On Wed, 25 Oct 2023 13:38:37 +0200, Alexander Bluhm wrote:
> Since 7.4 patch(1) does not work if an explicit patchfile is given on
> command line.
>
> https://marc.info/?l=openbsd-cvs&m=168941770509379&w=2
OK millert@
- todd
very utility.
I could go either way on that. It's true that this is default shell
behavior so you probably don't need to document it.
- todd
On Sun, 15 Oct 2023 13:53:46 -0500, Scott Cheloha wrote:
> Align timeout(1)'s execvp(3) failure statuses with those of GNU
> timeout. 127 for ENOENT, 126 for everything else.
Looks correct to me. OK millert@
- todd
age-ID if missing. I don't think it should be adding it to
relayed messages. Messages received on the submission port are
special, they need to be treated as local even though they originated
elsewhere.
- todd
hink so. Hard-coding ports is not great but there isn't a
way in the config file to indicate that explicitly.
- todd
On Wed, 04 Oct 2023 11:53:51 -0600, Todd C. Miller wrote:
> Yes, this should be fixed. One difference is that in FreeBSD,
> __swsetup() sets errno whereas in OpenBSD we set errno in the caller
> when cantwrite() is true. I think it makes sense to set both errno
> and the __SERR flag
this should be fixed. One difference is that in FreeBSD,
__swsetup() sets errno whereas in OpenBSD we set errno in the caller
when cantwrite() is true. I think it makes sense to set both errno
and the __SERR flag in the same place. We just need to decide which
place that is :-)
- todd
On Wed, 27 Sep 2023 10:59:26 -0600, "Todd C. Miller" wrote:
> I think we want support for arbitrary line lengths. There is only
> one place where we need to reallocate the line buffer.
The correct check is for "lp - line == linesz - 1". The code will
overwrite th
h with no breaking.
>
> Presumably nobody is relying on the current behaviour in scripts or other
> code, as it's always been broken.
I think we want support for arbitrary line lengths. There is only
one place where we need to reallocate the line buffer.
- todd
especially as regular malloc code did not
> change at all (the extra info needed is only collected if malloc flag
> D is set).
This is very cool. Being able to tell where the (now-freed) chunk
was allocated is a huge help in debugging this kind of issue.
- todd
L_TO_TIMESPEC if necessary.
- todd
On Mon, 11 Sep 2023 22:10:49 -0500, Scott Cheloha wrote:
> p_tv() is identical to p_ts() in every way except for the subsecond
> conversion constants.
>
> Better to write p_ts() once: in p_tv(), convert from timeval to
> timespec and call p_ts().
OK millert@
- todd
On Wed, 06 Sep 2023 10:43:34 +1000, Jonathan Gray wrote:
> tb updated us to the newer version a while ago
OK millert@
- todd
> 6878108432 79941f3aobj/options.o
> 390175 19024 85392 494591 78bff obj/pax
>
> I don't expect any regression on the ramdisks but a make release is
> running just in case.
Sure, nothing relies on that anyway.
- todd
HUP. Since our pipes are actually
bidiretional we can cheat and use select. Something like this:
- todd
lock_db() {
[ "${LOCKPID:-}" ] && return 0
# The installer doesn't have perl, so we can't lock there
[ -e /usr/bin/perl ] || return 0
nstead of the sleep() to tell when it goes away.
- todd
gh the minefield of matching bespoke
> "_t" types to real types and limits.
>
> Tweaks? ok?
OK millert@
- todd
>
> Is this a "so don't do that then", or is it expected to work?
> (There's no problem with "per-user crontab" files).
It is a bug. There is a missing call to Skip_Blanks() for the
/etc/crontab case. Instead of adding yet another unget_char() after
Skip_Blan
httpd(8) to append " \t " at the end of
> every header (legal per rfc 7230) and observing that ftp still works
> fine, including mtime handling.
OK millert@
- todd
If the mtime in the file header is larger than MAX_TIME_T, trucate
it to MAX_TIME_T, not INT_MAX. The existing assignment dates from
before we had a MAX_TIME_T definition in pax.
OK?
- todd
Index: cpio.c
===
RCS file: /cvs/src
On Mon, 26 Jun 2023 17:24:38 +0200, Paul de Weerd wrote:
> Having never heard of posix_spawn(3), I read the full manpage and
> (besides wondering "what's the point"), found that it's misspelled Ed
> Schouten's name:
Yes, that should be fixed.
