Re: using intel i5 freqency governors

2021-12-01 Thread Stefan Monnier
 How do I get the intel cpu "turbo boost" fully engaged when I'm
 running my script and go back into power save mode when the machine is
 idle?
>>
>> That should be the default behavior (i.e. if you don't touch any cpu
>> power configuration).
>
> Unfortunately, it clearly is _not_ the behavior I get.  .. and I
> hadn't touched any power config until I was trying things to get my
> 900 second execution time down around the 600 seconds it takes on the
> windows/cygwin machine.

You might want to try and figure out why you get `powersave` as default
governor, but until you've figured it out, you might like to force the
use of the `schedutil` governor instead which should give you the same
kind of speed as the `performance` governor under load while reverting
to low-power state afterwards.


Stefan



Re: [OT, deeply] Guix

2021-11-11 Thread Stefan Monnier
> So why not use it as an install tool? Then your entire configuration
> is recorded in and driven by a pretty simple text file. That's all you
> need, so we can pre-seed that config file for automatic installs. And
> we can customise that SAME file for installs configured in real-time,
> because salt has excellent templating capability in the config files.

I'm not interested in *installing*: I'm interested in evolving an
already installed system.  AFAIU Puppet can do that, don't know about
the other tools you mentioned.


Stefan



Re: [OT, deeply] Guix

2021-10-26 Thread Stefan Monnier
> I stumble upon this article about (supposedly) Guix's
> characteristics/advantages:
[...]
> , and was curious about the opinions of the educated Debian people on the
> matter.

I haven't read that article, but here's my opinion:
I would love to see Debian move towards a model like that of NixOS or
Guix.  One of the main benefits I see of those systems is that it has
a declarative description of what the system should contain.

Think of it this way: currently, you can more or less figure out which
packages you decided to install on your machine by going through the list
of installed packages and filtering out all those that are marked as
"automatically" installed.

But you can only manipulate this list indirectly, via `apt install` and
`apt remove`.

In contrast, with NixOS/Guix that list is available in a plain text
editable file.  And in order to add a package or remove a package you
can edit that list and then say something akin to "make" which will
add/remove the needed packages to bring the system to the state
described in the file.

Another, slightly more subtle, benefit is that you'll always get the
same system state from a given description.  In contrast, with Debian,
I have several Debian `testing` machines on which I have installed and
removed over the years the same set of packages, but not exactly in the
same order, and with a different interleaving of `apt upgrade`, and the
result is that they don't all have exactly the same packages installed,
and some suffer from problems that I don't see in the others.

With the NixOS/Guix approach, it's much easier to make sure the systems
are truly identical.

It's not been enough to convince me to switch, but I do wish Debian
would try and take some of those ideas.  I think part of it could be
done all within a new APT-like tool without any change to Debian itself
(the part that decides which packages should be installed and removed
based on some text file describing the desired configuration of the
system).


Stefan "happy Debian user since 2003"



Re: Sata Hard drive testing

2021-10-18 Thread Stefan Monnier
> Essentially, I have been experienced data loss, where nodes become
> unreadable, when I try to "fix it" with fdisk, it says it moves unreadable
> to the trash, basically deleting data.

"it says"?  Can you clarify what is this "it"?

Drive-level errors of "unreadable data" normally lead to errors that
appear in the kernel log and say scary things and "read sense" and
include hex numbers showing the actual bytes of the SATA command sent.

I don't expect such error messages to say anything about a "trash".
So maybe your errors occur elsewhere.


Stefan



Re: A .profile puzzle

2021-10-17 Thread Stefan Monnier
> normally when a storm comes through i turn off the PC anyways because
> I really don't want to have things fried (even if i do have the UPS
> and surge protection).

Hmmm does turning them off make any difference w.r.t a surge large
enough to pass through the surge suppression?

I thought the only effective way to make a difference is to
*unplug* them.


Stefan



Re: why pae kernel has only 3G memory

2021-10-16 Thread Stefan Monnier
> Any particular reason why you must use pae rather than 
> the standard 32 bit kernel?

Hmm... isn't "the standard 32 bit kernel" `linux-image-686-pae` ?

At least the `linux-image-686` package seems to describe itself as for use
only on those machines where `linux-image-686-pae` doesn't work.


Stefan



[OFFTOPIC] Re: mail service

2021-10-10 Thread Stefan Monnier
> In general, the circumstances which would require one to use a tool like
> protonmail are not commonly observed in connection with a list like
> debian-user.

While protonmail might be used for such situations, in my experience
most protonmail users I've seen are just people that are sufficiently
technically aware to know that they should stay away from gmail and
friends and look for a quality email provider (posteo being another
popular provider in that space).

So, my own bias would rather tend to expect good behavior (good
questions and good answers) from participants posting from
protonmail ;-)


Stefan



Re: slower Debian 11 after upgrade from 10

2021-10-05 Thread Stefan Monnier
> Sometimes an older kernel does better than a newer kernel, in the absence
> of proper firmware.  Sometimes it doesn't.

Sometimes a blob is moved from `main` to `contrib` or `non-free` because
it was in `main` by mistake (tho I can't remember that happening
between Debian 10 and Debian 11).


Stefan



Re: Slow graphics on Intel Xtreme?

2021-10-01 Thread Stefan Monnier
> This would seem to be running on the very edge of hardware: I'm surprised
> that anything graphical will run at all in 1GB of total memory and 64M
> of video RAM.

My old Thinkpad X30 is still quite functional (that's the machine I use
to plug into LCD projectors to display slides during talks and classes).

So, I can confirm that 1GB of RAM is still enough to run XFCE in Debian
testing (the PDFs are displayed via doc-view-mode in Emacs).  Firefox is
too sluggish on that machine, OTOH.


Stefan



Re: write only storage.

2021-10-01 Thread Stefan Monnier
>> Write only storage - DVD-R or equivalent Blu-Ray - but make sure to end the
>> session. Deletion - feed through a paper shredder.
> I already do that but currently that means I have roughly one month of
> backups on network accessible storage before I write to disc.

Rather than WORM you can just take normal disks and once you don't want
to write to them any more, you unplug them ;-)

If you still want to have read access to the data, then you make it
accessible via another server, ideally in another administrative domain
(and another physical location, since fires and other events can be just
as likely as ransomware).


Stefan




Re: (unable to start a new discussion) Re: Can surf the internet, but not my home network...

2021-09-17 Thread Stefan Monnier
> edges of text characters. So if showing a screen full of text to show 
> the error, smunch the daylights out of it, it will still be readable. 

Whatever happened to the idea of citing the actual text rather than
using an (unreadable) image?


Stefan



Re: /usr/sbin/reboot: disabled in systemd-nspawn container

2021-09-04 Thread Stefan Monnier
Greg Wooledge [2021-09-04 11:35:25] wrote:
> On Sat, Sep 04, 2021 at 04:49:24PM +0200, sp...@caiway.net wrote:
>> # file /sbin/reboot
>> /sbin/reboot: POSIX shell script, ASCII text executable
>
> That's not normal for a bullseye system using systemd for init.

Indeed.

> I'm not sure what you did or how your system got into this state.

Maybe

dpkg -S /sbin/reboot /usr/sbin/reboot

could give us a clue?


Stefan




Re: Debian 11 Live USB with persistence?

2021-08-19 Thread Stefan Monnier
to...@tuxteam.de [2021-08-19 09:11:00] wrote:
> On Wed, Aug 18, 2021 at 10:24:27PM -0400, Stefan Monnier wrote:
>> > I wanted to do something exactly like that some months ago.
>> > What I ended up doing is using a normal Debian installation
>> > with an overlay file system mounted over the root.
>> FWIW, you can do simpler and just use a normal Debian install on
>> a USB key.  That saves the trouble of the overlay filesystem.
>> And if you want to be able to revert easily to the initial state, then
>> you can probably get a similar result using an LVM snapshot.
> There's still some charm to Santiago's approach: the device is (nearly?)
> read-only after install.

I'm not sure I see the difference: the LVM snapshot can also be tagged
as read-only.

In both cases (LVM snapshot or overlayfs) the overall system is not
read-only.  That's of course indispensable for the "persistence"
property, but it's also a disadvantage compared to a Debian Live system
in that the /var subtree will sometimes be modified "gratuitously".

[ I usually try to reduce this aspect by using a tmpfs for /var/cache
  and also by moving /var/lib/apt/lists to /var/cache (with a symlink),
  but there are various files that can get written depending on the
  packages you have installed.  ]


Stefan



Re: Debian 11 Live USB with persistence?

2021-08-19 Thread Stefan Monnier
> I wanted to do something exactly like that some months ago.
> What I ended up doing is using a normal Debian installation
> with an overlay file system mounted over the root.

