherer wrote: > Hi, > > ospfd.conf(5) should mention what ospfd does automatically when > configured on carp interfaces. > > ok? ok jca@ > Remi > > > Index: ospfd.conf.5 > === >

Re: forbid config reloads in ospf6d

On Sun, Nov 05 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: > ospf6d consistently fails when I ask it to reload its config, even > though I have a very basic test setup: > > area 0.0.0.0 { > interface em0 { passive } > interface vether0 > } >

forbid config reloads in ospf6d

ospf6d consistently fails when I ask it to reload its config, even though I have a very basic test setup: area 0.0.0.0 { interface em0 { passive } interface vether0 } Fixing ospf6d doesn't seem trivial. Having it fail and exit doesn't seem to be a sufficient incentive, so I

Re: pr(1) - ignore LC_TIME

On Wed, Nov 01 2017, Mark Kettenis wrote: >> Date: Wed, 1 Nov 2017 17:04:44 +0100 >> From: Jan Stary >> >> Doesn't OpenBSD strftime() ignore LC_TIME anyway? > > One day we will support it. I see no reason to remove this call (or > any of the others

Re: awk setlocale(LC_NUMERIC)

On Wed, Nov 01 2017, Jan Stary wrote: > Why does awk need to fiddle with LC_NUMERIC? awk is not developed in OpenBSD land[1]. Since LC_NUMERIC doesn't have any effect on OpenBSD I'd rather just let this code be. [1] https://www.cs.princeton.edu/~bwk/btl.mirror/ > Jan > >

Re: ksh alloc(): switch calloc back to malloc

On Wed, Nov 01 2017, Theo Buehler wrote: > The plaintext history diff changed memory allocation in alloc() from > malloc() to calloc(). At that point, some of the new code in history.c > may (or may not) have depended on this. In the meantime, this code was > removed by jca

Re: makefs: malloc -> emalloc

On Wed, Nov 01 2017, "Michael W. Bombardieri" wrote: > Hello, > > makefs has an xmalloc.c with emalloc() function, but one thing > was still using malloc() directly. This patch makes malloc() > always happen through emalloc(). Fails to build, implicit emalloc decl. > - Michael > >

Re: nsd(8): make ip-transparent option work on OpenBSD

On Wed, Nov 01 2017, Florian Obser wrote: > OK? ok jca@ btw the diff below has already been posted to tech@: https://marc.info/?l=openbsd-tech=147747266305927=2 > diff --git server.c server.c > index c0835ce8c11..fe74f9a1b56 100644 > --- server.c > +++ server.c > @@

Re: unbound(8): make ip-transparent option work on OpenBSD

On Wed, Nov 01 2017, Florian Obser wrote: > This is useful on systems where IP addresses are dynamically > configured (dhclient(8), slaacd(8)) and are not yet up when unbound > starts. > > To quote the man page: > >ip-transparent: > If yes, then use

Re: kill net/slip.h

On Sun, Oct 29 2017, Martin Pieuchot wrote: > contains a handful of defines optionally used by > tcpdump(8). Nothing in codesearch.debian.net includes it. > > So I'd like to remove it, to reduce the number of network headers. > > ok? ok -- jca | PGP : 0x1524E7EE / 5135

Re: ->

On Sat, Oct 28 2017, Martin Pieuchot wrote: > Another diff to use in userland. Kernel and bootloader are still > untouched. > > ok? ok jca@ -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE

Re: urndis printf cleanup

On Fri, Oct 27 2017, Mikhail <mp39...@gmail.com> wrote: > On Fri, Oct 27, 2017 at 1:19 PM, Jeremie Courreges-Anglas > <j...@wxcvbn.org> wrote: >> On Fri, Oct 27 2017, Mikhail <mp39...@gmail.com> wrote: >>> add missed DEVNAME's and \n's >> >>

Re: elf_abi.h -> elf.h

On Fri, Oct 27 2017, Martin Pieuchot wrote: > I'd like to get rid of which is not portable at all. > > is at least present on Solaris, FreeBSD and OSX. > > ok? ok -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE

Re: urndis printf cleanup

On Fri, Oct 27 2017, Mikhail wrote: > add missed DEVNAME's and \n's They are not missing. The point is to have _attach() print a single line in dmesg. PS: your diff would not apply (mangled whitespace) > Index: dev/usb/if_urndis.c >

Re: tftpd(8): diff for ip path rewrite

On Tue, Oct 24 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: > On Mon, Oct 23 2017, Jan Klemkow <j.klem...@wemelug.de> wrote: >> On Sun, Oct 22, 2017 at 09:32:54PM +0000, Jeremie Courreges-Anglas wrote: >>> On Sat, Oct 21 2017, Jan Klemkow <j.klem...@wem

Re: tftpd(8): diff for ip path rewrite

On Mon, Oct 23 2017, Jan Klemkow <j.klem...@wemelug.de> wrote: > On Sun, Oct 22, 2017 at 09:32:54PM +, Jeremie Courreges-Anglas wrote: >> On Sat, Oct 21 2017, Jan Klemkow <j.klem...@wemelug.de> wrote: >> > On Fri, Oct 20, 2017 at 12:04:41PM +, Jeremie Courre

Re: close cron sockets in child processes

On Mon, Oct 23 2017, "Todd C. Miller" <todd.mil...@courtesan.com> wrote: > On Mon, 23 Oct 2017 17:27:15 +0200, Jeremie Courreges-Anglas wrote: > >> Here it is. Liberal use of O_CLOEXEC except for poke_daemon() which is >> more obvious (but I can add it t

Re: close cron sockets in child processes

On Mon, Oct 23 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: > On Mon, Oct 23 2017, "Todd C. Miller" <todd.mil...@courtesan.com> wrote: >> On Mon, 23 Oct 2017 09:05:48 +0200, Jeremie Courreges-Anglas wrote: >> >>> That doesn't work. Wh

Re: close cron sockets in child processes

On Mon, Oct 23 2017, "Todd C. Miller" <todd.mil...@courtesan.com> wrote: > On Mon, 23 Oct 2017 09:05:48 +0200, Jeremie Courreges-Anglas wrote: > >> That doesn't work. When here, we hold cronSock and dfd (should be 3 and >> 4). We need dfd to get fd (sh

Re: close cron sockets in child processes

On Fri, Oct 20 2017, "Todd C. Miller" wrote: > On Fri, 20 Oct 2017 16:25:32 +0200, Florian Riehm wrote: > >> cron(8) opens /var/run/cron.sock for communication with crontab(1). >> The forked cronjobs have the socked still open. >> This prevents restarting cron while a

Re: tftpd(8): diff for ip path rewrite

On Sat, Oct 21 2017, Jan Klemkow <j.klem...@wemelug.de> wrote: > On Fri, Oct 20, 2017 at 12:04:41PM +, Jeremie Courreges-Anglas wrote: >> On Fri, Oct 20 2017, Sebastien Marie <sema...@online.fr> wrote: >> > On Thu, Oct 19, 2017 at 08:58:12PM +0200, Jan Klemk

Re: close cron sockets in child processes

On Fri, Oct 20 2017, Florian Riehm wrote: > Hi, > > cron(8) opens /var/run/cron.sock for communication with crontab(1). > The forked cronjobs have the socked still open. > This prevents restarting cron while a job is running: > (CRON) DEATH (already running) Hah. > I think

Re: tftpd(8): diff for ip path rewrite

On Thu, Oct 19 2017, Stuart Henderson wrote: > On 2017/10/19 16:22, Theo de Raadt wrote: >> I am always worried by non-intuitive magic behaviour. >> >> It may serve some obvious purposes, but for someone else it is going >> to break things. >> >> I worry. > > The

Re: tftpd(8): diff for ip path rewrite

On Fri, Oct 20 2017, Sebastien Marie wrote: > On Thu, Oct 19, 2017 at 08:58:12PM +0200, Jan Klemkow wrote: >> >> Index: tftpd.c >> === >> RCS file: /mount/openbsd/cvs/src/usr.sbin/tftpd/tftpd.c,v >> retrieving

Re: tftpd(8): diff for ip path rewrite

On Wed, Oct 18 2017, Jan Klemkow wrote: > On Wed, Oct 18, 2017 at 08:37:48PM +, Jason McIntyre wrote: >> On Wed, Oct 18, 2017 at 10:25:13PM +0200, Jan Klemkow wrote: >> > This diff adds an option for client IP address path prefixes to the >> > tftpd(8). First, I used