- todd
Other implementations of "time -p" (both builtin and standalone)
only display two digits after the radix point. I'm a little concerned
about breaking scripts that consume out the output of "time -p".
Changing the precission of the non-portable output format is fine.
- todd
specifies only one optional string, while here for semplicity
> it's relaxed to allow anything.
This is a case where being liberal in what you accept seems fine.
OK millert@
- todd
On Tue, 07 Feb 2023 20:35:10 -0700, Todd C. Miller wrote:
> On Tue, 07 Feb 2023 17:17:02 -0700, Todd C. Miller wrote:
>
> > Yes, the bug is that the number is not displayed. The following
> > diff fixes that but there is still a bug because the resulting line
> > also
On Wed, 21 Jun 2023 19:11:09 +0200, Omar Polo wrote:
> On 2023/06/20 14:38:37 -0600, Todd C. Miller wrote:
> > > qid = ep+1;
> > > - if ((ep = strchr(qid, '|')) == NULL)
> > > - fatalx("Missing reqid: %s", line);
> > > - ep[0
ck is that we now need to use a few strncmp, but I think it's
> worth it.
This seems like a good approach, minor comments inline.
- todd
> diff /usr/src
> commit - 5c586f5f5360442b12bbc4ea18ce006ea0c3d126
> path + /usr/src
> blob - a714446c26fee299f4450ff1ad40289b5b3
code to be like open_memstream().
OK millert@
- todd
rything in sync.
Using the same order as smtpd.h makes sense. OK millert@
- todd
PD/issues/1213
>
> Diff below is straightforward and includes the documentation changes.
> I believe link-auth was forgotten in revision 1.61 of lka_filter.c
> when the mail-from/rcpt-to events got their fields swapped.
OK millert@
- todd
That probably requires a separate code path specifically for integers.
One thing at a time.
- todd
For context, see:
https://chaos.social/@Gottox/110527807405964874
https://github.com/chimera-linux/chimerautils/commit/1ecc1e99d4a309631e846a868b5a422f996704ac
. I have a few that I will
commit after this is in.
- todd
Index: usr.bin/seq/seq.c
===
RCS file: /cvs/src/usr.bin/seq/seq.c,v
retrieving revision 1.6
diff -u -p -u -r1.6 seq.c
--- usr.bin/seq/seq.c 25 Feb 2022 16:00:39 -
to substitute in the OpenBSD version and architecture.
- todd
On Sat, 03 Jun 2023 11:55:46 +0200, Omar Polo wrote:
> As per subject. While here I couldn't resist simplifying the "From "
> check too, although it is indipendent from the rest of the diff.
> (could commit separately if preferred.)
OK millert@
- todd
7;m also including sys/time.h in smtpd.h, as noted in
> event_init(3), since we're including event.h.
OK millert@
- todd
27;;' after
the "return 0".
- todd
t@ as well if you want to commit it.
- todd
That is also a lot easier to document.
- todd
re
>
> { local var=$(<$HTTP_SEC); } 2>/dev/null
>
> which is sufficiently opaque that I'm not sure it's an improvement.
OK millert@
- todd
On Wed, 24 May 2023 16:04:13 +0200, Omar Polo wrote:
> It makes `man -k any=S_ISVTX' slightly more useful by pointing at
> sticky(8) too other than strmode(3); may help if someone (like me :-)
> forgot about sticky(8) files.
OK millert@
- todd
splay the more useful message?
With diff:
$ echo $(< /nope)
ksh: /nope: No such file or directory
- todd
diff -u -p -u -r1.66 eval.c
--- bin/ksh/eval.c 13 Sep 2020 15:39:09 - 1.66
+++ bin/ksh/eval.c 24 May 2023 14:03:41 -
@@ -8,6 +8,7 @@
#include
#include
+#includ
lete output simpler by just asking
> the admin to delete the user, instead of both user and group.
Right.
- todd
s no other members.
This makes our userdel(8) behave more like the version on other
systems.
Opinions? This is something that has always bothered me and can
result in uid/gid mismatches if you remove a user, then re-add them
without removing the login group first.
Thoughts or strong opinions?
- t
> quite clear from the grammar in parse.y.
OK millert@
- todd
s);
> > + if (ret != 0)
> > + warnx("[Warning] unable to run `%s'", buf);
>
> more than "unable to run" I'd say "failed to run" or "command
> failed" / "exited with nonzero st
The way user(8) runs commands via system(3) is fragile. It does
not correctly handle paths with whitespace or shell metacharacters.