FWIW, you can do simpler and just use a normal Debian install on
a USB key.  That saves the trouble of the overlay filesystem.

And if you want to be able to revert easily to the initial state, then
you can probably get a similar result using an LVM snapshot.


Stefan



Re: nvme SSD and poor performance

2021-08-17 Thread Stefan Monnier
> P-S: If triming it is needed for ssd, why debian do not trim by default?

AFAIK trimming is not needed.  It can be beneficial in some cases, but
as a general rule, the SSDs should be able to provide great performance
without it.


Stefan



Re: Debian 11 bullseye Gdm3 nvidia 7200go nouveau glitches and more

2021-08-17 Thread Stefan Monnier
> From https://www.cnet.com/reviews/hp-pavilion-dv6300-preview/ I gather it  
> has a rather old processor (Celeron M 440 to Core 2DuoT7200) -- which one do
> you have exactly? Also, it seems there would be at most 2 GiB of RAM.

FWIW, I'm surprised it would only allow 2GB, since the previous
generation chipsets (for Core[non-2]Duo laptops) had a limit of 3GB.
So, I suspect it can actually take 4GB (or even 8GB).


Stefan



Re: how to to start X Window in debian-live-11.0.0-i386-standard.iso

2021-08-15 Thread Stefan Monnier
Greg Wooledge [2021-08-15 10:30:27] wrote:
> On Sun, Aug 15, 2021 at 04:13:37PM +0200, sp...@caiway.net wrote:
>> lou what about
>> # apt install xorg fluxbox
>> 
>> $ startx
>
> In theory that might work (you may also need firmware), but the amount
> of bandwidth you'd burn through downloading *all* those packages each
> and every time you boot the live image would be an appalling waste.

I expect the vast majority of Debian Live images nowadays are written to
USB flash storage rather than to read-only media, so I think it would
make a lot of sense for those live images to offer the ability to change
the image with things like `apt install`.

Personally, I've used (minimal) "normal Debian installs" instead of
Debian Live for that reason.


Stefan



Re: password set at installation of debian-10.10.0-amd64 not recognized

2021-08-15 Thread Stefan Monnier
> You don't really get that far if your root FS is unmountable.

Hmm... if it truly can't be mounted, then the system can't read
`/etc/passwd` and then whether there is a root account or not makes
no difference.

In the "usual" case where the root FS is readable but fsck found errors,
then indeed the normal boot steps will require a root password, but you
can normally circumvent this with `init=/bin/sh`.

If this fails because `/bin/sh` is affected by the filesystem's
inconsistency, then you need to resort to the initrd's debug shell,
e.g. with something `break=local` IIRC.

Of course, the existence of a root password can occasionally be handy
for things like `rsync` (strictly speaking, you can probably arrange
for rsync to first log in as a normal user and then use `sudo`, but
it requires a fair bit of fiddling, IIRC).


Stefan



Re: password set at installation of debian-10.10.0-amd64 not recognized

2021-08-14 Thread Stefan Monnier
> Nevertheless there are rare cases only root can make changes.

You mean cases where `sudo zsh -l` is not an option?


Stefan



[OFFTOPIC] Plonk (wss: Meta: behavior on list)

2021-08-13 Thread Stefan Monnier
to...@tuxteam.de [2021-08-13 19:11:43] wrote:
> On Fri, Aug 13, 2021 at 12:49:34PM -0400, Polyna-Maude Racicot-Summerside 
> wrote:
> [...]
>> Plonk ?
> Greg and The Wanderer already provided a definition. Just adding
> one standard reference, the Jargon File [1] in such things. Old
> Usenet lore.
> Cheers
>
> [1] http://www.catb.org/jargon/html/P/plonk.html 

How odd.  I always assumed that it was the comic-strip style
representation of the sound of hanging up the phone abruptly.


Stefan



Re: Disk for a small server

2021-08-11 Thread Stefan Monnier
David Christensen [2021-08-11 14:48:05] wrote:
> That is why there is the scientific method.  Please cite relevant article(s)
> with reproducible laboratory results and/or analysis of long-term real-world
> data regarding failure modes, effects, and hazards of non-ECC memory vs. ECC
> memory when paired with operating systems with vs. without storage stack
> integrity checking and correction.

And until such empirical data shows up, I'll give the benefit of the
doubt to the scientists/academics.


Stefan



Re: can't login via gdm

2021-08-11 Thread Stefan Monnier
>> I was attempting to modify
>> # for i in `find / -user 1001`; do chown 3001 $i; done
> Dear gods.  That violates at least two *major* rules of the shell.

It looks exactly like the kind of code I would write if I were trying to
"innocently" introduce a backdoor.


Stefan



[OFFTOPIC] Editing a file (was: percent char '%' in sudoers file)

2021-08-10 Thread Stefan Monnier
Roger Price [2021-08-10 11:11:24] wrote:
> On Tue, 10 Aug 2021, Bob Bernstein wrote:
>> Full disclosure: In a typical Bob fit of impulsivity I, yes, edited this
>> file using 'sudo nsno /etc/sudoers'.
> My impulse would be to use  VISUAL=/usr/bin/emacs visudo -f /etc/sudoers

You guys have amazing impulse control.
My fingers would have typed `e /etc/suders` before I would have had
a chance to do anything about it.


Stefan


PS: Of course the `e` above stands for Emacs, tho my root user has it
redirected to Zile.



[OFFTOPIC] Fun with pv and broken drive (was: `pv` rates)

2021-08-04 Thread Stefan Monnier
Stefan Monnier [2021-08-02 15:09:38] wrote:
> E.g. the average has moved up to ~80kB/s since my last message).

For your entertainment: I managed to bring the average rate up to about
1MB/s by running `lvs` in a loop at the same time (and that let me see
occasionally a transfer rate around 10MB/s).

I have no idea what it is that makes `lvs`s interference
so beneficial.  I tried a few other operations (like `du`, and `dd`) but
to no avail.

Without any interference from `lvs` I get a long term average slightly
below 100kB/s.


Stefan



Re: [OFFTOPIC] Mass storage prices and form factor

2021-08-03 Thread Stefan Monnier
> One cloud storage provider, Backblaze, regularly publishes reports on
> harddisk reliability (they obviously have a lot of data on that :-)

All of them seem to be 3½" as well.
Interesting.

It actually looks like the 2½" HDD market has been abandoned: 5 years
ago, the largest HDD were 5TB for 2½" and 10TB for 3½".

Since then, it seems like nothing new happened in the 2½" space, whereas
the 3½" drives kept growing (tho only about 15% per year).

Based on what I've heard, I suspect that to be competitive in the
datacenter, 2½" drives would need to offer about half the capacity of
3½" drives (of course, that also presumes manufacturers going through
the trouble of producing CMR drives in 2½" form factor).

OTOH for an end user who doesn't need more than 8TB of disk space, 2½"
drives can often be a better option (assuming you have enough SATA
ports).


Stefan



Re: ISO to external SSD

2021-08-03 Thread Stefan Monnier
> Second, the price of spinning disks is such that it makes no
> sense to buy anything smaller than 4TB, which will fit all this,
> and 6-8 TB are often a reasonable idea even for single users.

You seem to assume a 3½" form factor which either requires a "large"
desktop or an external enclosure.
Personally I consider this form factor dead every since I bought my
first 2TB 2½" disk.
[ For reference, the last 3½" disk I bought was a 1TB WD Green, which
  I fitted into my ASUS WL-700gE.  ]


Stefan



[OFFTOPIC] Mass storage prices and form factor (was: ISO to external SSD)

2021-08-03 Thread Stefan Monnier
Peter Ehlert [2021-08-03 08:27:26] wrote:
> On August 3, 2021 8:17:58 AM Stefan Monnier  wrote:
>>> Second, the price of spinning disks is such that it makes no
>>> sense to buy anything smaller than 4TB, which will fit all this,
>>> and 6-8 TB are often a reasonable idea even for single users.
>> You seem to assume a 3½" form factor which either requires a "large"
>> desktop or an external enclosure.
> Not really. My HP z620 work station ain't huge.

Clearly "large" is smaller than "huge", so we don't disagree.  FWIW of
my three desktops only one can accommodate 3½" disks (a Streacom F8
case) [ the other two are a 2006 Mac mini and a Librem mini.  ]

>> Personally I consider this form factor dead every since I bought my
>> first 2TB 2½" disk.
> If you use Lots of drives (4tb), and are on a limited budget, like me... The
> 3.5 form factor is more cost effective.

