15 March 2020, 21:04:35 "Rob Landley" :
> Why are you asking _me_ about this? Post it to the list. The restriction is an
> Android thing, not a toybox thing.
>
> I've said a dozen times that I want to convince android to create a posix
> container within which you can run binaries you build.
>
>
2 March 2020, 10:47:07 "Rob Landley" :
> > Hello, for a few weeks not one message from the toybox mailing list has
> > went to either my inbox or spam. Tried subscribe again, but it's silent
> > too. Please ask your host if they have any letters stuck in the
> > outgoing queue.
>
> This one wound
4 March 2020, 17:34:17 "Rob Landley" :
> > RFC 7208
>
> So your mail server implemented a new thing that doesn't work with the
> existing
> list, and this is my problem. Got it.
Pretty much. Sorry, if there were anything configurable about this on my side I
wouldn't be bothering you. (I do find
3 March 2020, 18:21:59 "Rob Landley" :
> https://support.google.com/a/answer/33786
Exactly. Can you please add this SPF thing to your domain records? Today asked
dreamhost for their postfix log on my address, and my provider for what to do
with that, and they both told me to write you about SPF
2 March 2020, 19:04:57 "Rob Landley" :
> >>> Hello, for a few weeks not one message from the toybox mailing list has
> >>> went to either my inbox or spam. Tried subscribe again, but it's silent
> >>> too. Please ask your host if they have any letters stuck in the
> >>> outgoing queue.
> >>
> >>
Hello, for a few weeks not one message from the toybox mailing list has
went to either my inbox or spam. Tried subscribe again, but it's silent
too. Please ask your host if they have any letters stuck in the
outgoing queue.
___
Toybox mailing list
2b729ee missed sprintf destination for user agent version suffix?
---
>From 6d36d756c76770e38a35b0f1c81dbd5c9bef9f37 Mon Sep 17 00:00:00 2001
From: Denys Nykula
Date: Tue, 11 Feb 2020 14:35:50 +0200
Subject: [PATCH] Fix wget version concat segfault.
---
toys/pending/wget.c | 3 +--
1 f
Hi, 24ae822 scripts/mkroot.sh writes a bad passwd so my nightly couldn't su:
cat > "$ROOT"/etc/passwd << 'EOF' &&
-root::0:0:root:/root:/bin/sh
+echo -e 'root::0:0:root:/root:/bin/sh
> ...
> su: user 'root': Bad file descriptor
What change did you intend in this bit? Also the output started
On Sat, Jan 11, 2020 at 12:00:41AM -0600, Rob Landley wrote:
> How about just:
>
> for i in /root /var/root /; do [ -e $i ] && EXPECT=$i && break; done
Looks fine.
___
Toybox mailing list
Toybox@lists.landley.net
> #!/bin/sh
Shell is very work in progress and not enabled in .config by default,
so build for example https://github.com/MirBSD/mksh into your setup.
> setsid cttyhack /bin/sh
And instead of this, exec /sbin/oneit /bin/sh.
> Quirky Ubuntu had apparently too old
> a gcc version to build.
> scripts/kconfig/lex.zconf.c:890:16: error: ‘EINTR’ undeclared (first use
> in this function)
What's your host libc and where are its headers? In apt that's libc6-dev.
___
Toybox mailing list
Toybox@lists.landley.net
> >> You may want to publish a DNS record
>
> It's dreamhost handling all that automatically.
Most warnings are about your domain settings, manual configuration:
https://help.dreamhost.com/hc/en-us/articles/215414867-How-do-I-add-custom-DNS-records-
> If I had any control over it I'd
> have
> The strange thing is that the binutils, gcc & musl lists don't seem to
> cause gmail to barf nearly as much as the toybox list does.
Mail-tester complains this way about this list server, rating its
likeliness of reaching destination as 32%:
> Listed in SORBS (last 28 days)
gh toy rm did behave like posix
while I was scripting earlier this year and there was some discussion here.
From 02c69847dee7543908f0928d9b170a8f0190c384 Mon Sep 17 00:00:00 2001
From: Denys Nykula
Date: Tue, 12 Nov 2019 16:55:35 +
Subject: [PATCH] Don't rm prompt for nonexistent, just warn.
-
With 81 character wide terminal, toy ls dislikes my etc dir, my home dir
and most other dirs on my netbook, unless given -1 or -l.
etc=`mktemp -d`; echo $etc; cd $etc
touch ImageMagick-7 Muttrc dbus-1 drirc.d dropbear fontconfig git group \
group- lynx.cfg lynx.lss mailcap mime.types mk.conf
ninja's a lot more widely used than you seem to think, but the mistake
is that you're considering it to be a replacement for make. it's not:
it's a replacement for make's back-end. so cmake, for example, can use
ninja now, and every cmake-using project may well be using ninja
> > be able to check "where is the cursor now" after
> > operations, and maybe get a dump of the simulated screen output.
Jarno, you could add vi an option that applies headlessly a key sequence
from a file? For example, a regression test with it would catch this
first line deletion error that I
On Mon, Oct 28, 2019 at 01:37:57AM -0500, Rob Landley wrote:
> On 10/27/19 10:08 AM, Denys Nykula wrote:
> > Script master is acme.js.org/world, comment at the beginning is example
> > how to run.
>
> This is usually the kind of thing people have a repo for somewhere. :)
Co
On Sun, Oct 27, 2019 at 03:27:46PM -0500, Rob Landley wrote:
> On 10/23/19 11:58 AM, Denys Nykula wrote:
> > This, yes. "Other implementations replace the symlink with the dir and
> > keep unpacking files there".
>
> Try now?
Works! Git related packages bui
On Sat, Oct 26, 2019 at 07:06:54PM -0500, Rob Landley wrote:
> On 10/16/19 6:19 PM, Denys Nykula wrote:
> > Yes, NetBSD package repository appears to be easier to bootstrap on Linux
> > than
> > those of GNU/distros. Package recipes are an independent layer from
> > t
> What behavior do you expect here?
> deleting the usr symlink and replacing it with a directory
This, yes. "Other implementations replace the symlink with the dir and
keep unpacking files there".
___
Toybox mailing list
Toybox@lists.landley.net
> the tar "no ../ files outside this directory" check is
> false positiving on "./" for some reason, gotta track that down.
Here's a reproducible toy tar crash, might or might not be connected.
git clone https://github.com/landley/toybox
wget http://musl.cc/x86_64-linux-musl-native.tgz
rm -r
On Wed, Oct 16, 2019 at 01:10:50AM -0500, Rob Landley wrote:
> Static linking against glibc tends to break stuff (like DNS lookup) and
> produce
> TONS of warnings.
Noticed. That's unfortunate.
> I had a linux from scratch based one of these, which I need to reproduce. But
>
> > Passing through HOSTCC makes sense, but passing through CFLAGS... is it
> > HOSTCFLAGS or CROSS_CFLAGS?
>
> Except now that I sit down to do it (in the kernel HOSTCFLAGS is a variable
> and
> CFLAGS seems to go to the cross compiler, modulo whatever EXTRA_CFLAGS and
> HOST_EXTRACFLAGS do),
> saw one of the Toybox talks and wanted to see how I could contribute
Completing landley.net/toybox/cleanup.html on toys/pending/{dhcp,route}.c
to help them out of pending would be most demanded I think.
___
Toybox mailing list
Toybox@lists.landley.net
Find stopped falling back to current dir without args:
# toybox find
r%': No such file or directory
(Another system) Segmentation fault
Because the inline array of "." becomes garbage outside its if.
From af4864a98dec8411bd58dec2543c2ec39591b301 Mon Sep 17 00:00:00 2001
From: Denys Ny
6f580999cf52f990 Mon Sep 17 00:00:00 2001
From: Denys Nykula
Date: Tue, 1 Oct 2019 20:32:55 +0300
Subject: [PATCH] Airlock XFCE.
Learn to bootstrap pkgsrc off toybox and build a coding desktop.
Remember to install kernel headers on host. And to copy host CFLAGS, or
armv7l binaries don't w
Nevermind, the very next line in the mkroot script explains how I needn't
have patched it, I just needed to pass variables this way:
make CROSS_COMPILE=$cc HOSTCC="${cc}cc -static" PATH="`dirname $cc`:$PATH" \
defconfig
...
make CROSS_COMPILE=$cc PATH="`dirname $cc`:$PATH" root \
musl toolchain from a glibc host without its own toolchain, I
append -static to both CFLAGS and full compiler path in HOSTCC.
From a8da293686dc539cf6dd9fa6b1f0dfba52fd8513 Mon Sep 17 00:00:00 2001
From: Denys Nykula
Date: Tue, 26 Sep 2019 22:58:23 +0300
Subject: [PATCH] Pass cflags and hostcc
output in wget because the nonexistent constant fails single
builds and mkroot.
From 6f47711ada5b1c43d7e518a06dd77516763e163e Mon Sep 17 00:00:00 2001
From: Denys Nykula
Date: Mon, 23 Sep 2019 05:15:02 +0300
Subject: [PATCH] Fix ftpget download, call from wget, noop gzip -n.
