Re: [DNG] PHP 8.1 depends on systemd?

2022-01-29 Thread James Cloos via Dng
Looking on my main web box (running deb sid w/ sysvinit), I see that I
have a dummy systemd-tmpfiles there:

:; apt show systemd-tmpfiles
Package: systemd-tmpfiles
Version: 1.0
Status: install ok installed
Priority: optional
Section: misc
Maintainer: Your Name 
Installed-Size: 9,216 B
Provides: systemd-tmpfiles
Homepage: httpgo.away
Download-Size: unknown
APT-Manual-Installed: yes
APT-Sources: /var/lib/dpkg/status
Description: make php work


so this is not new on deb's side; old enough in fact that i do not
recall having to do that.  (well maybe, now, but not until after i
wrote the above.)

-JimC
-- 
James Cloos  OpenPGP: 0x997A9F17ED7DAEA6
___
Dng mailing list
Dng@lists.dyne.org
https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng


Re: [DNG] Devuan with usr merge?

2021-11-13 Thread James Cloos via Dng
>>>>> John Morris via Dng  writes:

> So yes, it is time to eliminate /bin, /sbin and /lib.

the real result shod be eliminate /usr.

system packages should all use --prefix=/, local ones should default
to --prefix=/local, and closed src crap^Wstuff should use things like
/opt/FOO for a prefix.

linux, like att, calls bsd's /usr /home.  the ample disk sizes mean /usr
has lost its value and should be gone.

> Wish I could say the same thing about the X11 vs Wayland divide.  See
> the cold logic and theory in the Wayland argument but keep looking at
> the current reality and Wayland comes up short.

wayland has too many design flaws ever to be reasonsable.  weston might
be salvagable, at least as a basis for a proper av compositor.  but what
we need for that space is a daemon whose sole purpose is providing
interconnection between clients.  like a sip proxy with a better protocol.
unix, ip/udp and ip/tcp sockets at least, perhaps sctp and dccp, too.
both proxy-style and enabling the two to negotiate direct sockets.
but all type sockets by default and from the start.  when they use unix
sockets the can exchange a FD and there by arrange shared memory should
that be userful.  but ip sockets must remain fully usable for everything.

input, for example should be clients.  one for archaic stuff, by way of
the kernel, like the pre-usb stuff.  plus one daemon for each usb device.
and eventually for ethernet-connected devices.  (802.2cg w/ lp-wan style
ipv6 would work very well for input devices.  and ch for things like camaras.)

wayland's anti-network design, and not having things like input also be
clients make it too broken for a useful future.

(of course cg's 10 Mbit bandwidth mean normal ip/ethernet also would work,
but lp-wan's 127-octet mtu and therefor smaller v6 packets may be better
for simple input devices and input's real-time nature.)

-JimC
-- 
James Cloos  OpenPGP: 0x997A9F17ED7DAEA6
___
Dng mailing list
Dng@lists.dyne.org
https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng


[DNG] odd dist-upgrade difference

2021-05-27 Thread James Cloos via Dng
given the addition of breaks:mlocate to plocate, a sid box will delete
mlocate when upgrading plocate, but a ceres box instead will refuse to
update plocate when both are installed.

re-running /apt install plocate/ did the right thing.

not sure which behaviour is generally preferred,
but i prefer what sid did.

(same behaviour on multiple arch; only deb v. dev was different.)

-JimC
-- 
James Cloos  OpenPGP: 0x997A9F17ED7DAEA6
___
Dng mailing list
Dng@lists.dyne.org
https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng


Re: [DNG] iptables at risk when uprgading?

2019-11-16 Thread James Cloos via Dng
>>>>> "HB" == Hendrik Boom  writes:

HB> Does this mean that the upgrade from ascii to beowulf is not transparent
HB> and that I risk losing the iptables on my front-end machine when I do it?

Save the tables with iptables-save and ip6tables-save before upgrading
and try restoring them with iptables-restore and ip6tables-restore after
upgrading.

Or, save before and then, after the upgrade, run:


#!/bin/sh
update-alternatives --set iptables /usr/sbin/iptables-legacy
update-alternatives --set ip6tables /usr/sbin/ip6tables-legacy
update-alternatives --set arptables /usr/sbin/arptables-legacy
update-alternatives --set ebtables /usr/sbin/ebtables-legacy


and then restore them with iptables-restore and ip6tables-restore.

If you do not have arptables and ebtables installed, those two
lines will report errors, but it won't hurt anything.

After the upgrade, /usr/sbin/iptables will be a symlink to
/etc/alternatives/iptables.  After the update-alternatives calls,
/etc/alternatives/iptables will be a symlink to
/usr/sbin/iptables-legacy, so you won't then have to edit any callers.

Symlinks exist like that also for ip6tables, each's -save and -restore
as well as {arp,eb}tables.

Displaimer: not my work; someone posted the u-a calls on debian-devel
a few weeks ago.

I use it on some kvm and openvz which do not work well with nftables or
the nftable-using iptables. They got filtering working again for me.

-JimC
-- 
James Cloos  OpenPGP: 0x997A9F17ED7DAEA6
___
Dng mailing list
Dng@lists.dyne.org
https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng


Re: [DNG] /usr to merge or not to merge... that is the question??

2018-11-20 Thread James Cloos
For the question of whether merging / and usr is good:

It would be good except that rh pushes doing it backwards and debian
seems to be falling for that insanity, too.

/usr was intended for home directories.  Hense usEr.

/usr/bin and /usr/lib were created only because disks were small.

Disks, not partitions.

Since they no long are, and since linux went with /home for home dirs,
/usr should be eliminated.

*Everything* currently in /usr should instead be in /.

That is the only reasonable, rational and ethical way to merge them.

The default prefix should be /local and the default prefix used by
distributions should be /.  /opt/package should remain for non-src
stuff.

But the string '/usr' should never appear in any pathname.

-JimC
-- 
James Cloos  OpenPGP: 0x997A9F17ED7DAEA6
___
Dng mailing list
Dng@lists.dyne.org
https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng


Re: [DNG] what happened to usbmount?

2018-06-19 Thread James Cloos
>>>>> "HB" == Haines Brown  writes:

HB> Has an alternative been developed without the problems associated
HB> with usbmount?

pmount may be what you want.

I've not used it on devuan (or debian), but it works well on my gentoo
workstation.

-JimC
-- 
James Cloos  OpenPGP: 0x997A9F17ED7DAEA6
___
Dng mailing list
Dng@lists.dyne.org
https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng