The officially official Devuan Forum!

You are not logged in.

#1 Re: Desktop and Multimedia » XBell (beep) muted or disabled on Xorg (daedalus) » 2025-01-11 20:28:31

checkbox at the bottom for enabling system sounds.

"system sounds" are not the same as the dedicated PC speaker bell.

@cade: We are talking about the old-school hardware PC speaker, not some soundcard emulation, right?
Is the pcspkr module loaded, and does 'beep' work? Do you have an appropriate evdev link (i.e. /dev/input/by-path/platform-pcspkr-event-spkr)?

You're not the first to ask after this here, see this thread.

#2 Re: Hardware & System Configuration » [SOLVED] Entry in fstab but Authentication is still required to mount partition » 2025-01-09 11:30:24

I'm always looking for quick but sound solutions with easily sourced comprehensive technical documentation where necessary.

There's nothing quicker than 'man' (and 'apropos', to find the right manual), since those manuals are already on your system. '/' for string search in the viewer (it's pretty much just the 'less' pager).
More comprehensive manuals are often available with 'info', though that varies by package.

I still have no idea why everyone goes straight for random blog posts and online copies of manuals... 'man' is faster, available almost everywhere, and doesn't require a web browser.

Does what Sebastien describes present itself for my scenario for a tech illiterate user, who just needs to store their own media and document files on the partition?

It presents as a very niche attack vector, requiring a specific set of conditions and permissions. For most people it's irrelevant.

What would you all suggest?

I'd suggest you do whatever you like. There's no real reason not to set 'nodev' on a non-root "media and documents" filesystem, so you might as well.

Similarly, suid and nosuid are not expanded on in the first 3 pages I mentioned.

It instructs the kernel whether or not to honour the suid and cap_sys_admin bits on executables, what else is there to "expand on"?

Exploiting suid executables also requires specific conditions and root permissions at some point, but again, since disabling suid does no harm on a data filesystem, you might as well.

Am I correct in guessing that it was the nosuid that was preventing the default rw option from being enabled?

I highly doubt it. It was likely the lack of the 'user' flag, since that is exactly what it's for. 'rw' is a default mount option anyway.
It could also be some weirdness of your file manager of course, but since you didn't specify what you're using, in which desktop environment, or what kind of "Authentication dialogue" you saw, who would know.

One additional trap to be aware of: If you use the 'user' flag, only the user who mounted the filesystem will be allowed to unmount it. That means the combination of 'user' and 'auto' will likely result in root mounting the filesystem at boot, and authentication being required for an unprivileged user to unmount it later. See the 'users' and 'group' flags for alternatives.

For reference, entry for the 4TB spinning-rust RAID1 in my desktop is:

LABEL=rust                      /mnt/rust       ext4    noatime,user,noauto        0 2

Which results in:

/dev/md2 on /mnt/rust type ext4 (rw,nosuid,nodev,noexec,noatime,user)

And it mounts and unmounts just fine from a GUI file manager (dolphin), without elevated permissions.
Note:

man mount 8 wrote:

user
...
this option implies the options noexec, nosuid, and nodev

So we don't need explicit noexec, nosuid, or nodev, and:

man fstab wrote:

The kernel default is usually rw, suid, dev, exec, auto, nouser, and async.

So most of the time you don't need to add explicit (filesystem specific) 'defaults' for ext4 either.

If such a partition is to be shared between 2 Linux installations, can both installations own the partition or would it be better to give the user Read and Write access to the Root directory of the Data partition within each OS?

An OS installation doesn't "own" a filesystem in any way (unless it's ZFS, but I digress), *nix is a network OS at heart and permissions are all about users and groups.

Who can do what and where simply comes down to whether the users effective UID/GID (i.e. numeric user and/or group identifier) matches the file/directory owner UID/GID.
If you set up your user account to have the same UID on both installs, that user will have the same permissions for the same files regardless of which OS you boot... And I suggest you do set up identical usernames with the same UID, because anything else is a kind of a confusing PITA (e.g. "bob" (UID 1000) on machine 1 can write to the disk, but "bob" (UID 1001) on machine 2 can't, so Bob is annoyed).

Aside:

RedGreen925 wrote:

The mount point (LinuxDataExt-4TB) should be owned by the user to prevent file/directory creation problems.

