As noticed by landry, going to the ports tree for dependencies can fail
hilariously in case of a PORTSDIR_PATH.
Most specifically, assuming you are building something like firefox-102
in one of your ports directory, when building the debug package,
dependency look-up in pkg_create is going to
On Wed, Jun 08, 2022 at 09:44:19PM +0200, Marc Espie wrote:
> Naddy told me about an app that wants a gnu-m4 extension that
> requires >9 arguments to macros.
>
> I wrote a very quick patch that seems to do the work. There are probably
> lots of kinks to work out, it's been v
Naddy told me about an app that wants a gnu-m4 extension that
requires >9 arguments to macros.
I wrote a very quick patch that seems to do the work. There are probably
lots of kinks to work out, it's been very lightly tested.
(in particular, I haven't looked at stuff like $* and friends yet,
I've conducted a quick audit of the 300 or so uses of @extraunexec
in pkg_add.
All but 5 are actually glorified versions of
rm -rf
I propose eventually replacing them with
@extraglob pattern
with several advantages:
- we have file names that we know
- perl itself can expand the glob and
Due to some slowdown in updating texlive, I've looked up duplicate files inside
packages.
Most packages are okay, a few are not...
The following list shows the maximum number of times
a file appears in a given package (just looking at its checksum)
Note sdcc, which stores a crazy number of
On Sat, May 21, 2022 at 07:47:40AM -0500, Luke Small wrote:
> Perhaps it was rude sending off list stuff to the list. Your email sounded
> "less than friendly" and more of a professional challenge that you were
> definitely in the works to produce; much like Damien Miller’s challenge to
> prove
On Sun, May 15, 2022 at 08:46:00PM -0600, Theo de Raadt wrote:
> You'll turn into a smurf before you finish auditing xterm perfectly.
Should be in fortune ;)
Random generators have been analyzed for years.
Pick up "Concrete mathematics" by Don E. Knuth, read it, then come back
with actual science.
In the old pkg_add model (prior to 7.1), we create tmp files anyway, so the
code is fine.
7.1 and later, we try to avoid creating temp files, so a better name matching
correspondence would be cool
What this patch does:
the old code creates a simple hash where we associate ONE packing-list
On Wed, May 04, 2022 at 07:43:35AM -0400, Bryan Steele wrote:
> On Wed, May 04, 2022 at 01:20:10PM +0200, Frederic Cambus wrote:
> > Hi tech@,
> >
> > The base system includes the compiler-rt profile library for
> > source-based code coverage.
> >
> > So here is a diff to document support in
On Thu, Apr 14, 2022 at 04:28:37PM +0200, Alexander Bluhm wrote:
> Hi,
>
> When kbd -l is executed as regular user, it fails silently.
>
> $ kbd -l
> $ echo $?
> 0
>
> Error handling is a bit tricky. We want the first error if no
> device is available.
>
> $ ./kbd -l
> kbd: open
I got a new slightly different one.
Specifically, the old one has a "fun" failure mode in case
an old package got broken and is missing files.
The heuristics correctly deduces that the files are not
there, so it extracts in-place... then pkg_delete promptly
deletes what it believes to be files
I've changed pkg_add for updates and tied files a while back, which resulted
in creating a lot of temp files. The logic was complicated, and I didn't
want to make it more complex at the time until most of the issues were fixed.
Then we got tags, so that most packing-lists are innocuous these
I've been reading the newer C standard
(n2310.pdf to be specific).
That thing is scary.
I believe that, in the long haul, we will have to use the Atomic types stuff
that's apparently part of the C standard now.
Yeah, the switch to multi-threading made the standard way more complicated.
Sequence
On Mon, Feb 21, 2022 at 05:36:16PM +0900, Yuichiro NAITO wrote:
> Of course, all programs compiled without '-pg' work fine for me.
> I found this issue when I profile my application with gprof(1).
> For example, following example C source code fails to execute on OpenBSD 7.0.
Profile is partly
On Mon, Feb 14, 2022 at 11:37:55AM +0100, Marc Espie wrote:
> - you don't need to call unlock_db. When the process dies, the fd dies with
> it, and the lock.
>
> - I've committed glue to PackageInfo.pm to make lock_db self-contained
> (it will pull BaseState in on an "as neede
- you don't need to call unlock_db. When the process dies, the fd dies with
it, and the lock.
- I've committed glue to PackageInfo.pm to make lock_db self-contained
(it will pull BaseState in on an "as needed" basis, which also comes in
handy for other locks prior to having a state)
Index:
On Wed, Feb 09, 2022 at 07:30:46PM -0800, Andrew Hewus Fresh wrote:
> I was reminded that fw_update(8) updates the package database without
> locking currently. That can cause issues when running it concurrently
> with pkg_add, for example starting `pkg_add -u` in one terminal and
> `sysupgrade`
On Tue, Feb 08, 2022 at 01:59:51PM +, Stuart Henderson wrote:
> > > I think we should start updating Devel::PPPort in base. Then we
> > > can regenerate other ppport.h in base and ports. Note that some ports
> > > need newer Devel::PPPort than we have now. So updating seems
> > >
Or we can automate this with something like this:
Index: perl.port.mk
===
RCS file: /cvs/ports/infrastructure/mk/perl.port.mk,v
retrieving revision 1.32
diff -u -p -r1.32 perl.port.mk
--- perl.port.mk12 Dec 2021 19:25:39
On Fri, Jan 21, 2022 at 02:12:25PM +0100, Alexander Bluhm wrote:
> Hi,
>
> Since clang 13 each Perl or Perl XS module compile spits out a lot
> of -Wcompound-token-split-by-macro warnings. E.g. p5-Net-SSLeay
> produces 3882 warnings generated. You cannot spot anything useful.
> The problem is
On Mon, Nov 15, 2021 at 05:58:38PM +0100, Ingo Schwarze wrote:
> I don't know. A fairly reliable way to create security risks is
> complexity. Apart from the erratic run time behaviour that is likely to
> trip up sysadmins - LC_COLLATE can change the collation sequence even
> among ASCII
On Mon, Nov 15, 2021 at 03:43:47PM +0100, Jan Stary wrote:
> Here's a try (see below);
> one sentence one line while here.
>
> I would also replace 'results' with 'result' everywhere,
> but I am not a native speaker.
>
> Jan
>
>
> On Nov 10 18:46:08, h...@stare.cz wrote:
> > On Nov 10
We disabled the checksum check because it was rather expensive, but actually
we also store timestamps in packing-lists
This is a first draft at a better check. What this does:
- compare the fs timestamp with the packing-list timestamp. This should be
very cheap in most cases, since we're already
On Tue, Oct 19, 2021 at 10:42:04AM +0900, Yuichiro NAITO wrote:
> Following patch changes pkg_add to return a error code,
> if a package name is wrong.
>
> diff --git a/usr.sbin/pkg_add/OpenBSD/AddDelete.pm
> b/usr.sbin/pkg_add/OpenBSD/AddDelete.pm
> index 7a968cbf05d..39bee874ff1 100644
> ---
On Fri, May 28, 2021 at 06:56:42PM +0200, Marc Espie wrote:
> On Fri, May 28, 2021 at 07:55:39AM -0700, Chris Cappuccio wrote:
> > I tried to compile librdkafka on OpenBSD 6.5 (clang 7.0.1) and clang
> > compiled
> > the __thread parts with some built-in mechanism.
On Fri, May 28, 2021 at 07:55:39AM -0700, Chris Cappuccio wrote:
> I tried to compile librdkafka on OpenBSD 6.5 (clang 7.0.1) and clang compiled
> the __thread parts with some built-in mechanism. I upgraded the system to
> OpenBSD 6.9 and TLS is no longer supported by the in-tree clang. Was this
>
So, is there a way to get a work-around ? or should we push for proper
tls sooner than later ?
It always made me queasy having all those programs segfaulting on exit.
Not sure it's exploitable, but giving people the impression it's okay
to segfault.
I thought it might be bad code in the
On Fri, Apr 30, 2021 at 04:54:57PM +0200, Christian Weisgerber wrote:
> Marc Espie:
>
> > Until a patch from naddy, I wasn't even aware of getopts in sh(1)
>
> Let's start the discussion with this instead.
>
> This puts the deprecation notice in getopt.1 in a pro
On Fri, Apr 30, 2021 at 03:28:42PM +0100, Jason McIntyre wrote:
> my argument boils down to: sh(1) is small and has no examples. adding
> them changes the (deliberate) nature of the page. ksh(1) is what you
> read when you can;t get to sleep.