That's why I wrote "personally".  AFAIK the proportion of computer users
which "use lots of drives" is quite small.  Maybe it's higher among
Debian users, admittedly, but still your original statement above lacks
a qualification like "AFAIC" or "IMO" ;-)

OK, to add actual info to this message, here's a quick look at today's
lowest prices in my "usual" store:

   2TB $ 64 3½  HDD 32 $/TB
   3TB $ 65 3½" HDD 22 $/TB
   4TB $104 3½" HDD 26 $/TB
   6TB $139 3½" HDD (external)  23 $/TB
   8TB $199 3½" HDD 25 $/TB
  10TB $323 3½" HDD (external)  32 $/TB
  12TB $339 3½" HDD (external)  28 $/TB

I'm surprised at how stable the price per TB is over the 2-12 range.
It used to be the case that HDD price's curve was not nearly as linear
(which reflected the fact that production costs aren't (weren't?)
proportional to the drive's capacity).
I suspect that the profit margin varies widely over this spectrum.

   2TB $ 60 2½" HDD (external)  30 $/TB
   3TB $125 2½" HDD (external)  42 $/TB  (only "recertified" available)
   4TB $114 2½" HDD (external)  29 $/TB
   5TB $139 2½" HDD (external)  28 $/TB

It's also interesting to see that the price per TB is about the same for
2½" HDD as for 3½" HDD, whereas it used to be significantly higher.
[ And also that in the 2½" space, your best bet in $/GB is to buy
  a drive+enclosure, which implies you don't really know what you're
  getting other than the capacity of the drive.  :-(  ]

That makes 3½" form factor even more dead than I thought (two 4TB 2½"
drives should offer better performance than one 8GB 3½" drive and use
less space, not sure about power consumption).

I wonder what kind of drives are used nowadays in big datacenters (and
whether the prices they pay for them looks anything like the ones
above).

For completeness, here's the list for SSDs:

   ½TB $ 60 2½" SSD120 $/TB
   1TB $100 2½" SSD100 $/TB
   2TB $244 2½" SSD122 $/TB
   4TB $540 2½" SSD135 $/TB

so about 4x the price per TB (and here the linear price curve makes
a lot more sense).


Stefan



Re: `pv` rates

2021-08-02 Thread Stefan Monnier
>> If it weren't for the first sometimes changing to
>> 44.xKiB/s it'd be hard to know which is which (IIUC the average is
>> higher because occasionally the drive gives a more reasonable transfer
>> rate than that measly 45kB/s).
>
> So now we're left wondering how you came by this situation. Perhaps
> you slumped onto the Return key, then woke up after a few minutes,
> having missed the initial burst that gave rise to the average being
> more than 50% faster than the current rate.

I don't know either.  Reading the whole drive would take a few years, so
I'm only fetching the few parts which have changed since the last
backup, and when I look the rate seems to be always ~45kB/s, but
obviously there have to be bursts at higher speeds (presumably because
the drive's defect doesn't affect every cylinder in the same way or
something like that.  E.g. the average has moved up to ~80kB/s since my
last message).


Stefan



`pv` rates (was: burn iso to usb)

2021-08-02 Thread Stefan Monnier
Anssi Saari [2021-08-02 19:04:59] wrote:
> David Wright  writes:
>> On Mon 02 Aug 2021 at 16:14:15 (+0300), Anssi Saari wrote:
>>> Stefan Monnier  writes:
>>> >>> > cp /path/to/file.iso /dev/sdX
>>> >>> dd if=whatever.iso of=/dev/sdX  
>>> >> It's up to taste.
>>> >
>>> > Not at all.  The only right answer is:
>>> >
>>> > pv -parIe /dev/sdX
>>> 
>>> Actually I'm not sure how good it is to have both -a and -r, pv doesn't
>>> really show which rate counter is which...
>>
>> No need: the rate is far more variable than its average, as time passes.
>
> I guess it depends. Before that comment I tried it on an old USB
> stick. Read speed was pretty much constant if low. So I think it was the
> left rate counter that showed current rate but wouldn't bet on it.

[ I'm glad my silly intervention brings up a more constructive
  discussion ;-)  ]

If they're both pretty much constant, they're presumably both pretty
much equal, so it doesn't matter which is which ;-)

But indeed, it's not always the case.  I'm right now using `pv` to read
data off of a broken drive (a 2TB 2½" drive which apparently has
problems seeking, resulting in a transfer rate of about 45kB/s), and
it's currently showing me 45.0KiB/s and 70.6KiB/s both of which are
quite stable.  If it weren't for the first sometimes changing to
44.xKiB/s it'd be hard to know which is which (IIUC the average is
higher because occasionally the drive gives a more reasonable transfer
rate than that measly 45kB/s).


Stefan



Re: Updating kernels impossible when /boot is getting full

2021-08-02 Thread Stefan Monnier
> So really think hard before splitting off a filesystem outside of
> volume management. I believe it is more likely to cause problems
> than it is to avoid problems.

All my machines have a separate /boot partition (and everything
else in LVM).  These are all "historical accidents", because at the time
I set them up, the respective boot loader (LILO, Grub, U-Boot) didn't
know how to read LVM volumes, and I just never bothered to change.

But I fully agree with you: if your bootloader can read from LVM
(as is the case with Grub2), then you're better off without a separate
/boot partition.


Stefan "not sure if U-Boot can read from LVM yet"



Re: Updating kernels impossible when /boot is getting full

2021-08-01 Thread Stefan Monnier
Ilkka Huotari [2021-08-01 07:20:20] wrote:
> I'm using Ubuntu 21. My /boot partition size is 500M and it's getting full:

Notice that this is a Debian mailing-list, so questions about Ubuntu are
not really "on topic".

AFAIK Ubuntu handles the initrd files and kernels slightly differently
from what Debian does, so you'll probably be better served asking this
on an Ubuntu mailing-list or forum, otherwise you risk getting
confusing advice.


Stefan



Re: burn iso to usb

2021-07-31 Thread Stefan Monnier
>> > cp /path/to/file.iso /dev/sdX
>> dd if=whatever.iso of=/dev/sdX  
> It's up to taste.

Not at all.  The only right answer is:

pv -parIe /dev/sdX

Anyone who says otherwise is trying to trick you,


Stefan



Hardware life expectancy (was: Please help to test latest Debian 11 release candidate on real hardware)

2021-07-25 Thread Stefan Monnier
> Should I participate if my laptop is 10 years old?

I fondly remember showing my 2003 Thinkpad X30 to my students when it
turned 10 years old.  Given that I grew up in the glory days of Dennard
scaling, the standard rule of thumb was that you wanted a new machine
every 3 years or so and a machine older than 5 years was just
utter crap.

So being able to still make good use of a 10 year old machine
seemed surreal.

Nowadays, I'm still planning to use that same Thinkpad X30 to display
PDFs in the classroom (when I get to meet students physically again),
and more than half of my machine are older than 10 years old.
Better yet, they don't seem significantly slower than my newer machines.

So, yes, 10 year old machines and still very much relevant.


Stefan



Re: where can i find a list of wireless adapter that debian support

2021-07-23 Thread Stefan Monnier
> the tricky part of my search for ideal adapter(needn't non-free firmware)
> isthat many vendors claim they support linux, but i'm afraid they require
> non-free firmware
> is there some easy way to find out if it requires non-free firmware?

Buy it from a place that did this job for you.

E.g. my Librem mini came with a wifi adapter that doesn't require
non-free firmware, because the vendor went through the trouble (it
includes a bluetooth feature as well but that feature does require
a non-free blob, sadly).

I think you want to look for wifi adapters that work with the ath9k
driver (I don't know if that is sufficient to guarantee that they work
with a free firmware, tho).


Stefan



Re: MDs & Dentists

2021-07-21 Thread Stefan Monnier
Thomas Schmitt [2021-07-21 20:11:15] wrote:
> Stefan Monnier wrote:
>> Plain old HDDs and SSDs also work for "immutable" backups: just don't
>> keep them connected to the host after you've done the backup.
> But how do you keep the system from messing them up during the first
> backup after the malware took over ?

That's what

keep a handful of external SSDs to which you backup "in rotation"

is about.  Of course, if the time between the moment the malware takes
control and the moment when you notice it is longer than your rotation
you might be in trouble, indeed (tho that depends on the details of how
the malware operates w.r.t removable media).

Then again, this implies that all the work you've done between those two
moments would likely have been lost just as well if you backed up to
a BD-R.


Stefan



Re: MDs & Dentists

2021-07-21 Thread Stefan Monnier
Thomas Schmitt [2021-07-21 19:00:48] wrote:
> James H. H. Lampert wrote:
>> "Immutable backups." Interesting concept. But how? Optical media?
> Yes. BD-R single layer are affordable and can take 25 GB each. With some
> compression you can put the whole operating system and the most important
> user data onto a single medium.

Plain old HDDs and SSDs also work for "immutable" backups: just don't
keep them connected to the host after you've done the backup.
E.g. keep a handful of external SSDs to which you backup "in rotation".