Re: cannot take the address of stdout

On Thu, Oct 19 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: > When building an updated version of ports/security/polarssl, I hit the > following failure: > > [219/427] /usr/ports/pobj/mbedtls-2.6.0/bin/cc > -I/usr/ports/pobj/mbedtls-2.6.0/mbedtls-2.6.0/include

Re: ksh: remove the deprecated emacs-usemeta option

On Thu, Oct 19 2017, Sebastian Benoit <be...@openbsd.org> wrote: > Jeremie Courreges-Anglas(j...@wxcvbn.org) on 2017.10.18 22:16:48 +0200: >> >> This would make ''set -o emacs-usemeta'' a fatal error, which means that >> subsequent lines in your kshrc will not b

Re: ksh: skip long history lines

On Wed, Oct 18 2017, Ori Bernstein <o...@eigenstate.org> wrote: > On Thu, 19 Oct 2017 00:22:51 +0200 > Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: > >> Those variables are static so that error messages are only printed once >> in the shell lifetime.

Re: ksh: extra rewind call

On Thu, Oct 19 2017, Sebastian Benoit <be...@openbsd.org> wrote: > Jeremie Courreges-Anglas(j...@wxcvbn.org) on 2017.10.19 00:36:00 +0200: >> >> This call was added along with the magic check, but it not actually >> needed: history_load already calls rewind(3). I f

cannot take the address of stdout

When building an updated version of ports/security/polarssl, I hit the following failure: [219/427] /usr/ports/pobj/mbedtls-2.6.0/bin/cc -I/usr/ports/pobj/mbedtls-2.6.0/mbedtls-2.6.0/include -I/usr/ports/pobj/mbedtls-2.6.0/mbedtls-2.6.0/tests -O2 -pipe -Wall -Wextra -W

ksh: extra rewind call

This call was added along with the magic check, but it not actually needed: history_load already calls rewind(3). I feel like the magic check should also be in history_load(), so that a ksh process can recover at runtime from a binary->plaintext history file migration. Worth the trouble? ok for

Re: ksh: skip long history lines

Hi Ori, thanks for your feedback. Reply and updated diff below, On Wed, Oct 18 2017, Ori Bernstein <o...@eigenstate.org> wrote: > On Wed, 18 Oct 2017 22:33:51 +0200 > Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: > >> >> It would be nice to support arbitra

ksh: skip long history lines

It would be nice to support arbitrarily long lines, but a first step would be to skip them gracefuly. The code modifies the loop condition: no tests against ferror(3)/feof(3) are performed any more (I don't see their point). We could print a warning on ferror(), though, but that would be

ksh: remove the deprecated emacs-usemeta option

This would make ''set -o emacs-usemeta'' a fatal error, which means that subsequent lines in your kshrc will not be run. I think people were given enough time to cope with this (6.2 users get a warning). ok? Index: misc.c === RCS

ksh: history_load signature

The return value of this function isn't checked. A small change, the next step is to skip lines that are too long. ok? Index: history.c === RCS file: /d/cvs/src/bin/ksh/history.c,v retrieving revision 1.71 diff -u -p -p -u -r1.71

inetd & ipv4-compatible ipv6 addresses

We don't and won't support this, so let's not even try. ok? Index: inetd.c === RCS file: /cvs/src/usr.sbin/inetd/inetd.c,v retrieving revision 1.156 diff -u -p -r1.156 inetd.c --- inetd.c 4 Oct 2017 19:52:26 - 1.156

inetd: kill a few unsupported socket types

In my tests, only SOCK_STREAM and SOCK_DGRAM effectively work (even seqpacket on a unix socket is broken), so kill the others. We don't even have SOCK_RDM support. ok? Index: inetd.8 === RCS file:

Re: [patch] exiting bgplgsh with ^D

On Wed, Oct 04 2017, Denis Fondras wrote: > Hi, > > Here is a patch to make ^D (CTRL-D) escape from bgplgsh instead of displaying > help. This makes sense, but I'm wondering if this changes anything when actually using bgplg? Reyk, do you remember if there was a specific