No, no it should not. Mountpoint directory permissions should be root:root, they will be replaced with those of the root of the mounted filesystem (i.e. mount the filesystem, then chown, not the other way around).
Granting write on a mountpoint directory to an unprivileged user is just a super effective way to let that user fill up root when the filesystem is not mounted.

#3 Re: Hardware & System Configuration » [SOLVED] Entry in fstab but Authentication is still required to mount partition » 2025-01-09 04:11:21

nothing but defaults

And if the OP wants those mount options? Those are all perfectly reasonable flags to set on a "data files only" filesystem.

options with an ext4 system not the vfat options now set on it
...
The arch wiki shows those used with a vfat file system

VFAT has nothing to do with anything, and no it doesn't. It simply shows noexec,nosuid,nodev as part of a generic example fstab.
Those flags are reasonable for pretty much any filesystem that doesn't contain executables, suid executables, or device nodes (this is usually suggested as "security hardening", and of rather of dubious benefit). That they happen to be set on a vfat filesystem in that specific example means nothing.

Not only is there is nothing vfat specific in the OP's mount flags, some of them aren't even valid for vfat anyway. OTOH, they are all both valid and perfectly sensible for (non-root) ext4.

the extras were screwing it up

Rubbish. data=ordered and relatime are superfluous as they are defaults these days, and the rest are quite ordinary flags for a non-root filesystem.

There is nothing "complicated" here, 'man fstab', 'man mount', and 'man ext4' cover pretty much anything one would want to know... including the the 'user' mount option, which is most likely what the OP has been after all along.

Why does nobody read the manuals?

#4 Re: Installation » [SOLVED] Sources.list hates me! » 2024-12-31 18:18:33

greenjeans wrote:

I actually made a right-click extension "Edit file as Root" years ago that opens files in Pluma for me to edit, super handy. Much better than just working in the root account for long periods of time I reckon.

steve_v:
*opens a bunch of root-owned files in kate tabs, as unprivileged user*
*screws around for as long as needed, as unprivileged user*
*hits "save"*
*gets polkit authentication prompt*
*profits*

Much better than just having a GUI editor running as root for long periods of time I recon.

Aside, everything any of us might do in a GUI editor can also be done in emacs at a TTY.

#5 Re: Installation » [SOLVED] Sources.list hates me! » 2024-12-31 12:19:46

kdesudo was replaced with a config option for kdesu some time ago, but kdesu and friends are living on borrowed time because wayland (intentionally) doesn't support running GUI applications as another user at all.

While I generally disagree with wayland's intentionally missing features, in this case nothing of value is really lost.
GUI applications that need to do something with elevated permissions should handle this themselves (with e.g. a polkit auth prompt when writing a file), and I have never seen any good argument for running entire GUI programs as root. KDE applications (e.g. dolphin, kate) already do this, and I expect other DEs do as well or soon will.
If synaptic doesn't, then that's a bug that should be filed against synaptic.

As for mousepad... I'm still completely baffled as to why anyone would want to run a GUI text editor as root. Use a modern GUI editor that handles saving files with elevated permissions... Or just use GNU nano, because there's zero need for a GUI editor when working with config files to begin with.

Running libreoffice as root just to edit a text file is especially insane IMO, and I see no reason any desktop environment should support it or provide tools to enable it. kdesu and gksu were hacks, situations where they are useful (as opposed to pathological) are few and far between, and there's no reason for them to continue to exist.

"Run as administrator" for GUI applications (along with GUI logins as root) is a terrible idea inherited from Winblows and perpetuated by ex-Winblows users. It needs to die, and sooner rather than later.

#6 Re: Installation » [SOLVED] Sources.list hates me! » 2024-12-31 07:36:03

Red_Fir wrote:

the initial attempt was made as root and it came up in Libre Office Writer when moused.
Subsequent efforts made as su'd user in Mousepad

The only thing more confusing here than the use of libreoffice for editing a simple text file... Is the use of libreoffice as root.
Just.. Why? Use a real text editor, from an elevated terminal. Nano, vi(m), emacs, mcedit, all work just fine without any weird encoding issues or the hazard of runing enormous GUI bloat as root. Nano has been the default $EDITOR for years, and with good reason.