Stefan



Re: where can i find a list of wireless adapter that debian support

2021-07-19 Thread Stefan Monnier
loushanguan2...@sina.com [2021-07-19 20:32:39] wrote:
> i've found some adapter with search enginehopefully it can be installed in
> debian without firmware from non-free

AFAIK there is currently no wifi card that supports 11ac and for which
there exists a non-proprietary firmware.


Stefan



Re: Buster+Gnome on Pi4

2021-07-16 Thread Stefan Monnier
pcr [2021-07-16 20:28:01] wrote:
> Yesterday I had good luck with LibreCAD, did a nice drawing without any
> trouble; today, however, I failed to install REDUCE-algebra, because there
> is no version for ARM.  I tried to get the source and compile it myself, but
> that meant installing Subversion, and they don't have an ARM version
> either.

That's weird.  I'd expect `subversion` to be in Debian and to compile
for ARM without any particular problem.

> Today I was unable to get Zoom for ARM;
> I'm beginning to see a pattern :-)

As a general rule, Free Software should work just fine, but proprietary
software will be problematic because they only provide binaries for the
most common platforms.


Stefan



Re: How do I mount the USB stick containing the installer in Rescue Mode?

2021-07-15 Thread Stefan Monnier
Greg Wooledge [2021-07-15 07:00:40] wrote:
> On Thu, Jul 15, 2021 at 12:55:11PM +0300, Reco wrote:
>> "nofail" is really needed for removable devices, because whoever
>> designed systemd made an "interesting" decision to halt the boot process
>> (i.e. host is inaccessible by network, console access only) even if a
>> single filesystem mentioned in fstab fails to mount.
> This was the traditional behavior before systemd, so one can't really
> fault systemd for continuing the practice.

That's not my recollection.  AFAIK before systemd, you'd just get an
error message and the boot would just (try to) continue.

I don't think systemd's decision is bad.  But I think it's implementation is
not good enough: it should offer some kind of simple "continue
y/n?" prompt.
[ "Simple" for the user: the implementation might be not so simple.  ]

To be honest, I've added the `nofail` pretty much everywhere and hence
haven't faced this problem recently, so for all I know, the
implementation has already been improved.  But the behavior I saw back
when moving to systemd was definitely not pleasant.


Stefan



Re: MDs & Dentists

2021-07-14 Thread Stefan Monnier
jeremy ardley [2021-07-15 07:03:18] wrote:
> On 15/07/2021 6:26 am, Polyna-Maude Racicot-Summerside wrote:
>> Also can you find me one Linux distribution that is certified as medical
>> equipment for reliability ?
> You can't be seriously suggesting Windows is certified for any risk of life
> application!?

He's talking about official certification by some government entity or
some professional association or something like that.
In many situation, such a certification is required by law or by the
insurance companies.

It doesn't really matter what that certification means technically.


Stefan



Re: Bref rapport d'investissement pour mai 2021

2021-07-14 Thread Stefan Monnier
I can't reproduce this on Debian testing.
A problem in your config, maybe?


Stefan


steve [2021-07-14 09:14:48] wrote:

> Bonjour Xavier,
>
> Oui bien reçu. Mais c'est celui de mai, je pensais qu'on parlait de
> celui de juin (mais c'est probablement trop tôt).
>
> Par ailleurs, j'ai posté ce matin les documents pour le mandat de
> gestion.
>
> Excellente journée
>
> Meilleures salutations
> Steve
>
>
> Le 14-07-2021, à 07:03:31 +, xavier.crit...@ubs.com a écrit :
>
>>Bonjour Steve,
>>
>>Encore merci à vous pour la visite et votre confiance!
>>
>>Je viens de renvoyer le mail. Merci de me confirmer que vous l'avez bien reçu.
>>
>>Nous restons à votre disposition.
>>
>>Très bonne journée et courage pour cette dernière ligne droite avant les 
>>vacances!
>>
>>Xavier



Re: Working for free [was: Offensive variable names]

2021-07-14 Thread Stefan Monnier
Jeremy Ardley [2021-07-14 12:52:10] wrote:
> On 14/7/21 12:09 pm, Andrei POPESCU wrote:
>> On Ma, 13 iul 21, 20:54:22, Brian wrote:
>>> On Tue 13 Jul 2021 at 15:38:26 -0400, rhkra...@gmail.com wrote:
>>>> On Tuesday, July 13, 2021 11:27:03 AM Stefan Monnier wrote:
>>>>> BTW, for those who don't want to run their own server, there are still
>>>>> other reasons to avoid Github: my rule is to try and opt for the
>>>>> underdog so as to foster competition.
>>>> +1
>>> You are apparently well up in this field. Is there a particular
>>> udderdog you would recommend?
>> As far as I understand Gitlab is the only comparable alternative,
>> preferably self-hosted (like Debian's Salsa).
> AWS has a permanently free GIT repository for small projects (under
>  5 developers) and limited storage requirements
> https://aws.amazon.com/codecommit/

Then again, I'm not sure that Amazon qualifies as an "underdog".


Stefan


PS: Your reply's text was oddly "hidden" inside your signature (as
defined by the old convention that "\n-- \n" separates the main body
from the signature).



Re: Working for free [was: Offensive variable names]

2021-07-13 Thread Stefan Monnier
> We are in basic agreement. I'm not really a "developer" - I just host
> some relatively simple projects on Github. I agree that a deeper use of
> something like Github is something I'd have to carefully consider.

BTW, for those who don't want to run their own server, there are still
other reasons to avoid Github: my rule is to try and opt for the
underdog so as to foster competition.


Stefan



Re: Offensive variable names [was: Cool down ...]

2021-07-12 Thread Stefan Monnier
> It's only offensive to the people who are offended.  Theoretically all
> words are offensive since any word can be offensive to anyone just because 
> they deem it so.  Censoring (i.e. changing the language) of everything to
> appease everyone 1) isn't possible, 2) is foolish at best, 3) is a 
> waste of everyone's time, and 4) creates a power hungry mob of zealots
> looking to dismantle any word they deem offensive (e.g. paper machete).

It's neither black nor white.  Reduction to the absurd doesn't help make
good decisions.  After all, a breath of fresh air is just a bunch of
protons/neutrons/electrons moving, so it's no different from a bullet.
Or is it?


Stefan



Re: Buster no release file

2021-07-11 Thread Stefan Monnier
> Revert the change or communicate with the edior. Maybe he has a
> persuasive argument?

In my experience, "communicate with the editor" is the second step, the
first step being "try to figure out how to communicate with the editor".


Stefan



Re: ASTM Lab equipment protocol

2021-07-09 Thread Stefan Monnier
> Now that is easy.  I did not know about the archives.  Or the search.
> No results though.  So this might be a bit of a dead end.
> It doesn't mean that software doesn't exist.  Just that ASTM was never
> discussed on a Debian list.

So you get to be the first!  Still it seems that debian-science or
debian-med is more likely to include people who know about ASTM than
debian-user ;-)


Stefan



Re: Need help with ffmpeg installation - strange behaviour of my system - am I correct here?

2021-07-09 Thread Stefan Monnier
> 2. repair my situation - I cannot go to Debian 10 right now (I posted the
>   reason somewhere here - mouse ist awfully slow and jumpy))

You might want to try `aptitude` instead of `apt`: it will try to offer
ways to fix the problem (by removing package).
The solutions it offers can sometimes take a long time to come up
(because it performs a search through a large search tree), and it may
have effects you won't like, so carefully review the solutions proposed.


Stefan



Re: badblocks

2021-07-07 Thread Stefan Monnier
> I got a cheap SATA to USB external adaptor and used it to look at a 500Gb
[...]
> Might I think that there is something amiss with the USB/SATA adapter
> thing ?

In my experience, USB<->SATA adapters are not super-reliable (cheap or
not), the main problem stemming from power delivery, so you might like
to retry it on other USB ports, the more power it can deliver
the better.

[ Obviously, I presume that your adapter does not have its own power
  source.  ]


Stefan



Re: Bullseye installation problem (with Matrox GPU)

2021-07-07 Thread Stefan Monnier
> I did a little more experimenting and learned my G550 doesn't need
> nomodeset for the MGA X driver to work, but it only produces 1920x1080
> on my 2560x1440 screen, and xrandr still can't identify output names.

You can trying playing with "modelines" where you reduce the screen
refresh rate to 30Hz to keep the "dot frequency" below the card's limit.

Contrary to CRT, current displays tend not to need as high a refresh
rate to avoid flicker.  Of course a lower refresh rate may encounter
other problems (e.g. your display may simply refuse to display it).


Stefan



Re: Suggested way to ssh into obsolete devices (with old ssh crypto)?

2021-07-07 Thread Stefan Monnier
> I'm aware of that. My critique was specific to the "we take it out
> because it's dangerous to the user" part.