Re: syslogd close at EAGAIN

On Sat, Sep 16 2017, Alexander Bluhm wrote: > Hi, > > syslogd tries to ignore EAGAIN for tty. Unfortunately it closes > the file descriptor before checking the errno. So f_file contains > a bad file descriptor that may be reused at the next open. > > I think f_file

Re: syslogd preopen console

On Wed, Sep 13 2017, Alexander Bluhm wrote: > Hi, > > syslogd has special code for reporting errors before it has been > initialized. Then it tries to log to console. For every message > it reopens the console with file descriptor passing from the privsep > parent. Of

Re: sysctl_int(), sysctl_struct() & MP work

On Wed, Sep 13 2017, Martin Pieuchot wrote: > On 12/09/17(Tue) 15:53, Martin Pieuchot wrote: >> Diff below reduces the scope of the NET_LOCK(), this time in sysctl >> path. It is interesting for multiple reasons: >> >> - It reduces the contention on the NET_LOCK(), which

Re: syslogd close *:514 sockets

On Mon, Sep 11 2017, Alexander Bluhm wrote: > Hi, > > In the default configuration syslogd keeps two *:514 UDP sockets > open. > > udp 0 0 *.514 *.* > udp6 0 0 *.514 *.* >

Re: ksh: histptr and undefined pointer comparisons

On Thu, Aug 31 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: > I've noticed this some time ago and, while it doesn't seem to be > a problem in practice, it's still undefined. C says what happens with > pointers: > - within the bounds of an array > - one past the las

Re: SSHFP with EDNS0/DNSSEC

On Thu, Sep 07 2017, Andreas Bartelt <o...@bartula.de> wrote: > On 07/12/17 18:49, Jeremie Courreges-Anglas wrote: >> Eric Faurot <e...@faurot.net> writes: >> >>> On Wed, Jul 12, 2017 at 07:45:36AM +0200, Christian Barthel wrote: >>>> Hi, >

Re: ksh: history file is corrupt

On Thu, Aug 24 2017, Stuart Henderson <s...@spacehopper.org> wrote: > On 2017/08/24 17:42, Jeremie Courreges-Anglas wrote: >> On Wed, Aug 23 2017, Stuart Henderson <s...@spacehopper.org> wrote: >> > On 2017/08/15 11:57, Jeremie Courreges-Anglas wrote: >>

ksh: use PATH_MAX

ok? Index: c_ksh.c === RCS file: /d/cvs/src/bin/ksh/c_ksh.c,v retrieving revision 1.50 diff -u -p -r1.50 c_ksh.c --- c_ksh.c 21 Mar 2016 13:35:00 - 1.50 +++ c_ksh.c 2 Sep 2017 22:18:50 - @@ -99,7 +99,7 @@

ksh: histptr and undefined pointer comparisons

I've noticed this some time ago and, while it doesn't seem to be a problem in practice, it's still undefined. C says what happens with pointers: - within the bounds of an array - one past the last element of an array but nothing about a pointer one element before the first element of an array.

ksh(1): drop needless chunk

I don't think this is relevant. Index: ksh.1 === RCS file: /cvs/src/bin/ksh/ksh.1,v retrieving revision 1.195 diff -u -p -r1.195 ksh.1 --- ksh.1 30 Aug 2017 17:08:45 - 1.195 +++ ksh.1 30 Aug 2017 17:23:59 -

Re: strtok_r(3): mention standards

On Mon, Aug 28 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: > On Mon, Aug 28 2017, "Todd C. Miller" <todd.mil...@courtesan.com> wrote: >> On Mon, 28 Aug 2017 14:29:18 +0200, Jeremie Courreges-Anglas wrote: >> >>> IIUC strtok_r app

Re: ksh: add support for HISTCONTROL ignorespace & ignoredups

On Wed, Aug 30 2017, "Todd C. Miller" wrote: > Looks good to me. Do we want to document this as an extension in > the manual? We're not very good about documenting extensions to > traditional ksh behavior. I'm not against this but I'm not thrilled either. If anyone

Re: ksh: kill emacs-usemeta

On Wed, Aug 30 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: > On Wed, Aug 30 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: >> As mentioned in https://marc.info/?l=openbsd-tech=150401359712984=2 >> emacs-usemeta isn't used since some time already.