#7 Re: Hardware & System Configuration » (devuan 5.0) update gcc+? » 2024-12-31 07:27:10

/usr/lib/x86_64-linux-gnu/libstdc++.so.6: version `GLIBCXX_3.4.32' not found (required by minitube)

The correct answer to this class of problem is not screwing around with your install (aka creating a frankendebian) or pulling most of a different OS as a container, it's recompiling the problemtic software against installed libraries.

is there another way to upgrade this package easily without necessarily upgrading to testing

No, and attempting to do so will almost certainly bork your install.
The latest QT5 version of minitube is already in the Devuan repositories, I can't think of an "easier" way than using that. 4.0 appears to be nothing but a QT6 port and some cosmetic changes.

#8 Re: Other Issues » How recompile Kernel ? ( fresh install ) » 2024-12-22 08:47:48

the man don't about risks/error with -j

Indeed, because that option simply specifies the number of parallel jobs to run. The only "error" possible is that the build fails because things are compiled out-of-order (which means upstream needs to fix their makefile / build process), and the only "risk" is wasting time and trying again with those options omitted.
To put it another way: The function of '-j' and friends is not complicated, and I have no idea why anyone would need handholding over such an option. Try it and find out, like everyone else.

If I understand well the document is either 4.4.1 or 4.4.2 method , right ?

The choice is whether to use DKMS or module-assistant... That choice is up to you.

#9 Re: Other Issues » How recompile Kernel ? ( fresh install ) » 2024-12-20 22:23:03

Does someone already tried with -j$(nproc) ?

man make

#10 Re: Other Issues » How recompile Kernel ? ( fresh install ) » 2024-12-20 14:56:09

Same as you would on Debian.
OTOH, compiling a kernel is much the same on any distro and hasn't changed significantly in decades, the only real difference here is using the debian specific make targets (included in linux-source-[version]) to build .deb packages rather than manually installing the image/modules.

tried few way but failed...

And if you also fail to elaborate, nobody will know why.

#11 Re: Other Issues » [SOLVED] modprobe ikconfig -> FATAL: Module ikconfig not found in directory » 2024-12-15 09:09:08

any ideas

Only the obvious: Read the documentation (or a good wiki). The module this option builds is not named 'ikconfig'.

And the blindingly obvious:

CONFIG_IKCONFIG=m y

Selecting 'y' will build the option into the kernel image, so there definitely won't be a module to load.

Aside, no idea why you would want IKCONFIG as a module anyway. There's sod-all to be saved by not including a copy of .config in the kernel, so you might as well build it in.

#12 Re: Documentation » How to: Devuan 5 Daedalus an pipewire » 2024-12-15 07:16:29

Quite possibly tongue

While I do run plain ALSA on systems that have no need for a sound-server, something of the latter nature (i.e pulseaudio or pipewire) is practically a requirement if you want "modern" GUI-centric on-the-fly source/sink and sample-rate switching for things like bluetooth audio peripherals.
ALSA can of course do most of that itself, but it's somewhat awkward and requires a bit of experimentation and messing about in poorly-documented configuration files... Which few can be bothered with any more, because much like the extended "I just copy-paste and hope it works" in this thread, "If I can't click on it, it doesn't exist" is apparently a thing. roll

Pipewire is also essentially unavoidable if you want desktop audio/video capture in a wayland session, and as with all the other functionality still missing from the wayland spec, that's the way freedesktop intended it.

#13 Re: Documentation » How to: Devuan 5 Daedalus an pipewire » 2024-12-15 03:53:56

Why not start by enabling some logging and doing at least basic troubleshooting yourself? All I see here is "it stopped working".
Use of log files is in the daemon manual, and the very least, try the same command you're using in xinitrc manually and see if it's throwing any errors on stdout/stderr.

Nobody is psychic, and that debian bug report is so devoid of information as to be practically useless. Are you really running sid/experimental? If so, you should be prepared for random borkage and prepared to do some digging... That's why it's called "experimental".

And no, no I'm not going to look into this hose any of my systems by enabling the sid or experimental repos. If you want an install that doesn't regularly break after package updates, run stable (and backport things you want from sid yourself).