That's often an explanation but not the main motivation.
For the `none` cipher, I think it was, tho.

IIRC the problem was that using the `none` cipher causes the
authentication to be exposed in a way that is worse than using Telnet:
with Telnet you only expose the data you send to the wire, whereas with
SSH's `none` cipher you ended up exposing the data plus your
(valued) credentials.

> I'm torn on this one... Sometimes I've the impression that this leads to
> asocial software (i.e. nobody goes to any effort to make their software
> compatible to reasonable ranges of library (and other dependencies's)
> versions).
> Akin to the Flatpaks and Snaps of this world, perhaps with a less horrible
> dependencies management story).

Indeed, it has its downsides.


Stefan



Re: Suggested way to ssh into obsolete devices (with old ssh crypto)?

2021-07-06 Thread Stefan Monnier
>> It's entirely too common for obsolete encryption options that are
>> kept for "compatibility" end up being a vector for compromise, and
>> entirely reasonable to remove such options in order to provide the
>> most secure and maintainable tool for the vast majority of users.
> That's the attitude of authoritarian software: "my software is smarter
> than you".

I think the reality is a bit more subtle ;-)

In most cases, the real driver is a desire to keep the code simple and
to ease maintenance.  Removal of old, little used, and largely untested
functionality is part of what can be done for that.

>> If you want ancient crypto options, just run an ancient binary.
>> They're very easy to find in archive.debian.org.
> They're not as easy to run as soon as they start being outrun by
> their dependencie's versions, and you perfectly know that.

My experience running old Debian packages of Emacs under Debian testing
is not that bad.  Also, I think that if it's hard to do, it can be
blamed on Debian's package management (which should move towards
something more like NixOS to solve those problems).


Stefan



Re: Suggested way to ssh into obsolete devices (with old ssh crypto)?

2021-07-06 Thread Stefan Monnier
>> If they have buffer overflow-style holes, those should be fixed.
>> Other than that I can't see how they can be less secure than the "none" 
>> cipher.
> I guess since the "none" cipher isn't supported in debian's ssh

Good point.

> you will just drop this questionable line of argument?

It just moves it to first arguing in favor of supporting `none`.


Stefan



Re: Suggested way to ssh into obsolete devices (with old ssh crypto)?

2021-07-06 Thread Stefan Monnier
>> I think the first reaction should be to report it as a bug, so that the
>> old cipher is re-added.  I think the same argument in favor of including
>> the "none" cipher should apply to including old deprecated ciphers.
> The old ciphers are generally removed for a reason: because they are hugely
> insecure.

If they have buffer overflow-style holes, those should be fixed.
Other than that I can't see how they can be less secure than the "none" cipher.

I fully agree with removing them from the list of cipher that will be
automatically chosen for you.  But keeping them available upon explicit
request for those cases where it's the only cipher that works with
a particular other device makes a lot of sense.


Stefan



Re: Suggested way to ssh into obsolete devices (with old ssh crypto)?

2021-07-06 Thread Stefan Monnier
> Like you, I have been using CLI options to the ssh command to adjust the
> necessary algorithms if I need something "insecure".

You should be able to set that option for a specific (set of) hosts in
.ssh/config so you don't have to repeat it on the CLI every time.

> My thought is that once that no longer serves the purpose, I would
> setup a VM, container,

I think the first reaction should be to report it as a bug, so that the
old cipher is re-added.  I think the same argument in favor of including
the "none" cipher should apply to including old deprecated ciphers.


Stefan



Re: Memory allocation failed during fsck of large EXT4 filesystem

2021-07-05 Thread Stefan Monnier
Reiner Buehl [2021-07-05 10:21:13] wrote:
> Hi all,
> I have a corrupt EXT4 filesystem where fsck.ext4 fails with the error
> message:
>
> Error storing directory block information (inode=366740508, block=0,
> num=406081): Memory allocation failed
[...]
> The system has 4GB of memory and a 8GB swap partition. The filesystem has
> 7TB. Is there a quick way to enlarge the swap space to help fsck.ext4 to
> finish the repair? I do not have any unused partitions but have space for
> swap on other filesystems if that is possible.

I think you should report this as a bug in e2fsck.  While 7TB is
significantly larger than the partitions I have, 8GB of swap should
still be plenty for that (my first 1TB disk was connected to a machine
with 64MB of RAM (an asus wl-700ge) and fsck was slow but it worked), so
I suspect the error is not in the lack of memory space.


Stefan



Re: removing modules

2021-07-03 Thread Stefan Monnier
> The only thing with `MODULES=dep` is that it runs a slightly higher risk
> of ending up with an unbootable system after a hardware or
> filesystem/LVM/MD/partition change.  Just make sure you have some way to
> do a "rescue boot" in those cases (typically via a USB flash key with some
> minimal GNU/Linux system on it), but this comes in handy in many other
> circumstances so I recommend you have that at hand in any case.

BTW, you can try:

lsinitramfs -l /boot/initrd.img-  | sort -n +4

to see the parts of the initrd that are likely to take most of
the space.


Stefan



Re: removing modules

2021-07-03 Thread Stefan Monnier
Greg Wooledge [2021-07-03 14:03:44] wrote:
> On Sat, Jul 03, 2021 at 06:49:35PM +0100, mick crane wrote:
>> root@pumpkin:/boot# df -h /boot
>> Filesystem  Size  Used Avail Use% Mounted on
>> /dev/sdb1   236M  159M   65M  71% /boot
> This is the real issue.  This file system is simply too small if you
> plan to keep more than 2 kernels at a time.

Not at all.  Just adjust the /etc/initramfs-tools/initramfs.conf with

MODULES=dep
COMPRESS=lzma

and you'll have room for many more kernels.

> One thing I'll note is that your initrd* files are twice as big as mine.
>
> -rw-r--r-- 1 root root 30924690 Jan 29 17:34 initrd.img-4.19.0-13-amd64
> -rw-r--r-- 1 root root 33151811 May 13 07:19 initrd.img-5.10.0-6-amd64
> -rw-r--r-- 1 root root 33165977 Jun 28 07:08 initrd.img-5.10.0-7-amd64

Indeed, his initrd.imgs are unusually large.  Not sure what's going on
with them, but the above two changes should shave off at the very least
20MB from each one of his initrd.img.

The only thing with `MODULES=dep` is that it runs a slightly higher risk
of ending up with an unbootable system after a hardware or
filesystem/LVM/MD/partition change.  Just make sure you have some way to
do a "rescue boot" in those cases (typically via a USB flash key with some
minimal GNU/Linux system on it), but this comes in handy in many other
circumstances so I recommend you have that at hand in any case.


Stefan



Re: How to verify newly burned disc

2021-07-03 Thread Stefan Monnier
> The first hit at ebay.de when searching for "m-disc dvd" shows an offer
> for 10 4.7 GB discs at 85.11 € + 7.31 € for shipping from Japan :-)
[...]
> There is also an offer from Australia, 5 discs for "~" 21.74 € + ~ 42.97

Hmm... so that's in the order of about 1 €/GB
>From a quick look at SSD and uSD prices, it seems flash memory is about
an order of magnitude cheaper (and also a lot more convenient to
write :-) ).

I wonder if "archival quality" really serves its purpose here.  I mean,
I understand that flash is not reliable in the long term, but there's
also the problem of making sure you'll still have a working DVD reader
in the future.

I thought the only reliable way to archive digital data would be to
treat preservation as a *process*, where you "refresh" your archive
every N years by copying it over to a newer media (with enough
redundancy to detect and correct errors that may have crept in during
those N years).


Stefan



Re: Whole Disk Encryption + SSD

2021-06-29 Thread Stefan Monnier
>> > Along with SED, I suggest that you also implement Secure Boot.

>> Can someone give me pointers to actually known attacks (not
>> hypothetical ones, which I can invent myself without much difficulty)
>> that would have been prevented by Secure Boot?

> [2] https://en.wikipedia.org/wiki/Evil_maid_attack

Thanks.


Stefan



Re: Whole Disk Encryption + SSD

2021-06-28 Thread Stefan Monnier
> Along with SED, I suggest that you also implement Secure Boot.

Can someone give me pointers to actually known attacks (not
hypothetical ones, which I can invent myself without much difficulty)
that would have been prevented by Secure Boot?

I can see that subverting the early boot might be a good way for
rootkits to install themselves in a way that's hard to detect and/or
remove, but it's not like there aren't plenty of other ways to get
pretty much the same result.

IOW it sounds to me a bit like putting a reinforced steel frame around
a cardboard door.


Stefan



Re: add disk to LVM

2021-06-27 Thread Stefan Monnier
[...]
> `vgchange -an` usually does the trick), but AFAIK there is standard
^
   no
Damn!

