You are not logged in.
I'm working on new releases
Awesome news! Can't wait to see 'em, was playing with 32 bit Miyo all day yesterday. ![]()
One of these days, I hope to find time to build it with the latest version though and package it. For now though, the present version still works great on Jessie, Ascii, and Ceres...apart from having to reinstall it when upgrading from one to the next.
Been testing the new .99 version of the obmenu-generator script for a week or so, seems to function perfectly, still have not tested to see if it survives the ascii upgrade though.
Vuu-do 32 bit is about to happen finally, been working on it the last couple days, easier going now that I have a good bit of template in the 64 bit versions and lists of things and such.
First is a minimal Openbox version, may do a maxi for it too, and a mini JWM too at some point. I should have the new 32 done this week and uploaded. Working version idles nice and low like 32 should, around 88 mb of ram with conky and parcellite running, with those off I had it down to 80 mb. Nice and quick too.
Devuan does use /etc/os-release, but grub doesn't seem to check for that. (I grepped through all of /usr/lib/os-probes/mounted/ and /usr/lib/linux-boot-probes/ and couldn't find it.)
Yep, you have to manually add the lines to 90linux, but after that it picks it right up when you update-grub.
So as long as a distro provides an lsb release , grub should pick it up without any further code correct?
Current Refracta isos have /etc/devuan_version, but grub uses /etc/lsb-release instead, so it shows up as Refracta (or maybe Refracted Devuan). Ubunutu isos have /etc/debian_version, but I'm pretty sure it always shows up as Ubuntu in the grub menu (if you can find it.)
I actually haven't tried the default Devuan 1.0 iso, is there an /etc/lsb-release file in it now? I seem to remember you saying had to make one for Refracta...
Okay, so one way or another i'll get the os-prober package's failure to recognize Devuan fixed, but I need help, I need people who have other modern linux distros like ubuntu or mint etc., I need the identifier for the os, usually in /etc. For instance Devuan's is /etc/devuan_version and Debian is /etc/debian_version, Arch is /etc/arch_release.
This portion of the package hasn't been really updated in years, it still had distros that have been gone since 2003.
Here's what I have so far just adding Devuan (and Vuu-do, lol) and getting rid of the old discontinued ones:
#!/bin/sh
# Test for linux distributions.
set -e
. /usr/share/os-prober/common.sh
partition="$1"
dir="$2"
type="$3"
# This test is inaccurate, but given separate / and /boot partitions and the
# fact that only some architectures have ld-linux.so, I can't see anything
# better. Make sure this test has a high number so that more accurate tests
# can come first.
# Unless volumes to checked are already mounted, they will be mounted using
# GRUB's own filesystems through FUSE. Since these ATM doesn't support
# symlinks we need to also check in $dir/usr/lib* for distributions that
# moved /lib* to /usr and only left symlinks behind.
# TODO: look for ld-linux.so on arches that have it
if (ls "$dir"/lib*/ld*.so* || ls "$dir"/usr/lib*/ld*.so*) >/dev/null 2>/dev/null; then
if [ -e "$dir/etc/debian_version" ]; then
short="Debian"
long="$(printf "Debian GNU/Linux (%s)\n" "$(cat "$dir/etc/debian_version")")"