Rather than try to quote everything (which is also fragile) I think
it is safest to just exec the commands directly.
OK?
- todd
Index: usr.sbin/user/user.c
On Tue, 16 May 2023 19:26:07 +0200, Alexander Bluhm wrote:
> On Tue, May 16, 2023 at 11:15:31AM -0600, Todd C. Miller wrote:
> > Would it be possible to move the forward declaration of struct tdb
> > to netinet/tcp_var.h so it is not required in every driver?
>
> sure
Than
We can just use "cp -a skeldir/. homedir" to copy the skeleton dot
files to the new user's homedir. There's no good reason to use pax
when cp will do and this will simplify a future commit of mine.
- todd
Index:
closed
>
> So change most of them to fatalx which doesn't append errno. While
> here I'm also logging the actual error, via tls_config_error() or
> tls_error(), that before was missing.
>
> tls_config_new(), tls_server() and tls_client() failures are still
> logged with fatal(), which I believe it's correct.
OK millert@
- todd
All the callers of memsave() pass strlen(s) as the size argument.
We can eliminate the size argument and just use strdup(3) instead.
OK?
- todd
Index: user.c
===
RCS file: /cvs/src/usr.sbin/user/user.c,v
retrieving revision 1.128
K millert@. It would be nice to get these changes in portable
as well to avoid gratuitous differences.
- todd
;
> Why would you want to differentiate the address families here?
I was mostly just curious as I see FreeBSD seems to support this.
That made we wonder if there is hardware that only supports offloading
for IPv4.
- todd
The sigsuspend(2) man page doesn't spell out explicitly what happens
for signals that are discarded, either as the default action or
where the handler is set to SIG_IGN. I think it should.
OK?
- todd
Index: lib/libc/sys/sigsusp
onfig(8) we use "tcprecvoffload" and "tcpsendoffload". So, the
> user has a better insight of what this features are doing.
Is it possible to control these at the address family level? In
other words, is it possible to enable "tcprecvoffload" and
"tcpsendoffload" for inet but not inet6 or vice versa?
- todd
t; and month?
Sure. OK millert@ for this one too.
- todd
a const for a variable (no idea how it went there in
> -portable but it's not wrong so including that too.)
OK millert@
- todd
On Tue, 09 May 2023 21:45:30 +0200, Theo Buehler wrote:
> Some expressed concern that it should be done the other way around,
> i.e., leave the default at letsencrypt. Perhaps it's indeed better
> this way to avoid creating servers with bad certs.
OK millert@ for this version
- todd
On Mon, 08 May 2023 16:17:51 -, Tobias Stoeckmann wrote:
> Turns out that we have yet another possibility to trigger a theoretical
> signed integer overflow if pwd_tries is INT_MAX. This one avoids such
> situation as well.
OK millert@
- todd
.
OK?
- todd
Index: usr.sbin/cron/entry.c
===
RCS file: /cvs/src/usr.sbin/cron/entry.c,v
retrieving revision 1.54
diff -u -p -u -r1.54 entry.c
--- usr.sbin/cron/entry.c 6 May 2023 23:06:27 - 1.54
+++ usr.sbin/cron
value less than eiter the step value or the difference of the high
and low values (whichever is smaller).
- todd
Index: usr.sbin/cron/crontab.5
===
RCS file: /cvs/src/usr.sbin/cron/crontab.5,v
retrieving revision 1.41
diff -u -p -u
On Fri, 05 May 2023 17:05:05 -, Tobias Stoeckmann wrote:
> On Fri, May 05, 2023 at 11:00:12AM -0600, Todd C. Miller wrote:
> > This looks OK but I'd like to see an error message if waitpid()
> > really does fail. How about something like this, which also avoid
> >
tpid fails if the root user tries to
> sabotage the own passwd call. Let's just handle the error case here
> as well to avoid accessing undefined content of "res".
This looks OK but I'd like to see an error message if waitpid()
really does fail. How about something like
urally from the random value being in num1?
Also, using the % operator with the random value results in modulo
bias, which we would like to avoid. If we use a random offset based
on the step interval instead there is no need for the modulus.
