On Thu, Dec 15, 2016 at 09:47:53AM -0500, sven falempin wrote:
> > maintain tmpfs
> >
> So fixing the BUG section of the man page ?
That's not what "maintain" means.
Your diff doesn't apply here, can you resend?
ok kspillner@. It would be nice to replace some of those exit(1) with
exit(EXIT_FAILURE), but that should be a separate diff that also updates the
err/errx's as well.
On Oct 1, 2014, at 13:06, Fritjof Bornebusch frit...@alokat.org wrote:
On Wed, Oct 01, 2014 at 06:41:25PM +0100, Nicholas
On Fri, Sep 12, 2014 at 12:38:07AM -0700, Doug Hogan wrote:
Hmm this doesn't look right to me. ressl_config is not allocated the
same way as the other variables. The other variables such as ssl,
sslhost, etc are local to that function and allocated there.
ressl_config is a global and only
On Fri, Sep 12, 2014 at 09:39:30AM -0500, Kent R. Spillner wrote:
Of course, you're right. Is there any reason why ressl_config must be
global? It's only used in url_get. It should be harmless to create a
new one each time as long as we always free it.
Here's a diff for that:
Forgot
While reviewing tedu@'s libressl config cleanup diffs I noticed we're
not explicitly freeing ressl_config in ftp(1).
Ok?
Index: fetch.c
===
RCS file: /work/cvsroot/src/usr.bin/ftp/fetch.c,v
retrieving revision 1.129
diff -p -u
On Wed, Sep 10, 2014 at 02:42:43PM +0200, Marcin Piotr Pawlowski wrote:
Yes, I think that it could be is possible to double clean the node cache.
Updated diff with suggestion from Stefan.
ok kspillner@, but would prefer a better name than ieee80211_clean_cached. :)
Perhaps
On Wed, Sep 10, 2014 at 07:39:00PM +0200, Stefan Sperling wrote:
I think ieee80211_clean_cached() is fine.
The function only removes notes in CACHE state. These are nodes
that are visible but haven't otherwise interacted with us.
There are additional states, all of which refer to nodes which
487.24 user 367.82 sys
On Sun, Sep 07, 2014 at 03:19:23PM -0400, Jean-Philippe Ouellet wrote:
On Wed, Sep 03, 2014 at 03:31:50PM -0500, Kent R. Spillner wrote:
Need to re-roll for -current?
Yep, sorry. I knew the the dup3 changes would break it when I saw them
go in but I've been
Need to re-roll for -current?
/usr/src/sys/kern/kern_descrip.c: In function 'dodup3':
/usr/src/sys/kern/kern_descrip.c:312: error: 'struct filedesc' has no member
named 'fd_ofileflags'
cc -Werror -Wall -Wstrict-prototypes -Wmissing-prototypes -Wno-main
-Wno-uninitialized
acpidock(4) updates its sensor's value when docking and undocking, but it
never sets its sensor's status. As a result, it's difficult to hook into
docking/undocking changes with sensorsd(8) (to change the value of
machdep.lidsuspend, for example).
This diff sets the status to SENSOR_S_OK when
Haven't tested, just eyeballing the diff; I have two questions inline below.
(Apologies if iOS Mail.app mangles the quoting :()
On Aug 14, 2014, at 17:39, Vadim Zhukov persg...@gmail.com wrote:
2014-08-11 4:37 GMT+04:00 Vadim Zhukov persg...@gmail.com:
This patch adds support for handling
On Fri, Aug 15, 2014 at 02:13:55PM -0700, Philip Guenther wrote:
In case someone is bored (heh) and looking for a user land bug to fix...
Given this:
printf '1\n2' a
printf '1\n2\n3' b
Compare the output of
diff -u a b
with the output of
diff -uw a b
The former gives
);
}
/*
On Wed, Jul 16, 2014 at 05:55:55PM -0500, Kent R. Spillner wrote:
*Bump*
On Jul 10, 2014, at 12:33, Kent R. Spillner kspill...@acm.org wrote:
Ping.
On Thu, May 01, 2014 at 01:22:56PM -0500, Kent R. Spillner wrote:
After sending my previous reply I
:03AM -0500, Kent R. Spillner wrote:
Sorry, let me split this into smaller diffs to ease review.
First up, the diff below removes a redundant cast: buf is declared a
char * so there's no need to cast it to a char *. I noticed the same
issue in ufs.c, too, so fix it in both places.
Index
, buf, buf_size)))
return rc;
dp = (struct direct *)buf;
On Tue, Jul 22, 2014 at 01:01:05PM -0500, Kent R. Spillner wrote:
Next, use NULL instead of casting 0 to pointer types.