#14 Re: Installation » Devuan 5 not Booting on 2005 Toshiba Tecra A7 Intel Centrino Duo » 2024-12-10 10:25:24

can't find online where the battery resides on the Motherboard.

In laptops it's often remote. Follow the wires to the battery wrapped in shrink-sleeve stuffed in some random corner of the chassis.
There's a small chance this could have an integral battery (e.g. some variant of the infamous Dallas RTC chip), but I doubt it. Either way, it's likely dead by now. Open the machine up and have a look, it's probably obvious.

"could be caused by a failing hard drive"

Or a dodgy connection, as the error message indicates. Oxidation and/or general crud in SATA connectors can lead to all manner of fun.

Would using GParted from the Live DVD by using the Test feature on the drive's partitions be the same as using "chkdsk /f" in a Windows environment?

Likely close enough, assuming that the live media has ntfs and/or fat (you don't say what filesystem you're running) fsck tools.
That said, a filesystem check probably isn't what you want for suspected hardware problems anyway. Pull the SMART diagnostics from the drive, then run a long (media read) self-test or badblocks. See 'man smartctl' and 'man badblocks' respectively. Both are considerably slower and more thorough than chkdsk.

Could I put the Hard Drive in an enclosure and boot to Windows with the Dell Laptop I have and use "chkdsk /f" from there on the Toshiba's Hard Drive?

Probably. Really not sure why we're talking about Windoze and Windoze filesystem tools to begin with though.

Aside, also really not sure why all the buggering about with an old mechanical hard drive.  It'll be miserably slow compared to even the cheapest SSD, and considering its age there's a very good chance it's just plain dodgy.

I suggest you forget about the original drive and software (i.e. image the drive if there's stuff you want on it, then bin it), slap some cheap second-hand RAM and a bargain-basement SATA SSD in, then move on to getting a real OS installed.
The box is not worth the time and effort trying to source "officially supported" or branded components, and messing with 20 year old hard drives is a mugs game unless you're restoring rare vintage hardware... which this is not.

#15 Re: Installation » Devuan 5 not Booting on 2005 Toshiba Tecra A7 Intel Centrino Duo » 2024-12-09 04:51:42

Pretty much all of the above are questions specific to your particular machine, unless someone here has that exact model the best you will get is generic guesswork.

SATA 2.0, was introduced in April of 2004 and supports bandwidths up to 300 MB/s. So, maybe this Laptop could support SATA 2?

The ICH7 (best guess at the IO controller in this machine without one to play with) datasheet suggests SATA2 @ 300MB/s.
To determine current negotiated SATA link speed, boot linux and look in dmesg.
Even if the chipset is only SATA1, you'd still benefit from going to an SSD. Few mechanical laptop drives can really sustain 150MB/s, and none can beat an SSD when it comes to seek time. Any modern SATA drive should be backwards compatible with whatever the chipset supports.

higher capacity Hard Drive (WD 500GB Blue, 5400 RPM) was detected by the BIOS and could be recognised

AFAIK the only real limitation that applied in 2006 was 2TB. Most BIOS implementations from the time will actually work with larger drives, you just won't be able to address the rest of the space.

on page 17, Revision Number 002 (April 2006) indicates that support for 4-GB physical memory @ 667 MHz was included

DDR2 SODIMMs are cheap and widely compatible. Your best bet is to just buy some and try them.

Toshiba did make a 2GB module whose naming convention matches that of the 1GB module Part Number that is in the User Manual
...
There is also a PA3513U-1M2G PC2-5300 module for the 667MHz frequency. However, the RAM is actually unbranded.

There's usually no need to go after "branded" modules or those specifically mentioned in the manual, you just need supported technology, voltage, and (rarely an issue with desktop/laptop modules) buffering and number of ranks. Faster is always better, assuming 667MT modules aren't rare and at a premium price.

Still no reply from the Ebay seller in China.

Considering the item value, no ebay seller is going to know, care about, or make any effort to investigate compatibility with some random laptop from 2006.

Did installing the 2nd Hard Drive cause something to be changed in the BIOS?