- todd
On Thu, 04 May 2023 21:41:26 -, Klemens Nanni wrote:
> On Thu, May 04, 2023 at 03:30:30PM -0600, Todd C. Miller wrote:
> > This fixes two issues with the parsing of random values:
> >
> > 1) A random value with a step is now rejected. For example:
> >
> >
* echo max minute is 59
Whereas before it would work most (but not all!) of the time.
OK?
- todd
diff -u -p -u -r1.53 entry.c
--- usr.sbin/cron/entry.c 21 May 2022 01:21:29 - 1.53
+++ usr.sbin/cron/entry.c 4 May 2023 21:19:40 -
@@ -498,12 +498,17 @@ get_range(bitstr_t
Fix a bug introduced in rev 1.86 (fchown removal). Currently, if
the second seteuid(2) were to fail (not really possible) we would
leave a temporary file in the spool dir. This will be ignored by
cron but we still want to clean it up.
- todd
Index: usr.sbin/cron/crontab.c
On Thu, 04 May 2023 07:32:18 -0700, Navan Carson wrote:
> Any chance the syntax could be:
>
> ~/20 * * * * command
>
> To align with how ~ is used currently.
That is already a valid syntax, though not a terribly useful one.
It currently results in a random number with a step of 20.
- todd
(and reloaded).
The man page bits are from job@
Opinions? Does the proposed syntax seem OK?
- todd
Index: usr.sbin/cron/crontab.5
===
RCS file: /cvs/src/usr.sbin/cron/crontab.5,v
retrieving revision 1.41
diff -u -p -u -r1.41 crontab
) < 0;
}
In this case you probably want to assign c as an unsigned char.
E.g.
while ((c = (unsigned char)*v++) != '\0') {
...
}
Or better yet just declare c as unsigned char instead of int.
- todd
to not include the whole message, which may
> well be quite large.
Make sense to me. OK millert@
- todd
archive. I
have not yet verified that it gets extracted correctly. If there
is interest I will do some more testing.
- todd
Index: bin/pax/gen_subs.c
===
RCS file: /cvs/src/bin/pax/gen_subs.c,v
retrieving revision 1.32
diff -u -p -u
:xx:xx:xx:xx:xx
though I don't think my AP switches channel widths.
- todd
) is now being called with buflen -1, which can avoid one
> extra character of processing effort for a long input string.
I think that will be wrong for the "name" property. From
sys/dev/ofw/fdt.c:OF_getprop
if (len < 0 && strcmp(prop, "name") == 0) {
data = fdt_node_name(node);
if (data) {
len = strlcpy(buf, data, buflen);
...
So passing in buflen is probably correct.
- todd
self. Use an in-line Oo ... Oc instead, ok?
OK millert@
- todd
On Tue, 28 Mar 2023 16:19:42 +0200, Omar Polo wrote:
> sigh... forgot to advance the pointer after strrchr otherwise argv[0]
> would have been /ksh instead of "ksh".
OK millert@ for this version.
- todd
"sh" accordingly to $SHELL.
It might be best to use the basename of the actual shell for argv[0].
Our ksh for instance has slightly different behavior when invoked
as sh.
OK millert@ for the diff as-is.
- todd
> Index: region.c
>
On Sat, 25 Mar 2023 20:13:35 +0100, Otto Moerbeek wrote:
> Last arg also is a pointer, so pass NULL.
Looks like it's been that way since the initial CSRG commit.
OK millert@
- todd
itly NUL-terminate token.
This results in less code and is more easily audited.
I've also removed the duplicate check for *(pbuf+1) != '{'.
OK?
- todd
Index: usr.sbin/smtpd/mda_variables.c
===
RCS file: /cvs/src/usr.sbin/sm
ackets and your fix pf.conf, syslogd
> still does not send messages to remote syslog server.
>
> Better continue trying also in that case. It restores pre-6.5
> behavior.
OK millert@
- todd
On Wed, 08 Mar 2023 09:02:08 -0600, joshua stein wrote:
> In the non-fail case, done is set to NULL and then free()d.
> free(NULL) is legal but maybe worth removing?
Please leave this as-is. I don't think it is worth appeasing
cppcheck in this case.
> diff --git usr.bin/ssh/scp.c usr.bin/ssh/
utput
OK millert@
- todd
ERM to mountd(8), it should be done
> manually as there is too much involved with RPC daemons to make it
> automagic.
>
> Can this be flipped so `rcctl stop mountd' works again?
I don't think anything has changed in that respect. However, I'm
not sure why this was disabled in the first place.