>
> why is it wrong to add your example to ksh(1)? why
On Fri, Apr 30, 2021 at 03:28:42PM +0100, Jason McIntyre wrote:
> On Fri, Apr 30, 2021 at 04:07:55PM +0200, Marc Espie wrote:
> > On Fri, Apr 30, 2021 at 02:44:01PM +0100, Jason McIntyre wrote:
> > > On Fri, Apr 30, 2021 at 11:54:16AM +0200, Marc Espie wrote:
> > > &g
On Fri, Apr 30, 2021 at 02:44:01PM +0100, Jason McIntyre wrote:
> On Fri, Apr 30, 2021 at 11:54:16AM +0200, Marc Espie wrote:
> > Until a patch from naddy, I wasn't even aware of getopts in sh(1)
> >
> > Unless I made some mistakes, this translates the example in
On Fri, Apr 30, 2021 at 12:03:00PM +0100, Raf Czlonka wrote:
> Hi Mark,
>
> You and me both ;^)
>
> Until recently, I thought that getopt(1) was POSIX, whereas it is
> in fact getopts(1), and it is not a shell built-in there, but a
> utility[0].
Nope, it is a shell built-in... the "wording" of
Until a patch from naddy, I wasn't even aware of getopts in sh(1)
Unless I made some mistakes, this translates the example in getopt(1)
manpage.
It's likely some stronger wording might be adequate, I suspect some
of the BUGS section in getopt(1) does not apply to the sh(1) built-in.
Index:
@@ -0,0 +1,122 @@
+.\"$OpenBSD$
+.\"
+.\" Copyright (c) 2008 Marc Espie
+.\"
+.\" All rights reserved.
+.\"
+.\" Redistribution and use in source and binary forms, with or without
+.\" modification, are permitted provided that the following conditions
+.\&quo
I do think we want to write something specific for .for loop variables
which are actually very special compared to the rest.
I'm not incredibly happy with the way netbsd explains it, not surprisingly.
On Sun, Oct 11, 2020 at 06:03:55PM +0200, Landry Breuil wrote:
> On Sun, Oct 11, 2020 at 05:03:31PM +0200, Marc Espie wrote:
> > On Sun, Oct 11, 2020 at 04:53:18PM +0200, Christian Weisgerber wrote:
> > > Marc Espie:
> > >
> > > > The new design:
&
On Sun, Oct 11, 2020 at 06:03:55PM +0200, Landry Breuil wrote:
> On Sun, Oct 11, 2020 at 05:03:31PM +0200, Marc Espie wrote:
> > On Sun, Oct 11, 2020 at 04:53:18PM +0200, Christian Weisgerber wrote:
> > > Marc Espie:
> > >
> > > > The new design:
&
On Sun, Oct 11, 2020 at 04:53:18PM +0200, Christian Weisgerber wrote:
> Marc Espie:
>
> > The new design:
> >
> > The idea behind "libset" is to be able to specify a "set" of wantlib that
> > corresponds to our package, AND to just write
I want to introduce a new mechanism (libsets) to supplement wantlib...
Note that the keyword already exists in pkg_add.
the story so far:
wantlib evolved from lib-depends
basically, you tell a given package that it requires some libs, (WANTLIB)
and pkg_create will register those libs in the
On Wed, Jun 24, 2020 at 05:32:57PM +0100, Jason McIntyre wrote:
> On Tue, Jun 23, 2020 at 07:28:09PM -0400, sven falempin wrote:
> > Dear readers,
> >
> > It may not be very obvious that 'dry run' mode of pkg_add
> > actually downloads packages.
> > It is a good feature and maybe the pkg_add man
On Wed, Jun 17, 2020 at 11:34:20AM +0200, Solene Rapenne wrote:
> I propose a small diff for pkg_add when using http/https mirrors.
> Don't wait 30 seconds for the ftp process to stop when closing
> file handler, send SIGHUP immediately after closing the file handler.
>
> Running pkg_add -u
On Tue, Jun 16, 2020 at 08:43:02PM +0100, Raf Czlonka wrote:
> Ping.
>
> CC'ing espie@ as he committed the initial code.
>
> Cheers,
>
> Raf
>
> On Sun, Jun 07, 2020 at 07:30:39PM BST, Raf Czlonka wrote:
> > Hi all,
> >
> > I've been running openbsd-backgrounds on all of my desktop machines
On Thu, Jun 11, 2020 at 07:55:44AM -0600, Theo de Raadt wrote:
> I think this conversation about major numbers, and now minor numbers,
> is incredibly silly.