> support for doing those things for you currently (and if you forget to
[...]


Stefan



Re: add disk to LVM

2021-06-27 Thread Stefan Monnier
> What would be the way to do that using LVM ?

I think you already had the answers you need, but I'll just point out
that it is possible to use LVM on a "removable" disk (a disk which you
sometimes take out of a machine to put into another), and as a matter of
fact, it would make a fair bit of sense in several situations.

Sadly, currently the LVM tools don't support doing this very well if you
intend to do it without rebooting the machine: it's pretty easy to
connect a disk with an LVM VG on it and start using it (you just need
to do an "LVM scan" beforehand, but that's cheap and can be automated
without too much trouble) and it's not too hard to "unregister" the LVM
structures before disconnecting the disk either (basically
`vgchange -an` usually does the trick), but AFAIK there is standard
support for doing those things for you currently (and if you forget to
`vgchange -an` before you disconnect the disk then you end up with
stale LVM structures which can become quite annoying and can prevent
you from accessing that drive from that machine until you reboot it).

Maybe it's time people start reporting these hurdles as bugs and provide
patches to `pmount`, `systemd.mount`, etc...
So we can start using LVM as the "standard" partitioning mechanism for
GNU/Linux.


Stefan



Re: A feasible method to add examples to man pages?

2021-06-22 Thread Stefan Monnier
> That's what I get for blindly copying and pasting the "directions" from
> Github. I should've have just told Monnier that the whole shebang
> can be installed locally, which renders inoperative his nitpick about
> not having the Net when he needs to cheat the most (or most often).

The difference is that the manpages get installed automatically along
with the software, so they're always there, always uptodate, without me
having to think about it beforehand, without having to choose where to
install them, nor remember where I installed them.

Don't get me wrong: all it means is that I think we want to keep *some*
examples in the manpages.  IOW, I was just pointing out that the main
benefit of manpages is that they're local and work in almost all
situations.  They don't need to contain all the info you'll ever want to
know, but I usually welcome a few examples in there ;-)


Stefan



Re: A feasible method to add examples to man pages?

2021-06-21 Thread Stefan Monnier
> Many see the need for examples to be associated with man pages.
[...]
> Would it be proper/reasonable/practical/??? to add an entry of the form
> https://cht.sh/XXX ?

I think it can make sense for manpages to include a link to some further
doc in the form of a wiki or some other webpage, but I often enough find
myself using `man` while trying to fix a system's network or X11/Wayland
config, which makes it either impossible to access the Internet at that
moment, or painful to browse a webpage.


Stefan



Re: Wiping an unencrypted SSD in preparation for encryption

2021-06-15 Thread Stefan Monnier
> But in any case, I'm not sure about booting Grub on an SSD from the
> BIOS, because AIUI Grub uses sector addresses to find its core.img,
> and AIUI sectors get shuffled around by the SSD controller.

That shuffling is purely internal and hence completely invisible
(barring bugs and the need to sometimes make sure data is *really*
erased), so that's a non-issue.


Stefan



Re: thunderbird

2021-05-31 Thread Stefan Monnier
>> The most used client is probably Element, formerly called Riot
>> (https://element.io/). The link to its Linux version is not placed
>> nicely visble, it is below the Desktop Clients for Win and Mac
>> behind the comment "Also available on Linux", which points to a
>> repository which can be used for installation and updates. This
>> Element client works flawlessly.
> Source code? That page is a maze of twisty little marketing-heavy
> passages, all alike :-D

Element is not in Debian, AFAICT, but it is in F-Droid.
FWIW, I use Nheko on the desktop, which *is* in Debian.
The matrix server also is in Debian (and supported by FreedomBox).


Stefan



Re: thunderbird

2021-05-31 Thread Stefan Monnier
> Thanks, but I'll prefer a decentralised, end-to-end encryptable,
> well established messaging infrastructure with a rich choice of
> user and transfer agents. It's called e-mail.

Actually, matrix.org aims to be that as well.  It probably doesn't yet
qualify for "rich choice" (e.g. there is a matrix-client.el but it
doesn't support encryption yet), but other than that, it's pretty
much there.


Stefan "using Matrix as a replacement for WhatsApp"



Re: which command can show if usb 3.0 is used

2021-05-31 Thread Stefan Monnier
> If output shows ehci, then you are using usb 2, and if output shows xhci,
> then you can use usb 3.
   ^^^

Can, indeed, but that still depends on other things, such as the actual
setup of the physical port.  IIUC most modern Intel chipsets don't come
with EHCI any more and their xHCI handles all the ports, not all of
which are necessarily USB3.

Furthermore, "USB3" is an umbrella term which can refer to various things.


Stefan



Re: which command can show if usb 3.0 is used

2021-05-31 Thread Stefan Monnier
> hwinfo --usb-ctrl
>
> And after doing
>
> lsusb -t

Thanks.  I wasn't familiar with the `-t` arg to `lsusb`, it's
very informative (e.g. it gives you the actual speed negotiated, so
I can see that my USB C phone still uses 480Mb/s)


Stefan



Re: xv (was Re: Changing background automatically, Mint 20.)

2021-05-24 Thread Stefan Monnier
> printing images is so *20th* century. 
  ^
  2D


-- Stefan



Re: Debian-friendly laptop

2021-05-20 Thread Stefan Monnier
> While building the latest linuxcnc for my pi, I came across this mentio 
> of coreboot, so I checked their site for asus mainboard support, to find 
> asus unmentioned.
>
> So, will it work for an Asus Z370-A II mainboard?

I have no idea, but you can check the doc and or code to find out.
The statistically most likely answer is "no" because most chipsets and
motherboards are woefully underdocumented so it's very hard to port
Coreboot to it.

That's why I find Purism and System76 to be important: being
manufacturers they have a bit more leverage to get some of that
documentation (presumably under NDA) such that they can port Coreboot to
the particular chipset they're using, and that can then be helpful for
other machines using the same (or similar) chipset.


Stefan



Re: Debian-friendly laptop

2021-05-20 Thread Stefan Monnier
> Or perhaps more than a bit more ;) I see that the Librem 14 version 1
> starts at $1470, for an i7 10710U, 8GB RAM, 250GB SSD, 14" 1920x1080
> screen, no WLAN. I'm pretty sure you can do a whole lot better than
> that from the standard brands.

Indeed, you pay more, but it funds development of coreboot, saves you
from having to wonder if the wifi's chipset will work out of the box,
as well as brings to Intel's attention the existence of real customers
willing to pay to have *less* code running in the ME.


Stefan "typing this on his Purism mini"



Re: OT: minimum bs for dd?

2021-05-16 Thread Stefan Monnier
>> On Sun, May 16, 2021 at 01:31:49PM -0500, Richard Owlett wrote:
>> > I'll bite ;}
>> > When is it the right tool?
>> 
>> When you're using it to convert ebcdic to ascii, while swapping bytes and
>> reblocking an ancient file from a barely readable archival tape.
>> 
>> > When is it not?
>> 
>> When copying a file.
>> 
> When copying a file and writing it to another medium, perhaps eg when writing
> a DVD .iso file directly to a USB stick, it's ideal.

Not sure about ideal:

cat /dev/sdb

is one char longer than

dd /dev/sdb

but it's often faster (you can speed up `dd` by providing a larger
`bs=` argument, but then you've lost the length advantage ;-)


Stefan



Re: AppleWEBkit

2021-05-12 Thread Stefan Monnier
to...@tuxteam.de [2021-05-12 09:42:33] wrote:
> On Wed, May 12, 2021 at 03:28:18AM -0400, Gene Heskett wrote:
> [...]
>
>> The recent thread involving the mac G5's intrigues me though, so its war 
>> story time:
>> 
>> We bought a pair of them in a quite graphically capable editing kit for 
>> commercial production, paid $25G's for the pair [...]
>
> This G is for "grand", not "giga", right?

And here I was, thinking we were talking about the acceleration you
experience when the price finally hits you,


Stefan



Re: Suspend on old MacBook with fresh Bullseye installation

2021-05-11 Thread Stefan Monnier
Andreas Tille [2021-05-11 20:18:55] wrote:
> Hi Stefan,
>> > I also used hibernate as well as the LXDE controls or closing the lid -
>> > the result is always the same:
>> >  
>> >   1. Black screen
>> >   2. Network disabled (ssh session freezes, no ping from other host)
>> >   3. Fan keeps on moving silently
>> >   4. Can't wake up to normal operation neither by
>> >   - opening the lid
>> >   - pressing any key
>> >   - pressing power button 
>> 
>> Hmm... hibernate "should" work reasonably well, regardless of the
>> underlying firmware.  Are you booting via EFI or via the "legacy BIOS"
>> (which at some point was referred to as BootCamp)?  You may want to try
>> the other one, just to see if it helps.
>  
> I used the installer default which is EFI.  I admit I have no idea how
> to convince that MacBook to use something else.  Its the first time I've
> got my hand on any Apple hardware - so no idea how I can switch to
> something else from hardware side (nor whether I need to change
> something on the Debian installation.