Index: sys/lib/libsa/ufs.c
);
+ bcopy(buf, namebuf, link_len);
}
/*
On Tue, Jul 22, 2014 at 02:16:26PM -0500, Kent R. Spillner wrote:
Simplify if-conditions by removing explicit != 0.
Index: sys/lib/libsa/ufs.c
Thanks for the feedback, Ingo! Sorry for the confusion. I'll spend some more
time with queue(3). :)
On Jul 17, 2014, at 1:21, Ingo Schwarze schwa...@usta.de wrote:
Hi,
Kent R. Spillner wrote on Wed, Jun 04, 2014 at 10:01:12AM -0500:
config(char *) contains a hand-rolled version
*Bump*
On Jul 10, 2014, at 12:34, Kent R. Spillner kspill...@acm.org wrote:
Ping.
On Wed, Jun 04, 2014 at 10:01:12AM -0500, Kent R. Spillner wrote:
config(char *) contains a hand-rolled version of getlist(char *). The only
difference
is that the hand-rolled version includes a NULL
*Bump*
On Jul 10, 2014, at 12:33, Kent R. Spillner kspill...@acm.org wrote:
Ping.
On Thu, May 01, 2014 at 01:22:56PM -0500, Kent R. Spillner wrote:
After sending my previous reply I noticed that you already committed
your diff, so here are my comments again in the form of a proper diff
Ping.
On Wed, Jun 04, 2014 at 10:01:12AM -0500, Kent R. Spillner wrote:
config(char *) contains a hand-rolled version of getlist(char *). The only
difference
is that the hand-rolled version includes a NULL check before the strcmp.
Replace this
with a call to getlist(char *) instead
Ping.
On Thu, May 01, 2014 at 01:22:56PM -0500, Kent R. Spillner wrote:
After sending my previous reply I noticed that you already committed
your diff, so here are my comments again in the form of a proper diff:
* Use NULL instead of casting 0 to pointer types
* Remove unnecessary (char
On Thu, Jul 10, 2014 at 11:50:18AM -0700, Matthew Dempsky wrote:
Diff below defines MAP_ANONYMOUS as a synonym for MAP_ANON, keeping
MAP_ANON as canonical in following BSD heritage.
If MAP_ANON is going to be the canonical value then shouldn't you
change MAP_PRIVATE to MAP_ANON instead of
On Thu, Jul 10, 2014 at 08:29:04AM -0600, Philip Guenther wrote:
CVSROOT: /cvs
Module name: src
Changes by: guent...@cvs.openbsd.org2014/07/10 08:29:03
Modified files:
usr.bin/rdist : common.c config-data.h
Log message:
Assume POSIX: write() takes size_t and
The diff below fixes the spell-o in OpenBSD::PackageRepository::Persistant's
name.
All regress tests still pass on amd64.
On a related note, the regress tests now require user interaction because the
packages created added during the tests are unsigned. I haven't looked into
that yet but am
config(char *) contains a hand-rolled version of getlist(char *). The only
difference
is that the hand-rolled version includes a NULL check before the strcmp.
Replace this
with a call to getlist(char *) instead, and move the NULL check there to
protect other
callers as well.
I think we can
After sending my previous reply I noticed that you already committed
your diff, so here are my comments again in the form of a proper diff:
* Use NULL instead of casting 0 to pointer types
* Remove unnecessary (char *) cast on buf because buf was already
declared as char *
* Simplify if ((rc
Ping.
On Wed, Apr 23, 2014 at 10:15:10PM -0500, Kent R. Spillner wrote:
The diff below removes an unncessary memset() on line 253 of conf.c.
cwm used to support reloading the config file, but okan@ removed that
functionality about a year ago in favor of simply restarting the whole
thing
Ping.
On Wed, Apr 23, 2014 at 12:44:38PM -0500, Kent R. Spillner wrote:
I think I sent this out a long time ago but never followed up on it. :(
According to cwmrc(5) you can configure an autogroup like so:
autogroup group windowname,windowclass
However, parse.y doesn't actually
Ping.
On Wed, Apr 23, 2014 at 10:48:38PM -0500, Kent R. Spillner wrote:
The diff below removes the check for siz == 0 in xmalloc() because it is
unnecessary.