>
> Do we have a major number on the open(2) system call? No, and the main
> reason is because the code isn't a draft, it was completed.
>
On Thu, Jun 11, 2020 at 03:42:27PM +0300, Paul Irofti wrote:
> On Thu, Jun 11, 2020 at 02:05:44PM +0200, Marc Espie wrote:
> > On Thu, Jun 11, 2020 at 01:13:07PM +0300, Paul Irofti wrote:
> > > On 2020-06-11 02:46, Christian Weisgerber wrote:
> > > > Paul Irofti:
>
On Thu, Jun 11, 2020 at 01:13:07PM +0300, Paul Irofti wrote:
> On 2020-06-11 02:46, Christian Weisgerber wrote:
> > Paul Irofti:
> >
> > > This iteration of the diff adds bounds checking for tk_user and moves
> > > the usertc.c stub to every arch in libc as recommanded by deraadt@.
> > > It also
On Wed, Jun 10, 2020 at 05:38:36PM +1000, Jonathan Gray wrote:
> On Wed, Jun 10, 2020 at 09:19:47AM +0200, Martin Pieuchot wrote:
> > On 09/06/20(Tue) 20:19, jo...@armadilloaerospace.com wrote:
> > > Looking for some guidance to avoid proposing any unpopular diffs.
> > >
> > > Style(9) says not
On Tue, Jun 09, 2020 at 08:19:53PM -0700, jo...@armadilloaerospace.com wrote:
> Also, style(9) says that prototypes should not have variable names
> associated with the types. I try to use good names in the headers
> for documentation purposes; what is the thinking behind the rule?
function
On Mon, Jun 01, 2020 at 03:37:45PM -0600, Anthony J. Bentley wrote:
> Marc Espie writes:
> > > To setup the right alt key as compose, you can either:
> > >
> > > - run 'setxkbmap -option compose:ralt' somewhere in your session
> > > startup script
> &
On Mon, Jun 01, 2020 at 03:46:09PM +0200, Matthieu Herrb wrote:
> On Mon, Jun 01, 2020 at 03:28:52PM +0200, Stéphane Aulery wrote:
> > Hello,
> >
> > Le 01/06/2020 14:55, Matthieu Herrb a écrit :
> > >
> > > >
> > > > (I have just tried with a test user with nothing configured besides
> > > >
I occasionnally use setxkbmap 'us(intl)' in order to have diacritics
as dead keys.
I have LC_CTYPE=en_US.UTF-8
If I type 'c
I get a ç in programs like firefox or chrome, BUT I get
ć in xterm ?
why are things different ? which program is right ?
(I have just tried with a test user with
On Sat, May 30, 2020 at 05:10:49PM +0200, Jeremie Courreges-Anglas wrote:
> On Sat, May 30 2020, Marc Espie wrote:
> >> - have some magic I don't know in ELF handling that would allow to either
> >> tweak the default location/introduce ${WRKOBJDIR} in that debug info.
>
> - have some magic I don't know in ELF handling that would allow to either
> tweak the default location/introduce ${WRKOBJDIR} in that debug info.
Thinking some more about it, that 3rd option is possibly not that
far-fetched
we do pass most debug-info thru dwz for shrinkage, so there's one
In a trace:
> > > #3 0x15e48c95459e in WebVfx::shutdown ()
> > > at /usr/obj/ports/webvfx-1.2.0/webvfx-1.2.0/webvfx/webvfx.cpp:193
Now, this is NOT the default location for WRKOBJDIR, but we are shipping
packages with debug information...
This could be a bit cumbersome, if in order to
On Tue, May 26, 2020 at 12:13:02AM +0200, Ingo Schwarze wrote:
> Regarding the FreeBSD headers, i like them even less. There are both
> terrible design choices and terrible implementation choices. Regarding
> design, it appears that you *must* #include after other
> headers - if you include it
On Fri, May 29, 2020 at 03:48:41PM +0200, Ingo Schwarze wrote:
> Hi Marc,
>
> do i understand correctly that you intend to say: there are
> legitimate reasons to use these xlocale.h functions in portable
> software, in particular in portable libraries, so we should probably
> have them all, or
On Thu, May 28, 2020 at 05:44:31PM +0300, Paul Irofti wrote:
> Hi,
>
> Here is a new iteration of the diff which includes support for MD high
> resolution clocks. Currently only implements TSC on amd64. If the
> MD function is not defined, it fallsback to the syscall.