Re: ksh: kill emacs-usemeta

On Wed, Aug 30 2017, Jeremie Courreges-Anglas <j...@wxcvbn.org> wrote: > As mentioned in https://marc.info/?l=openbsd-tech=150401359712984=2 > emacs-usemeta isn't used since some time already. I have a diff that > just hides the flag but I changed my mind. The people who use the

ksh: add support for HISTCONTROL ignorespace & ignoredups

This implements a feature I miss from bash. What I care about: - not clog the history with repeated commands - easily prevent some commands to go in the history The diff below implements HISTCONTROL for this, user-facing behavior based on bash. The default ksh behavior doesn't change. I did

ksh: kill emacs-usemeta

As mentioned in https://marc.info/?l=openbsd-tech=150401359712984=2 emacs-usemeta isn't used since some time already. I have a diff that just hides the flag but I changed my mind. The people who use the saved output of "set +o" from an earlier ksh might be negatively affected, but that's a bit

Re: csh(1): zap redundant assignment

On Wed, Aug 30 2017, Anton Lindqvist wrote: > Comments? OK? Sure. > Index: lex.c > === > RCS file: /cvs/src/bin/csh/lex.c,v > retrieving revision 1.24 > diff -u -p -r1.24 lex.c > --- lex.c 23 Jan 2017

Re: [patch] kbd should give an error unless we're root

On Tue, Aug 29 2017, Jesper Wallin wrote: > Hi, > > I noticed that kbd(8) lets you list all available keyboard encodings > with the -l flag. However, we get no error and no list unless we're > running this as root. I apologize if this is intended and I fail to > understand

ksh: completely drop setlocale(3)

Since last year ksh no longer sets the locale in main.c, so the diff below is a no-op but removes 6k of text from a static ksh(1). The emacs-usemeta flag is unused since 2012. I'm working on a diff to kill it gracefully. ok? Index: emacs.c

Re: strtok_r(3): mention standards

On Mon, Aug 28 2017, "Todd C. Miller" <todd.mil...@courtesan.com> wrote: > On Mon, 28 Aug 2017 14:29:18 +0200, Jeremie Courreges-Anglas wrote: > >> IIUC strtok_r appeared in IEEE Std 1003.1c-1995, the diff below >> documents that. That would be the first use

strtok_r(3): mention standards

IIUC strtok_r appeared in IEEE Std 1003.1c-1995, the diff below documents that. That would be the first use of .St -p1003.1c-95 in the tree. (Is this one available online? My search skillz have failed me so far...) Other systems: - linux, "POSIX.1-2001, POSIX.1-2008", FreeBSD - freebsd, "IEEE

ksh: walking valid history entries

Right now the code uses two methods to access "valid" entries in the history array. One method only looks at entries in [history;histptr], the other method walks the array until it finds a NULL pointer. The latter method appears fragile to me. *IIUC* the test at the beginning of

Re: [phess...@openbsd.org: add pinebook support to u-boot]

On Sun, Aug 27 2017, Peter Hessler wrote: > - Forwarded message from Peter Hessler - > taken from Icenowy Zheng's repository, > commit bd656cab0c3e0669e977641c15095f32d7b0731f > > This gives us basic support for the Pine64 Pinebook > > OK?

ksh: deallocate dead history entries

Let's try to free allocated lines when resizing / reloading. When using a big HISTSIZE and multiple concurrent shells, resource usage climb up rather quickly. The initial version didn't bother with resetting the lines pointers to NULL, but better safe(r) than sorry, I guess. ok? NB: after

ksh: history file is corrupt

On Wed, Aug 23 2017, Stuart Henderson <s...@spacehopper.org> wrote: > On 2017/08/15 11:57, Jeremie Courreges-Anglas wrote: >> CVSROOT: /cvs >> Module name: src >> Changes by: j...@cvs.openbsd.org2017/08/15 11:57:57 >> >> Modified files: >&