I was curious about the check for siz == 0 in xmalloc() when I first saw
it, so I dug in further and came to the conclusion it's
I think I sent this out a long time ago but never followed up on it. :(
According to cwmrc(5) you can configure an autogroup like so:
autogroup group windowname,windowclass
However, parse.y doesn't actually accept that syntax; you have to put
quotes around windowname,windowclass so they're
Moar bikeshedding:
Index: sdiff.c
===
RCS file: /work/cvsroot/src/usr.bin/sdiff/sdiff.c,v
retrieving revision 1.30
diff -p -u -r1.30 sdiff.c
--- sdiff.c 26 Nov 2013 21:08:12 - 1.30
+++ sdiff.c 23 Apr 2014 17:58:07
The diff below removes an unncessary memset() on line 253 of conf.c.
cwm used to support reloading the config file, but okan@ removed that
functionality about a year ago in favor of simply restarting the whole
thing. So while it used to be possible to call conf_init() multiple
times, it is now
The diff below removes the check for siz == 0 in xmalloc() because it is
unnecessary.
I was curious about the check for siz == 0 in xmalloc() when I first saw
it, so I dug in further and came to the conclusion it's unnecessary:
* It errors out immediately, so aside from the zero size specific
Personally, when sysmerge asks if I'd like to merge something I'd
prefer a little breathing room at the prompt. (Will this break
anything?)
Index: usr.bin/sdiff//sdiff.c
===
RCS file: /work/cvsroot/src/usr.bin/sdiff/sdiff.c,v
Sorry, my webmail client ate the whitespace. mutt is a casualty of the
Valhalla rampage (whines about permissions on /dev/arandom) and kerberosV
removal. I'll rebuild mutt from ports and resubmit.
-Original Message-
From: Kent R. Spillner kspill...@acm.org
Sent: Tuesday, April 22, 2014
rm: /usr/libdata/perl5/site_perl/amd64-openbsd/kerberosV: is a directory
Index: faq/current.html
===
RCS file: /work/cvsroot/www/faq/current.html,v
retrieving revision 1.502
diff -p -u -r1.502 current.html
--- faq/current.html22
+475,7 @@ USAGE:
int_usage();
PROMPT:
putchar('%');
+ putchar(' ');
/* Prompt user again. */
continue;
On Tue, Apr 22, 2014 at 01:33:48PM -0500, Kent R. Spillner wrote:
Sorry, my
I notice a lot of people have suggested use an emulator, as if that had never
occurred to the OpenBSD developers before, but nobody has volunteered to verify
that the available emulators are good enough to actually replace real hardware.
Also, I don't understand why anyone thinks emulation
On Jan 18, 2014, at 16:25, Sia Lang silverlangu...@gmail.com wrote:
If the tests are as good as this project claims them to be, the process
should take exactly one test cycle. If that's the case, then the test regime
suck big time. Logic brother. Logic.
I don't know what tests you're
Still haven't tested, but I also saw:
+password_pwd = malloc(password_pwd_len + 1); /* +1 for \0 */
+
+/* extract the password */
+for ( cnt = 0 ; cnt password_pwd_len ; cnt++ )
+password_pwd[cnt] = password[cnt];
+password_pwd[password_pwd_len] = '\0';
Use
+/* the string generated by yubikey is 44 bytes long */
+password_pwd_len = strlen(password) - 44, cnt;
Haven't tested your latest diff, but I think you have a copy-pasto here (,
cnt).
usr.bin/indent could use a little love, but first it needs some regression
tests. The diff below is my initial attempt at this but before I progress too
much further I'd like some feedback on the general direction I'm going,
especially that Makefile. Is there a way to generalize all of the
One more tiny usr.bin/indent diff: drop some casts that are no
longer needed.
Index: indent.c
===
RCS file: /work/cvsroot/src/usr.bin/indent/indent.c,v
retrieving revision 1.23
diff -p -u -r1.23 indent.c
--- indent.c26 Nov 2013
usr.bin/indent nearly builds clean with -Wall -Werror. The diff below
adds both to CFLAGS and fixes the minor fallout. Arguably, this
doesn't buy us that much, but it makes me feel a little warmer before
I start making too many more changes.
I believe initializing tabs_to_var to zero is safe
we should probably not have more than one argument in our pages for the
.In lines, and, arguably, mandoc should behave the same as groff.
Regardless, the comment doesn't add much. #include ufs/ufs/quota.h is
pretty self-explanatory that it's for UFS quotas. :)
Hey, dude-
The advice is appreciated, but why is it better?
What I need is stability. I now have 5.2-STABLE with the PCI bus number
resource tracking and secondary PCI root segment detection patches
retrieved from CVS. These patches were applied to CVS not long after
tagging 5.2-STABLE.
Hey, dude-
Can you please provide a unified cvs diff?