To use the "legacy BIOS" style boot, you need:
- setup an MBR (ideally you have less than 4 partitions, so you can
  have a simple MBR with one pseudo-partition which is the "EFI
  protective" thingy and then the other partitions as before).
  You can set such an MBR without having to fully reinstall
  or repartition.  rEFIt used to come with a tool that did that, IIRC.
- install the non-EFI grub on your disk.
- potentially tell your Mac explicitly to use this boot method.
  A tool like rEFIt can be used (rEFIt is old, it's been superseded at
  least by rEFInd and probably others by now but I haven't kept up, sorry).


Stefan



Re: Suspend on old MacBook with fresh Bullseye installation

2021-05-11 Thread Stefan Monnier
> I also used hibernate as well as the LXDE controls or closing the lid -
> the result is always the same:
>  
>   1. Black screen
>   2. Network disabled (ssh session freezes, no ping from other host)
>   3. Fan keeps on moving silently
>   4. Can't wake up to normal operation neither by
>   - opening the lid
>   - pressing any key
>   - pressing power button 

Hmm... hibernate "should" work reasonably well, regardless of the
underlying firmware.  Are you booting via EFI or via the "legacy BIOS"
(which at some point was referred to as BootCamp)?  You may want to try
the other one, just to see if it helps.

How did you initiate the hibernate?

What does `cat /sys/power/disk` say?
Try changing it, e.g. via `echo shutdown >/sys/power/disk`
(that won't help for suspend but it might help for hibernation since
step 3 above suggests that the machine didn't *really* turn itself off).

> I was wondering about the black screen and suspected that the video
> driver might be responsible for the problem.  Despite I'm perfectly fine
> with the nuoveau driver I simply tried nvidia-legacy-340xx-driver (only
> in unstable) hoping that this might solve the suspend problem.

FWIW, I'd expect `nouveau` to behave a bit better w.r.t suspend/hibernate.


Stefan



Re: PC fan getting very loud because of CPU load

2021-05-08 Thread Stefan Monnier
deloptes [2021-05-08 21:33:47] wrote:
> Dan Ritter wrote:
>> It is also the case that fans are cheap. Replacing one for
>> $10-20 is generally good for another 5-10 years.
> But the question is why it runs > 100% - the fan is not that important in
> the case. It could be replaced with not so loud one, but the CPU will still
> run at 100+ % when firefox is running.

Actually, it could be related: on some of my laptops, when the fan
wasn't working well, I often saw both 100% CPU and a loud fan at the
same time, and both were due to the CPU's temperature (the system reacts
to too-high a temperature by increasing the fan speed, of course, but
also reducing significantly the CPU's frequency which in turn can cause
it to hit 100% usage even if there isn't a terribly high load).


Stefan



Re: What is the best (and free) Linux softphone?

2021-05-05 Thread Stefan Monnier
I think Linphone used to have a CLI interface, so you might want to look
into it.  I haven't used it in a long time, so I don't know what is its
current status.  I basically gave up on SIP (partly because of very
spotty support for encrypted communications and for async messages like
SMS) and recommend Mumble, Matrix, or Jitsi for discussions over IP
instead; tho none of those support the "call" feature, so they require
some previous arrangement about when the communication will take
place :-( (Matrix trie(s|d) to provide that, but it doesn't really work
for me).


Stefan



Re: [OFFTOPIC] Re: how to use fetchmail with MS Office 365 / davmail?

2021-05-04 Thread Stefan Monnier
> I'm also interested to know how good a service you actually get within
> buildings, where most of us are most of the time. I see that wireless
> repeaters are recommended according to a home's floor area. Are they
> repeating 30GHz round the house, or conventional 2/5GHz? If the
> latter, there's no need for replacing any of your normal Wifi devices
> at all. You just get a cell-modem instead of a cable- or ADSL-modem.

I just hope you're right ;-)


Stefan



[OFFTOPIC] Re: how to use fetchmail with MS Office 365 / davmail?

2021-05-03 Thread Stefan Monnier
> It doesn't seem sensible to put a cell-connection into each TV
> when they're all immobile. OTOH cars and pets go places.
>
> And is 20GB of data per day a "reasonable usage" on a mobile data plan?
> Whereas 1TB per month on a fixed line is quite normal.

These arguments seem stuck in the present.

After all we already have "stationary cell phone" services to replace
land-line phone services and I think most "phone" companies would be
looking forward to a future where there's no "last mile" any more, there
are only cell towers instead.


Stefan



Re: Ghost cronjob

2021-05-03 Thread Stefan Monnier
Mart van de Wege [2021-05-03 20:11:25] wrote:
> Stefan Monnier  writes:
>>> root@galahad:~# grep btrbk /etc/ -rl
>>
>> Have you `grep`d in `/var/` as well?
>> [ E.g. `/var/spool/crontabs` ]
>>
> Yep, nothing there, aside from the usual suspects (apt & dpkg files).
>
>>> And yet I find this in /var/log/btrbk.log:
>>>
>>> 2017-03-12T20:16:28+0100 startup v0.24.0 - - - - # btrbk command line 
>>> client, version 0.24.0
>>
>> Any other mention of activity around that time in some other log file?
>>
>>
> Not that I can see. I am going to see what patching btrbk to log PPID
> shows up tonight.

My usual "trick" is to log a full `ps --forest -ef`.
[ Instead of patching, you can rename it and replace it with a script
  that runs the "real" btrbk.  ]


Stefan



[OFFTOPIC] Re: how to use fetchmail with MS Office 365 / davmail?

2021-05-03 Thread Stefan Monnier
>> > There: now your smart-ass TV is a monitor again.
>> At least until they start using a cell-connection for Internet access
>> (which would seem only natural in the world of TVs, which historically
>> got their programs over the air) :-(
> Cars do that already. Why shouldn't TVs? Or pet collars?

Don't know about pet collars, but for TVs I'm pretty sure it's only
a question of time.


Stefan



Re: Ghost cronjob

2021-05-03 Thread Stefan Monnier
> root@galahad:~# grep btrbk /etc/ -rl

Have you `grep`d in `/var/` as well?
[ E.g. `/var/spool/crontabs` ]

> And yet I find this in /var/log/btrbk.log:
>
> 2017-03-12T20:16:28+0100 startup v0.24.0 - - - - # btrbk command line client, 
> version 0.24.0

Any other mention of activity around that time in some other log file?


Stefan



Re: HTML syntax.

2021-05-03 Thread Stefan Monnier
> Is there an editor which checks that HTML opening and closing tags are 
> paired and nested properly.  An opening tag without matching close, a 
> closing tag without matching open and crossed tags should be flagged 
> by line number or color.

Since Emacs's built-in `nxml-mode` does that, and Emacs is far from
being the most popular HTML editor, I'd tend to assume that any editor
worth its name provides that feature.


Stefan



Re: how to use fetchmail with MS Office 365 / davmail?

2021-05-03 Thread Stefan Monnier
> There: now your smart-ass TV is a monitor again.

At least until they start using a cell-connection for Internet access
(which would seem only natural in the world of TVs, which historically
got their programs over the air) :-(


Stefan



Re: When to reboot after dist-upgrade?

2021-05-03 Thread Stefan Monnier
> I'm trying to distinguish when a system reboot is an absolute need
> and when it is absolutely safe to keep the system running/working
> after a `sudo apt-get update && sudo apt-get dist-upgrade`, once
> I have already performed a complete restart of all needed services
> through `sudo needrestart' options in Debian testing.

I think the only reason you might "absolute"ly need to reboot is if
a newer kernel fixes a security hole to which your current system is
vulnerable.  Since your kernel might include security holes to which you
happen not to be vulnerable (e.g. because you just happen not to use
that part of the kernel, or for some other reason), it's in generally
extremely difficult to determine with a 100% certainty whether or not
a reboot is "absolute"ly needed.

Me, I basically only reboot in 2 cases:
- the power went out


Stefan



Re: Google vs. DDG

2021-05-02 Thread Stefan Monnier
deloptes [2021-05-02 17:44:42] wrote:
> Stefan Monnier wrote:
>> For these kinds of searches, I tend to go to Wikipedia first.
>> One of the reasons is that I can do that via the Wikipedia app which is
>> much more snappy than my browser on my smartphone.
>> Another is that I have much more confidence in the Wikimedia foundation
>> when it comes to treating me as a product.
> What confidence? Anonymous people correcting each other posts. You can't
> trust scientists anymore not to speak of Wikipedia

Please re-read what I wrote.  I did not talk about confidence in the contents.