Investigate why downloads
Got a weird tar which confuses toybox. Comparing to the one toybox makes
from same files, notice the missing slash in directory, spaces in size where
nulls usually are, and lettered username beside uid. Has this any relation
to the matter of policy thing mentioned in tar.c? BLFS of last year says
Portage can't upgrade toybox with sed linked to toybox:
# make oldconfig
scripts/genconfig.sh
sed: bad pattern '/^\*\//q;/^config [A-Z]/,$p'@9 (;)
kconfig/conf -o Config.in
Config.in:5: can't open file "generated/Config.in"
make: *** [kconfig/Makefile:18: oldconfig] Error 1
Pkgsrc tries to
> the stuff rob's talking about is for the linux console before you
> start X11. (and is more like
> https://www.x.org/archive/X11R6.8.1/doc/xmodmap.1.html in the X11
> world.)
Requested setkeycodes toy would still be useful, I remap laptops'
chronically broken Ctrl to CapsLock, setkeycodes 3a
> throw symlinks all over the place if you like
Contrarily, I make fewer symlinks.
> (And they've
> apprently rewritten the documentation in rust, which I was unaware was an
> option?)
RST is ReStructured Text. Sort of like Markdown. Has tooling for tables of
contents and interlinks, generates
Hi Elliott, maintainers of Android apps with native executables are tweeting to
Rob but Android is your area of work so please comment:
https://twitter.com/topjohnwu/status/1171308880904646656
> I'm curious about @landley's (author of toybox) thoughts on the direction
> Android is heading. He had
> It's probably easier to maintain a makefile for llvm, cmake is terrible.
Unsure if easier to maintain, but easier to make, yes. Will try and see.
> If you don't mind /share and /include at the top level.
Looking up files in shell a lot, I do want /local and these dirs in root. Not
fond of
> Ideally I'd do http://landley.net/aboriginal/about.html#hairball and build
> debian and such under toybox, but
> ...
Portage with musl and some netbook stuff builds here weekly, it mostly works
under toybox, and where it doesn't because of GNU extensions I complain or send
patches to you and
> > which awk 2>&1 >/dev/null || ln -fsv nawk /bin/awk
> > which vi 2>&1 >/dev/null || ln -fsv vim /bin/vi
>
> Those last two are on the todo list. But after toysh and route, which mkroot
> needs. (And vi might be after make, and promoting half the stuff currently in
> pending.)
Implementing sh
> Which is where we come back to the question, what and how are others
> dealing with this problem and is there a need for a simple and basic
> installer inside toybox?
In gentoo I have this /etc/portage/bashrc hook symlink toybox to every
command missing or linked to nowhere, after package
> when i rewrite the formatting
> stuff based on the one i wrote a while ago, do you want that as part
> of man or would you rather have a separate [man-only] nroff that man
> calls?
Part, if it's not general purpose. Thanks for continuing the toy.
___
Background infinite loop starting with the system, checking
/sys/class/power_supply/*/charge_now once a minute and running
echo mem >/sys/power/state to re-suspend if its value is below a
certain point and previous suspend according to system logs was
more than N minutes ago? Might need some
> enabled toysh in your config
> didn't select "Disable recursive execution"
> I'm not going to special case env.
Both settings are intentional. Had learned that one of common env usages is to
specifically to opt out of shell builtins and force PATH lookup for one
command, like in answers below.
exec(argv);
Is it clean to set toys.stacktop = 0 in env, so it runs my system shell?
How to approach writing a test for this?
---
>From 1ef072946aa860bfc0a762782ab4f82d4734b693 Mon Sep 17 00:00:00 2001
From: Denys Nykula
Date: Fri, 5 Jul 2019 14:27:24 +
Subject: [PATCH] Have env exec, not rec
Thanks for the review! Addresed all points. Should've written more test
cases and caught the need to handle empty string and more slashes. Yes,
musl rename() seems alright, the problem is instead with getbasename().
From fc92db71a1cc11cb152cb4aa90d589264a54c0ac Mon Sep 17 00:00:00 2001
From: Denys
Leaving the Makepost address because somebody abused my email provider
and got Mozilla and GitHub really unhappy.
---
>From 91c5ba7d250188afa5b40631b15bc3a778691288 Mon Sep 17 00:00:00 2001
From: Denys Nykula
Date: Sat, 22 Jun 2019 14:57:59 +0300
Subject: [PATCH] Fix mv with trailing sl
45 matches
Mail list logo