>
> There is the question
On Tue, May 26, 2020 at 12:13:02AM +0200, Ingo Schwarze wrote:
> Hi,
>
> my impression is there are two totally independent topics in this
> thread.
>
> 1. Whether and how to make things safer, ideally terminating the
> program in a controlled manner, if it uses functions that are
> not
On Fri, May 22, 2020 at 02:31:38PM +0200, Mark Kettenis wrote:
> > Date: Fri, 22 May 2020 13:50:44 +0200
> > From: Marc Espie
> >
> > I've had to neuter some setlocale() in mlt, since our code is
> > definitely NOT thread-safe. No biggie, since we do not have su
On Fri, May 22, 2020 at 04:02:31PM +0200, Mark Kettenis wrote:
> > From: "Todd C. Miller"
> > Date: Fri, 22 May 2020 07:23:55 -0600
> >
> > On Fri, 22 May 2020 14:57:11 +0200, Marc Espie wrote:
> >
> > > From a security standpoint, is t
On Fri, May 22, 2020 at 02:31:38PM +0200, Mark Kettenis wrote:
> > Date: Fri, 22 May 2020 13:50:44 +0200
> > From: Marc Espie
> >
> > I've had to neuter some setlocale() in mlt, since our code is
> > definitely NOT thread-safe. No biggie, since we do not have su
I've had to neuter some setlocale() in mlt, since our code is definitely NOT
thread-safe. No biggie, since we do not have support for LC_NUMERIC right now.
I think we might want the thread-specific functions, namely stuff like
strtod_l, or sprintf_l and friends.
Even if they do not do anything
I haven't used wsmoused in ages, but I tend to think ideas like (by default)
scaling to the font size have great merit.
There should be a debugging toggle (maybe) to get raw values.
It should work out-of-the-box preferably. ;)
I do think newcomers will tend to miss examples, even if /etc/examples is
somewhat documented elsewhere.
Also, queries to mandocdb will probably benefit... say "what commands do
have an example configuration file ?" kind of what you should be able to
ask with advanced queries.
On Thu, Apr 09, 2020 at 06:33:25PM +0200, Marc Espie wrote:
> Our make (bsd make) has "alternate" dependency operators :: and !
>
> Twenty years ago, we made it so that ambiguous constructs like
> File::Find.3p: File/Find.pm
>
> would no longer be ambiguous thanks to
Our make (bsd make) has "alternate" dependency operators :: and !
Twenty years ago, we made it so that ambiguous constructs like
File::Find.3p: File/Find.pm
would no longer be ambiguous thanks to the added space.
Now, I believe that ! should be trumped by : if both are present in the
same
On Thu, Apr 09, 2020 at 02:44:14PM +0200, Jeremie Courreges-Anglas wrote:
> On Thu, Apr 09 2020, Jeremie Courreges-Anglas wrote:
> > find(1) uses ARG_MAX to compute the maximum space it can pass to
> > execve(2). This doesn't fly if userland and the kernel don't agree, as
> > noticed by some
There's some inconsistency:
lists/xserv/mi:./usr/X11R6/share/aclocal/xorg-macros.m4
lists/xshare/mi:./usr/X11R6/lib/pkgconfig/xorg-macros.pc
both should be in xshare
This does explain the failures of tightvnc on some bulk machines.
By default, proot does NOT copy xserv in the chroot, and I
On Sun, Mar 22, 2020 at 01:09:17PM +0100, Denis Fondras wrote:
> On Sun, Mar 22, 2020 at 11:41:55AM +0100, Marc Espie wrote:
> > If tar can't create intermediate directories due to permission
> > issues, the resulting message is confusing:
> >
> > ./tar xf gcc.tar
If tar can't create intermediate directories due to permission
issues, the resulting message is confusing:
./tar xf gcc.tar gcc-8.3.0/include/obstack.h
tar: Unable to create gcc-8.3.0/include/obstack.h: No such file or directory
(here I have gcc-8.3.0 owned by root and no permissions)
The
On Sun, Feb 09, 2020 at 01:37:33PM +, Jason McIntyre wrote:
> On Sun, Feb 09, 2020 at 02:27:23PM +0100, Marc Espie wrote:
> > On Sun, Feb 09, 2020 at 12:41:43AM +0100, Sebastian Benoit wrote:
> > > Ingo Schwarze(schwa...@usta.de) on 2020.02.09 00:33:06 +0100:
> > > &
On Sun, Feb 09, 2020 at 12:41:43AM +0100, Sebastian Benoit wrote:
> Ingo Schwarze(schwa...@usta.de) on 2020.02.09 00:33:06 +0100:
> > Hi,
> >
> > Jason McIntyre wrote on Sat, Feb 08, 2020 at 10:15:08PM +:
> >
> > > - i'm ok with adding the path to these files to a FILES section
> >
> > So,
On Sat, Feb 01, 2020 at 04:57:26PM +0100, Ingo Schwarze wrote:
> Hi Jeremie,
>
> Jeremie Courreges-Anglas wrote on Sat, Feb 01, 2020 at 01:37:32PM +0100:
> > On Fri, Jan 31 2020, Ingo Schwarze wrote:
> >> ngc...@gmail.com wrote on Fri, Jan 31, 2020 at 10:14:52PM +0900:
>
> >>> Reduce scope of a
On Thu, Jan 30, 2020 at 11:14:37AM +, Stuart Henderson wrote:
> Putting this here for discussion... good idea? bad idea? does it need
> more checks for expected file contents?