Stefan



Re: Google vs. DDG

2021-05-02 Thread Stefan Monnier
>> when I just try to search for "language:en stieglitz" (without the
>> quotes)
> YES! That's what I was looking for. I wasn't sure DDG had this
> feature (although I hoped for it).

Sadly, it doesn't work when combined with `!w`.
[ and I think it'd make sense to allow the shorter `lang:de`.  ]

And of course, the term `language` is an obvious English bias.


Stefan



Re: Google vs. DDG

2021-05-02 Thread Stefan Monnier
> For example I wanted to know what is Stieglitz in German - it is kind
> of bird, but I wanted to know how it looks like.  DDG results did not
> even come close to a bird.

For these kinds of searches, I tend to go to Wikipedia first.
One of the reasons is that I can do that via the Wikipedia app which is
much more snappy than my browser on my smartphone.
Another is that I have much more confidence in the Wikimedia foundation
when it comes to treating me as a product.


Stefan



[OFFTOPIC] Re: pci 0000:00:01:0: MSI quirk detected; subordinated MSI disabled ...

2021-04-30 Thread Stefan Monnier
> Now I wonder how this might enable random access to the nth
> character.  I will keep looking around.

Another part of the question is: why would someone give you the position
information in terms of characters rather than in terms of (say) bytes,
or words, or ...


Stefan



Re: how to use fetchmail with MS Office 365 / davmail?

2021-04-30 Thread Stefan Monnier
> viewing material because it's about ten years old. But when we
> inevitably replace it and have no choice about accepting a 'smart' TV,

There are still normal TVs around.


Stefan



Re: Google vs. DDG

2021-04-28 Thread Stefan Monnier
>> FWIW, I'm pretty sure that such anectodal evidence is of no importance
>> because you can also come up with examples where the situation
>> is reversed.
> Can you?

I meant "you" in a very general sense: I'm pretty sure it's possible,
but no, I haven't done the necessary work (and I'm not very interested
in doing it either).

> I'm interested in examples of such cases.

I suspect that a good way to find such an example might start by trying
to think of websites which Google would specifically want to avoid
(e.g. for legal or political reasons) while Microsoft wouldn't.
This might be because it's related to a country which kicked out one of
the two while the other managed to sign some kind of agreement or
something like that.

> Incorrect (or at least outdated):
>
> DuckDuckGo gets its results from over four hundred sources. These
> include hundreds of vertical sources delivering niche Instant Answers,
> DuckDuckBot (our crawler) and crowd-sourced sites (like Wikipedia,
> stored in our answer indexes). We also of course have more traditional
> links in the search results, which we also source from multiple
> partners, though most commonly from Bing (and none from Google).
>
> https://help.duckduckgo.com/duckduckgo-help-pages/results/sources/

Thanks for the precisions.


Stefan



Re: Google vs. DDG

2021-04-28 Thread Stefan Monnier
> Here's a sort of example I just ran into. When trying to find
> information about Thetis hardware security keys, DDG simply couldn't
> find the company's website: searching DDG for "thetis key" turns up (in
> the first page of hits) a bunch of Amazon listings, and a bunch of
> reviews of, and articles about, security keys that mention Thetis.
> Searching for the same thing on Google, OTOH, returns the company's
> website (https://thetis.io) as the first hit (along with a convenient
> list of pages on the site).

FWIW, I'm pretty sure that such anectodal evidence is of no importance
because you can also come up with examples where the situation
is reversed.

This is simply because the subset of the internet that is indexed by the
two search engines is not simply in a subset relation.  So the question
is not whether such things happen, but how often they happen for your
use-case one way compared to how it happens for your use-case the
other way.

And of course this question is only relevant as one of the properties
distinguishing the two search engines.  Obviously, the purpose of DDG is
not to give better search results.

BTW, as far as I know, DDB doesn't do its own indexing but it relies
internally on Bing, so in the above is explained by the difference
between Bing and Google.  Technically they could probably just as
well rely on Google (or on both).


Stefan



Re: OT: Freestanding spreadsheet program?

2021-04-24 Thread Stefan Monnier
> I think he also wanted something that doesn't require a desktop environment.

AFAIK Gnumeric works fine in "naked X11".


Stefan



Re: OT: Freestanding spreadsheet program?

2021-04-23 Thread Stefan Monnier
> Back in the good (bad?) old days of TRS-80, all we had was VisiCalc. Simple.
> Today, is there a useful spreadsheet program that does not rely on all the
> baggage associated with either an "office suite," or
> a "desktop environment?"

I can mention `gnumeric` and if you're into Emacs I can also also
suggest SES.


Stefan



Re: installing two versions of a user application

2021-04-23 Thread Stefan Monnier
> You probably want to look inside the control archives rather than the
> data archives:
>
> https://en.wikipedia.org/wiki/Deb_%28file_format%29
>
> If both control archive files are using a package name of
> "master-pdf-editor", you can try Stefan's suggestion of renaming one
> and seeing if they'll then coinstall.

Exactly: the immediate problem you're encountering is that the package
names themselves (regardless of the files they contain) collide.
This info is the `control` part of the .deb archive.

The way I do it is to use Emacs to open the `foo.deb` file.
This should show you something like:

MMode  Size  Date   Filename
- --  --    
  -rw-r--r--   4  2021-03-01 12:23  debian-binary
  -rw-r--r--1128  2021-03-01 12:23  control.tar.xz
  -rw-r--r--  615380  2021-03-01 12:23  data.tar.xz
- --  --    
  616512 3 files

where the `data.tar.xz` part contains the actual files and
`control.tar.xz` is the part you want to modify.
So you select the `control.tar.xz` which itself will look somewhat
similar and will contain a `control` which you can open and in which
you'll find presumably something like:

Package: master-pdf-editor
[...]

followed by other kinds of metainfo about the package.  Just edit that
first line (e.g. by adding a `4` or `5`), then do `C-x C-s` to save this
`control` back into `control.tar.xz`, then kill this buffer with `C-x
k RET`, then save the new `control.tar.xz` back into the actual foo.deb
file with `C-x C-s`.

Then retry to install those debs with:

dpkg --install foo.deb

I recommend you do those modifications on a copy of the file, just so
its easier to recover from mistakes.


Stefan



Re: installing two versions of a user application

2021-04-23 Thread Stefan Monnier
> $ sudo apt install ./master-pdf-editor-4.3.89_qt5.x86_64.deb
> ./master-pdf-editor-5.7.08-qt5_included.x86_64.deb
> Reading package lists... Done
> Building dependency tree
> Reading state information... Done
> Note, selecting 'master-pdf-editor' instead of
> './master-pdf-editor-4.3.89_qt5.x86_64.deb'
> Note, selecting 'master-pdf-editor' instead of
> './master-pdf-editor-5.7.08-qt5_included.x86_64.deb'
> master-pdf-editor is already the newest version (4.3.89-2).
> The following packages will be upgraded:
>   master-pdf-editor

I suggest you look inside those .deb files: they seem to both name the
package they install `master-pdf-editor` but dpkg does not allow
installing different versions of a given package at the same time.
So you'll need to edit those `.deb` to change the package name from
`master-pdf-editor` to (say) `master-pdf-editor4` and
`master-pdf-editor5`, for example.


Stefan "who does that kind of thing in order to install emacs25,
emacs26, and emacs27 at the same time"



Re: Strange emacs behavior after upgrade to bullseye

2021-04-20 Thread Stefan Monnier
>> Emacs implements its locking protocol using symlinks with names
>> that look like `.#` and whose content looks like
>> `u...@host.pid:BOOT_TIME`.
>
> Ah, good old dot-locking.  Well, perhaps the OP can test whether it's
> possible to create a symlink in that directory.

That's probably part of the problem, yet the error doesn't seem to come
from the code which would *create* the symlink but rather the one that
would delete it :-(


Stefan



Re: Strange emacs behavior after upgrade to bullseye

2021-04-20 Thread Stefan Monnier
> Because the error says it cannot use LOCKS.
> Because you can't do Unix file locking on a non-Unix file system.
> The error does NOT say "Permission denied".

FWIW, the error comes from Emacs's own locking code which doesn't seem
to use unix file locking, so the problem comes from elsewhere.

Emacs implements its locking protocol using symlinks with names
that look like `.#` and whose content looks like
`u...@host.pid:BOOT_TIME`.


Stefan "still not sure exactly where it goes wrong"



Re: Smart TV on WiFi as Extra Display

2021-04-19 Thread Stefan Monnier
> not support stretching your desktop to it) and it has a Java viewer 
> applet that can be used to connect to it from a web browser.

There's a good chance the browser hardcoded in the TV doesn't support
Java applets.


Stefan



  1   2   3   4   5   6   7   8   9   10   >