- todd
The SIGHUP handler only sets a flag these days, there is no longer
any need to block it while using the exports list.
OK?
- todd
Index: sbin/mountd/mountd.c
===
RCS file: /cvs/src/sbin/mountd/mountd.c,v
retrieving revision 1.90
is set. Also, there is no
reason to use a critical section now that the SIGHUP handle just
sets a flag. Something like this (untested), should be fine.
It needs testing by someone who actually uses NFS though...
- todd
Index: sbin/mountd/mountd.c
On Wed, 15 Feb 2023 09:03:55 -0700, "Todd C. Miller" wrote:
> It should not be removing .shstrtab. What happens if you tell
> llvm-strip to preserve .shstrtab? E.g. --keep-section .shstrtab?
Nevermind, I misread the readelf output, the stripped binary does
actually have .shstrtab.
- todd
It should not be removing .shstrtab. What happens if you tell
llvm-strip to preserve .shstrtab? E.g. --keep-section .shstrtab?
- todd
On Tue, 14 Feb 2023 17:47:00 -0800, Andrew Hewus Fresh wrote:
> With the perl update, we get a new version of unicode available to
> update this file as well. This was just running the script with the new
> perl version.
OK millert@
- todd
On Sun, 05 Feb 2023 10:43:58 -0500, aisha wrote:
> The auth_getchallenge function doesn't seem to exist in the code base.
> OK to remove this reference?
Yes, that should be removed. OK millert@
- todd
Unfortunately we cannot use strtonum(3) here since there may be
non-digit characters following the number. So, strtoll(3)
it is then.
- todd
Index: lib/libc/time/strptime.c
===
RCS file: /cvs/src/lib/libc/time/strptime.c,v
appear to be insufficient.
My inclination is to just convert everything to use strtonum(3).
- todd
ms
> odd?)
It just returns a boolean value, 1 for OK, 0 for not OK. There is
a result parameter for the output value. The code is effectively
the same as _conv_num. I dislike that it uses int64_t instead of
time_t though.
- todd
Is there a reason you didn't just change the gmtime_r() in the 's'
case to localtime_r()? That seems like the simplest fix. Using
strtol() for what may be a 64-bit value on an 32-bit system looks
wrong.
- todd
especially in
> our world where so much other dangerous stuff has been stopped.
I wonder if it makes sense to have a version of sysctl.conf that
only gets used for the next reboot and then is removed, kind of
like /etc/rc.firsttime. Maybe call it /etc/sysctl.once.
- todd
On Fri, 20 Jan 2023 09:32:38 -0700, Bob Beck wrote:
> So isdigit(3) says in the first paragraph that
>
> 'The complete list of decimal digits is 0 and 1-9, in any locale.'
>
> Later on it says:
>
> 'On systems supporting non-ASCII single-byte character encodings,
> different c arguments may corres
On Wed, 04 Jan 2023 13:13:28 -0800, Nathan Houghton wrote:
> This patch removes a few remnants of the unsupported diff -l
> option from diff.c and the diff manual page.
>
> The diff.c usage() print is already correct, so no changes are
> needed there.
Thanks, committed.
- todd
On Mon, 29 Aug 2022 13:51:13 +0200, =?utf-8?B?0L3QsNCx?= wrote:
> In that case, how about this scissor-patch?
> It has the added benefit of removing the existing floating-point usage.
That version looks good to me, committed.
- todd
Updated version. I kept "The BLOCKSIZE environment variable" in
the -P description since it is the first time the man page metions
BLOCKSIZE.
- todd
Index: bin/df/df.1
===
RCS file: /cvs/src/bin/df/df.1,v
retrieving rev
), systems that already implement it
will not remove it. So I think the only concern is some theoretical
future POSIX system.
- todd
In POSIX mode, df(1) does not honor the BLOCKSIZE environment
variable.
Any comments on the wording?
- todd
Index: bin/df/df.1
===
RCS file: /cvs/src/bin/df/df.1,v
retrieving revision 1.48
diff -u -p -u -r1.48 df.1
--- bin/df/df.1
On Sat, 31 Dec 2022 10:33:26 -0500, Scott Cheloha wrote:
> The progress meter in scp(1) and sftp(1) updates periodically, once
> per second. But using alarm(3) to repeatedly rearm the signal causes
> that update period to drift forward:
OK millert@
- todd
1 - 100 of 1084 matches
Mail list logo