At the least, it'll cause drive auto-detection to run and probably trigger an ESCD update. As for why it hangs on first-boot... Laptop BIOSes are weird and often buggy. Always have been, probably always will be.
Be thankful you don't have to deal with "system configuration changed, insert reference diskette" any more tongue

#16 Re: Installation » Devuan 5 not Booting on 2005 Toshiba Tecra A7 Intel Centrino Duo » 2024-12-05 11:22:21

mtbvfr wrote:

I saw 2 failed notices regarding apparmor. Is that critical?

No. For some truly bizarre reason Debian (and by extension Devuan) enables apparmor by default, but doesn't ship a default configuration for it that actually works properly.
Ignore it, disable it, or fix whatever it's moaning about, your choice.

mtbvfr wrote:

According to 2 early pages of the User's Manual the Laptop can use 2 x PC4200, 200-pin, SO-DIMMs up to 2048MB each but according to a page towards the end, on Memory Expansion it mentions that only the following memory modules can be used.

The former will likely be the chipset limitation (which can be confirmed elsewhere, e.g. intel docs), the latter will be configurations the OEM tested with.
It's not particularly uncommon for a board to support, at least theoretically, DIMMs that weren't readily available when it was released.

As for getting some, I usually just go to ebay for stuff like this, and eat the (fairly reasonable for small items like memory) shipping (to NZ).

#17 Re: Other Issues » Install packages / software to an offline machine... » 2024-12-03 11:12:43

blackhole wrote:

You cannot use apt / apt-get / aptitude to install a deb file directly.

AFAIK aptitude can't install from a file, but apt/apt-get has supported this for years. It'll even resolve dependencies for local packages (but not using local packages, at least not automatically, same as dpkg).

#18 Re: Installation » [SOLVED] Make run nftables on Devuan rpi » 2024-12-03 11:01:51

And what makes you think nftables is a daemon?

#19 Re: Installation » Power-profiles-daemon doesn't work on Devuan » 2024-12-01 10:48:25

Power-profiles-daemon is expected to be started on-demand via dbus activation... Which is obviously only going to work with systemd.
If something is trying to start it that way, you'll see errors to the effect of "nothing provides the net.hadess.PowerProfiles name on dbus" or somesuch.

On Gentoo we have an (openrc) init script to start PPD explicitly at boot, but Devuan appears to be doing the usual "just pull from Debian & don't bother to check it works properly" thing Devuan does, so the package only contains a systemd .service definition.
If you want it to work, you'll probably have to write a sysv init script for it yourself or come up with some other way to start it on demand.

As for TLP, TLP does far more than PPD, but it's more work to set up and isn't (currently) integrated into the major (GNOME, KDE) DEs.
It works fine for my needs on the one ancient laptop I have ever bothered to own though, same as it always has.

#20 Re: Installation » Devuan 5 not Booting on 2005 Toshiba Tecra A7 Intel Centrino Duo » 2024-12-01 10:31:12

rolfie wrote:

if only 512M of RAM is installed, this isn't good enough to start from

512MB is fine for a netinstall, though the live desktop will likely choke. What it's not good enough for is pretty much any modern graphical web browser.
Frankly, for anything <2GB I wouldn't be running a full DE (even XFCE), and for <1GB It'd be a straight-up CLI-only install.

That's not to say you can't run a GUI with 512MB RAM, just that you probably won't like it very much.

#21 Re: Installation » Devuan 5 not Booting on 2005 Toshiba Tecra A7 Intel Centrino Duo » 2024-12-01 03:53:01

mtbvfr wrote:

How did you manage to have "mtbvfr wrote:" entered into the Message box?

By manually copy-pasting the text and the username like it's 1999, because golinux childishly disabled the quick-quote functionality some time back on account of a few users excessive use of it annoying him.

The full quoting annoyed a few (or more accurately, one) people. Removing the quick-quote links creates extra work any time a quote is warranted, and removes such niceties as timestamps and back-linking to the quoted post.

mtbvfr wrote:

I don't have another 32-bit machine to test with.

Most (if not all) 64bit systems should still be able to boot the 32bit install media (at least in legacy/CSM mode). Even if the live kernel won't boot for some reason, you should still at least get to the boot menu.

golinux wrote:

All the extra " and spaces are to make the structure viewable when posting. Maybe there is a better way to do that . . .

In any capable and half-way modern board software, the post editor has a "disable BBCode" checkbox. This one apparently does not, having it as a global toggle reserved for administrators... Which you are, and so should know.