The first patches I sent last week were cvs diffs, but I saw quilt
was recently added to ports to I switched to using it to manage all
of the patches (I have a few other things I'm working on, too, which
I haven't sent out for review yet).
Hey, dude-
I too would prefer to use nitems, to be consistent with the rest of the
code. Also reduces the number of gratuitous changes, and of course the
size of the diff.
I chose the guard element approach because it leads to the smallest diff,
but I can move the definition of
Hey, dude-
This breaks how name_to_kbfunc is currently being used;
for (iter = 0; iter nitems(name_to_kbfunc); iter++) {
if (strcmp(name_to_kbfunc[iter]-tag, ...
I think you can see where this goes wrong with your patch
applied (hint: passing NULL to strcmp is not advised ;-)
Hey, dude-
This looks really ugly to me. Save argv[] approach below.
That's much nicer. I'm running with your diff here; no regressions.
Thanks again!
Best,
Kent
Hey, dude-
On Sun, Oct 28, 2012 at 04:14:07PM -0400, Okan Demirmen wrote:
A slightly altered version of your diff was applied -thanks!
Cool, thanks!
I noticed you and jmc@ removed the .Xr XClassHint 3 pointer.
What do you think about the addition below?
Basically, I was really confused at
Hey, dude-
will anyone miss reload support? one can always re-exec cwm, or any
other wm for a matter of fact.
I use it every once in a while, but I won't miss it. Out of curiousity,
though, what's the motivation? Too many hacks for handling config
changes that should be visible on the
Hey, dude-
Below is a new diff which only sets no if class matches and name is NULL or
if class and name both match.
Since we must explicitly check for name == NULL I decided to also go ahead
and skip setting no if any previous class-only matches were found.
It's too bad the
Hey, dude-
It seems people do really use that. How about this diff?
In short: do u_exec(cwm) in conf_reload(). This should probably be
u_exec(somewhere-argv0-is-saved).
or use __progname ; also rename from 'reload' to 'restart'
I like this approach a lot. Here's an update which also
Howdy-
It seems people do really use that. How about this diff?
In short: do u_exec(cwm) in conf_reload(). This should probably be
u_exec(somewhere-argv0-is-saved).
or use __progname ; also rename from 'reload' to 'restart'
I like this approach a lot. Here's an update which also
Howdy-
Here is an initial attempt at adding a menu for searching/executing internal
functions. It's useful for those times you want to an unbound function, and
even when the function is bound to some key combo you haven't memorized yet
it can be faster than looking up the keybinding in the
Hey, dude-
-Control automatic window grouping, based on the name and/or class
-properties, where
+Control automatic window grouping based on either the class property,
+or on the name and class properties, where
I like your wording, and I think it is clearer. However, after
this change now
Bump.
On Sun, Feb 19, 2012 at 10:23:41AM -0600, Kent R. Spillner wrote:
Howdy-
This patch adds support for the -L|-P options to ln; see:
http://pubs.opengroup.org/onlinepubs/9699919799/utilities/ln.html
Based on FreeBSD's code manpage
(http://svnweb.freebsd.org/base/head/bin/ln
Hrmmm... Nevermind, this patch is clearly bogus: we shouldn't always set no
when the class matches. :(
I'll fix and resend.
On Oct 25, 2012, at 18:08, Kent R. Spillner kspill...@acm.org wrote:
Merhaba-
I use a couple of different autogroups that match on just the client's class
to help
Merhaba-
Below is a new diff which only sets no if class matches and name is NULL or
if class and name both match.
Since we must explicitly check for name == NULL I decided to also go ahead
and skip setting no if any previous class-only matches were found.
Best,
Kent
Index: group.c
Merhaba-
I use a couple of different autogroups that match on just the client's class
to help me organize my desktop (e.g. Firefox, VirtualBox, XTerm, etc.), but
I also have some special windows in these classes that I want to group
separately (*ahem*). Currently, group_autogroup returns the
Hey, dude-
That's my current pet peeve with the installer, since it doesn't
work the same as
normal boot, and it gets me every time (reboot manually, go
to upgrade, see network
config stall, hit ^C, see # and go 'oh fuck').
What
about overriding the 10s default link-timeout in dhclient.conf
Howdy-
This patch adds support for the -L|-P options to ln; see:
http://pubs.opengroup.org/onlinepubs/9699919799/utilities/ln.html
Based on FreeBSD's code manpage (http://svnweb.freebsd.org/base/head/bin/ln/),
except the regress stuff (which is all my fault).
Best,
Kent
Index: bin/ln/ln.1
64 matches
Mail list logo