>
> Index: sysupgrade.sh
> ===
> RCS file:
So, basically, if we start arbitrary commands, then
the classic loop
/* Wait for the child to exit. */
while (waitpid(cpid, , 0) == -1 && errno == EINTR)
continue;
is not quite enough.
See the small note in manpages (not only us, but everyone):
WIFSTOPPED(status)
So I did a full bulk build with the following diff.
No failure due to old keyword.
This contains exactly:
- abort directly instead of setting pgn->must_make = false
- tag as "SPECIAL_DEPRECATED" all old keywords
- remove the conditionals with those OP. We just keep OP_INVISIBLE because
it's
More specifically, I'm running with this patch
The specific choice of keywords to deprecate is up for grabs, the
infrastructure for actually being able to error out on these keywords is
probably something I should commit anyhow.
diff --git a/gnode.h b/gnode.h
index 283fead..04e5462 100644
---
The convoluted logic that resets must_make does not make any sense to me.
It's just as simple to set built_status to ABORTED directly.
Note that in the compat make case, there are two instances of using
must_make left, one in arch.c, which I have yet to understand, and one
in
On Fri, Jan 10, 2020 at 10:52:03PM +0100, Alexander Bluhm wrote:
> On Fri, Jan 10, 2020 at 01:58:47PM +0100, Marc Espie wrote:
> > Bleh, I forgot to synch two patches I already committed. Here's a patch
> > that applies cleanly.
>
> I did run this make through a full regress
On Thu, Jan 09, 2020 at 01:09:59PM +0100, Marc Espie wrote:
> So my development branch is getting a bit too far
> ahead compared to what's committed.
>
> Here's a big diff to test. Comments as to what's going on
> and the changes this contains:
>
> - buffer changes: add s
So my development branch is getting a bit too far
ahead compared to what's committed.
Here's a big diff to test. Comments as to what's going on
and the changes this contains:
- buffer changes: add support for "permanent static buffers"
that don't get reinit'd all the time (that's a mostly
-q is just plain broken in compat mode, it's easy to fix though.
The second issue is that, if a Makefile as an .END target, that one
will always be run, even in query mode, contrary to .BEGIN.
So let's fix that as well.
Fairly obvious patch
Index: compat.c
The only way errorJobs ever gets filled is by engine.c
around line 680 with a snippet that reads
if (!keepgoing) {
if (!silent)
printf("\n");
job->next = errorJobs;
On Thu, Jan 02, 2020 at 11:19:30AM +0100, Ingo Schwarze wrote:
> Hi Kurt,
>
> Kurt Mosiejczuk wrote on Wed, Jan 01, 2020 at 08:21:04PM -0500:
>
> > cp(1) uses -R for recursive copy. scp(1) uses -r. This diff adds -R
> > as an alias for -r to scp(1) for those assuming consistency with cp(1).
>
>
We don't normally hit this thanks to the "expensive job" heuristics
(meaning that we don't start the second target while the first is
building), but still there is a race, and if for whatever reason we
remove that heuristics we hit it fairly hard.
The patch is fairly trivial, just run both
If we look at job_handle_status, we see there's one case where it
doesn't place the job on the errorJob list, but doesn't try to reclaim
the memory.