#22 Re: Installation » Devuan 5 not Booting on 2005 Toshiba Tecra A7 Intel Centrino Duo » 2024-11-30 22:59:14

As already suggested, make sure the drive boots on something else... preferrably something more modern.
Blocksize shouldn't make any difference, and AFAIK XP didn't do any off-but-not-really-off "Fast startup" bollocks.

FWIW, I have found USB boot pretty hit-and-miss on machines from that era. There were several ways to make removable (CD, USB) devices bootable, and which worked with what BIOS was kinda down to luck.

Personally I'd just burn a CD, since that machine looks to have a fixed optical drive and CDs were still pretty much the standard install media at the time.
It'd also be interesting to try a bootable USB image from the period (e.g. knoppix or the like). If that works, you'd have somewhere to start with figuring out why the BIOS doesn't see the Devuan image as bootable.

greenjeans wrote:

I think I have some to work on that are even older

I have Jessie (last release supporting i586) running on a K6-2 from ~1998. It's quite usable, so long as you don't do anything silly like trying to run a modern GUI.
I'm sure it'd run Daedalus too, if only the "i386" build still meant "i386". neutral

IIRC I used PLOP on a floppy disc to boot the CD, because BIOS CD boot was flaky as all hell and USB boot hadn't been invented yet (also only USB 1.0, which would have been painfully slow anyway).

#23 Re: Off-topic » systemd v257 requires the merging of /bin and /sbin » 2024-11-15 12:24:46

I never said anything about capability. One can be the best programmer in the world, and still behave like an asshole.

Much of Lennart's code is excellent, some of it is even a good idea... The attitude exposed by his (and various other obvious parties as well) continual attempts to force everyone to get in line and do things according to "the plan" for GNU/Linux is that of an asshole.

His comments are replete with "I know best, do as I say" sentiment, and examples of "force x to do y, because it's the my future" are not at all difficult to find.
Usually these boil down to "Gentoo is resisting again, assimilate harder", or "Let's make not doing things our way as annoying as possible", while still waving the "it's open source, change it if you want" flag.

The headline is a garden-variety example of the latter. Doubtless any upstream bug reports or patches will be closed wontfix, with the usual "it's open-source, fork it if you really care that much"... Knowing full-well that this sanctimonious "taint" message is not quite annoying enough for anyone to actually do so.

Ordinarily all this would be a simple case of "dev is a twat, ignore", but he has unfortunately sold his ideas well enough to big corporate (IBM, Microsoft, etc.) that they now see them (systemd in particular) as the leverage they have always wanted to gain control of systems-level GNU/Linux development.

Much of the current direction within systemd is toward "trusted" computing, immutable operating systems, and homogenising the GNU/Linux landscape... These things sound like they are for the benefit of the end-user, but the natural end product - the computer as a cookie-cutter "secure" (for them) appliance, controlled by the vendor - is a holy-grail for big IT.

nothing but an MS registry for the Linux ecosytem

Uhh, sure... If you've been living under a rock for the last couple of decades. systemd does just a mite more than "registry" (a title better handed to the likes of, say, gconf) these days.

#24 Re: Off-topic » systemd v257 requires the merging of /bin and /sbin » 2024-11-15 08:47:52

systemd v257 requires the merging of /bin and /sbin tries to strong-arm distributions into another pointless filesystem layout change, by inserting scary-sounding messages to incite bug reports from clueless users.
FTFY^

In other news: The sky is blue, the sun rose in the east again, and Lennart is still an asshole.

#25 Re: Installation » New bpo Kernel Upgrade 6.10.11+bpo-amd64 fails to boot » 2024-11-10 11:36:11

A non-booting kernel or broken driver really isn't a big deal, and frankly I have no idea why people seem to get so flustered by it (see recent Debian drama WRT out-of-tree nvidia driver "completely breaking" roll systems).
Pretty much every distro keeps old kernels around (and generates bootloader entries for them) for exactly this scenario.

Even without that, you can just boot a livecd/usb and set root= appropriately or chroot in. Unless one is silly enough to encrypt / and not keep a record of the key OFC... Or use secure retarded-boot, but that's really a bootloader issue and only makes using live media more annoying.

Board footer

Forum Software