Re: ifstated: stop tracking interface indexes

On Tue, Aug 15 2017, Rob Pierce <r...@2keys.ca> wrote: > On Mon, Aug 14, 2017 at 11:26:46PM -0400, Jeremie Courreges-Anglas wrote: >> On Mon, Aug 14 2017, Rob Pierce <r...@2keys.ca> wrote: >> > ifstated currently tracks and maintains the index of each monitored

Re: ksh(1) history lines allocation

On Tue, Aug 15 2017, Rob Pierce wrote: [...] > I was able to reproduce the problem with a HISTSIZE of 10 which at 125000 > entries rendered my system unusable. With the patch I am running fine with a > HISTSIZE of 12 and have come back several times after hitting the

Re: ifstated: stop tracking interface indexes

On Mon, Aug 14 2017, Rob Pierce wrote: > ifstated currently tracks and maintains the index of each monitored interface > and does not maintain interface names. This means we need to re-index on > interface departure and arrival. > > The following diff moves away from indexes to

ksh(1) history lines allocation

So I tinkered with the way ksh(1) tracks memory allocation, trying to make it faster in the general case. One approach used a RB tree, I wrote since a simple hash table implementation which seems to work rather well. But the actual problem I'd first like to solve is a corner case. I use

Re: rtadvd: no longer decrement lifetimes in real time

This one fell through the cracks... On Sat, Aug 12 2017, Florian Obser wrote: > Stop supporting prefix lifetimes that decrement in real time. > It complicates the code, it's off by default and RFC 4861 section > 6.2.1 lists it as MAY. > After this we can stop regenerating

Re: ksh: remove special handling of export PS1 as root

On Fri, Aug 11 2017, Philip Guenther wrote: > Currently, if ksh starts as root and PS1 was exported but doesn't contain > '#' or '\$' (which will be expanded to '#') then it'll override it with > the root default of '# '. This is kind of a corner case, because PS1 is > not

Re: slaacd: simplify route message handling

On Fri, Aug 11 2017, Florian Obser wrote: > When read(2)'ing from the routing socket only one message is returned. > Clue provided by jca and claudio. > > OK? ok jca@ -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE

Re: rtadvd: open ioctl socket up front

On Fri, Aug 11 2017, Florian Obser wrote: > No need to constantly re-open a socket. Just open it up front and keep > it around. ok jca@ -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE

Re: rtadvd: simplify route message handling

On Thu, Aug 10 2017, Florian Obser <flor...@openbsd.org> wrote: > On Tue, Aug 08, 2017 at 11:08:14AM +0200, Jeremie Courreges-Anglas wrote: >> >> So rtadvd has this complex code that tries to parse multiple routing >> messages, when it actually reads only one message at

rtadvd: simplify route message handling

So rtadvd has this complex code that tries to parse multiple routing messages, when it actually reads only one message at a time from the routing socket. The diff below attempts to acknowledge this and tries to be as mechanical as possible, variable renaming/gc'ing can happen later. Better use

rtadvd: reschedule timer when a new prefix is added

ra_timer_update() just updates rai->timer.tm, we need to call evtimer_add(3) for the new timeout to be scheduled. I missed this one when converting rtadvd to libevent. ok? Index: config.c === RCS file:

Re: libc syslog dead variable

On Tue, Aug 08 2017, Alexander Bluhm wrote: > Hi, > > Since sendsyslog(2) handles the LOG_CONS parameter, conp in syslog(3) > is a dead variable. > > ok to remove? ok jca@ > bluhm > > Index: lib/libc/gen/syslog_r.c >

Re: ifstated: add handing of departed interfaces

On Sun, Aug 06 2017, Rob Pierce wrote: > The following diff adds support for detecting the state change of a departed > interface. ifstated is not a very verbose daemon, so this diff quietly does > the right thing (i.e. there is no exttra warning about a departing interface). But

Re: ifstated: consistent use of log.c

On Mon, Aug 07 2017, Rob Pierce <r...@2keys.ca> wrote: > On Sun, Aug 06, 2017 at 06:47:38PM +0200, Jeremie Courreges-Anglas wrote: >> On Thu, Aug 03 2017, Rob Pierce <r...@2keys.ca> wrote: >> > As a result ifstated.c no longer needs err.

<    1   2   3   4   5   6   7   8   9   >