Index: job.c
===
RCS file: /cvs/src/usr.bin/make/job.c,v
retrieving revision 1.143
So, a job is the series of shell commands run by make
a:
cmd1
expensive_cmd2
cmd3
so the granularity of expensive should be at the command level,
not the job level.
So, instead of running the code that may start new jobs at
the end of remove_job, let's run it after each
On Sat, Dec 28, 2019 at 07:48:45PM -0500, Bryan Steele wrote:
> With FreeDB announcing[0] that the service will be shutting down as of
> March 31st of 2020, and the only other alternative (MusicBrainz) already
> having shutdown their freedb/cddb gateway in favour of their own API
> early this
On Fri, Dec 13, 2019 at 02:19:26PM +0100, Matija Skala wrote:
> Dne petek, 13. december 2019 ob 13:34:57 CET je Marc Espie napisal(a):
> > On Fri, Dec 13, 2019 at 08:45:08AM +0100, Jasper Lievisse Adriaanse wrote:
> > > Hello Matija,
> > >
> > > Could you ple
On Fri, Dec 13, 2019 at 08:45:08AM +0100, Jasper Lievisse Adriaanse wrote:
> Hello Matija,
>
> Could you please provide a testcase for inclusion in
> src/regress/usr.bin/pkg-config too?
> Also, is there a particular port or pkg-config file in the wild that you ran
> into which exhibits this
On Thu, Dec 12, 2019 at 02:51:18AM -0700, Theo de Raadt wrote:
> Marc Espie wrote:
>
> > On Wed, Dec 11, 2019 at 01:15:40PM -0700, Theo de Raadt wrote:
> > > I'm not sure I understand the goal of the signal handler.
> > >
> > > sdiff is moving
On Wed, Dec 11, 2019 at 01:15:40PM -0700, Theo de Raadt wrote:
> I'm not sure I understand the goal of the signal handler.
>
> sdiff is moving forward through the file, only. If you are in a pager,
> you want to increase the width for the later output not yet visible?
> the normal way one does
This patch allows sdiff to auto-detect tty width,
by passing -w auto
More importantly, sdiff will adjust its variables on the fly for
subsequent lines (I don't know if redrawing the current
lines when the tty changes is advisable).
It's pretty straightforward, tested thru sysmerge.
Doesn't seem
On Mon, Dec 02, 2019 at 08:31:02AM +, Miod Vallat wrote:
> grep(1), when invoked with the -R option but no path, displays a
> "recursive search of stdin" warning and acts as if -R had not been
> specified.
>
> GNU grep, in that case, will perform a recursive search in the current
> directory,
This should get rid of the weird error (don't know how to make
.../all/debug-*.tgz) when switching from !DEBUG_PACKAGES to DEBUG_PACKAGES.
Basically, this introduces a "build two targets at once" in bsd.port.mk.
make has some glue to figure out whether this is a "duplicate the work" or
"one
On Wed, Nov 27, 2019 at 06:21:01PM +0200, Lauri Tirkkonen wrote:
> All that said I do understand if there is reluctance to merge the
> jobserver stuff since it doesn't actually help the current situation in
> most cases. Nevertheless it has been personally beneficial to me in
> identifying areas
On Wed, Nov 27, 2019 at 05:31:48PM +0200, Lauri Tirkkonen wrote:
> I'll give you some more background that I maybe should have given
> earlier already: in my hobby OS Unleashed, we use bmake and earlier did
> some (slightly hacky) modifications to subdir.mk to enable paralellizing
> jobs in
On Fri, Nov 15, 2019 at 03:29:29PM +0200, Lauri Tirkkonen wrote:
> On Fri, Nov 15 2019 15:24:57 +0200, Lauri Tirkkonen wrote:
> > Your points are valid and I agree with them completely. There are
> > clearly problems with lock contention,
>
> and I should mention here that I would probably not
On Wed, Nov 27, 2019 at 12:42:48PM +0200, Lauri Tirkkonen wrote:
> In a lesson to always proof-read *before* sending your message:
>
> On Wed, Nov 27 2019 12:31:51 +0200, Lauri Tirkkonen wrote:
> > - a full release build with the jobserver enabled passed after
> >disabling
On Mon, Nov 25, 2019 at 04:11:53PM +0200, Paul Irofti wrote:
> Hi,
>
> Few people complained (hi landry@!) that stripped binaries are slightly
> larger now than they used to be when debug packages are enabled.
>
> My investigations show that this is because objcopy --strip-debug is
> less
1 - 100 of 548 matches
Mail list logo