# RPM derived distributions may also have a redhat-release or
# mandrake-release, so check their files first.
elif [ -e "$dir/etc/devuan_version" ]; then
short="Devuan"
long="$(printf "Devuan GNU/Linux (%s)\n" "$(cat "$dir/etc/devuan_version")")"
elif [ -e "$dir/etc/vuudo_version" ]; then
short="Vuu-do"
long="$(printf "Vuu-do GNU/Linux (%s)\n" "$(cat "$dir/etc/vuudo_version")")"
elif [ -e "$dir/etc/altlinux-release" ]; then
short="ALTLinux"
long="$(cat "$dir/etc/altlinux-release")"
elif [ -e "$dir/etc/magic-release" ]; then
short="Magic"
long="$(cat "$dir/etc/magic-release")"
elif [ -e "$dir/etc/blackPanther-release" ]; then
short="blackPanther"
long="$(cat "$dir/etc/blackPanther-release")"
elif [ -e "$dir/etc/ark-release" ]; then
short="Ark"
long="$(cat "$dir/etc/ark-release")"
elif [ -e "$dir/etc/arch-release" ]; then
short="Arch"
long="$(cat "$dir/etc/arch-release")"
elif [ -e "$dir/etc/vine-release" ]; then
short="Vine"
long="$(cat "$dir/etc/vine-release")"
elif [ -e "$dir/etc/pld-release" ]; then
short="PLD"
long="$(cat "$dir/etc/pld-release")"
elif [ -e "$dir/etc/fedora-release" ]; then
short="Fedora"
long="$(cat "$dir/etc/fedora-release")"
elif [ -e "$dir/etc/redhat-release" ]; then
short="RedHat"
long="$(cat "$dir/etc/redhat-release")"
elif [ -e "$dir/etc/SuSE-release" ]; then
short="SuSE"
long="$(head -n 1 "$dir/etc/SuSE-release")"
elif [ -e "$dir/etc/gentoo-release" ]; then
short="Gentoo"
long="$(cat "$dir/etc/gentoo-release")"
elif [ -e "$dir/etc/turbolinux-release" ]; then
short="Turbolinux"
long="$(cat "$dir/etc/turbolinux-release")"
elif [ -e "$dir/etc/pardus-release" ]; then
short="Pardus"
long="$(cat "$dir/etc/pardus-release")"
elif [ -e "$dir/etc/kanotix-version" ]; then
short="Kanotix"
long="$(cat "$dir/etc/kanotix-version")"
elif [ -e "$dir/etc/slackware-version" ]; then
short="Slackware"
long="$(printf "Slackware Linux (%s)\n" "$(cat "$dir/etc/slackware-version")")"
elif [ -e "$dir/sbin/pkgtool" ]; then
short="Slackware"
long="Slackware Linux"
elif [ -e "$dir/etc/frugalware-release" ]; then
short="Frugalware Linux"
long="$(cat "$dir/etc/frugalware-release")"
elif [ -e "$dir/etc/lfs-release" ]; then
short="LFS"
long="$(printf "Linux From Scratch (%s)\n" "$(cat "$dir/etc/lfs-release")")"
else
short="Linux"
long="unknown Linux distribution"
fi
label="$(count_next_label "$short")"
result "$partition:$long:$label:linux"
exit 0
else
exit 1
fiOkay, tested one machine, Compaq CQ58, about a 2012 model, dual-core AMD APU (at a tortoise-like and underwhelming clock speed of 1.0 ghz), no issues with Devuan openbox 64 bit, mouse and keyboard work fine so far.
Wow, this seems serious, looks like it may not occur on all machines, I will logout and test one of my other partitions and post back.
EDIT: Hey, can you post the model/make of your machine?
Tagging this thread as this is of interest to me too, I need to learn more about setting a system up for other keyboards, locales, and languages.
Will be trying this out on mine later, thanks for posting!
I was just posting to let everyone know that I did a fresh install on a Thinkpad T440 with xfce4 and it would not sleep on lid close. Using this thread and some searching I got it to work.
First was running visudo as root to allow my user to run pm-suspend
# username suspend permission username ALL=NOPASSWD: /usr/sbin/pm-suspendthen I created an acpi event by creating /etc/acpi/events/lid
event=button/lid action=/etc/acpi/actions/lid_suspend.sh %e(make sure to make it executable with "chmod +x lid")
Finally I created the folder and file /etc/acpi/actions/lid_suspend.sh
#!/bin/bash if [ "$3" = "close" ] ; then /usr/sbin/pm-suspend fi(chmod +x lid_suspend)
I did "service restart acpid" for good luck and it worked!
Thank you everyone for your help!
MiyoLinux wrote:All of my MiyoLinux systems are now on ceres. Lo and behold, Radio Tray is in the repository. That alone is enough to make me upgrade to Ceres! LOLOL!
I'll have to check if it's available in ascii.
Well I feel like an idiot. Radio Tray is available in Jessie.
LOLOLOL! Not sure how I missed it before...
I have to have my internet radio too, somafm usually, Radiotray is a nice program.
I use Exaile, has i-radio built in and even has somafm and some others pre-programmed, so that's what I usually use since I already have the player for mp3's and CD's anyway. But i'm skating perilously close to Gnome territory using it and from what i've read the newest version is completely gtk3, which may bring it's own set of problems.
Is Deadbeef still around?
greenjeans wrote:but I might do a JWM version in 32, been playing with it and what comes out may not be the JWM you are used to.
Oooo...I hope it happens! I'd really like to see the Greenjeans' hackism on this!
Well it would be a lot easier if I had a 32 bit JWM Miyolinux to hack on.
lololol!
I read your short tutorial in another thread, purely brilliant, I would never have thought of it like that but installing lightdm first like that on a cli iso would indeed pretty much pull in everything you need for a working desktop other than the WM/DE itself...stroke of genius that, quick and dirty, I love it! That's probably how i'll start.
Have you checked out Trizen's other menu program? If you haven't, it's called simply "menutray", and that's what it does, dynamic menu that fetches .desktop files just like obmenu-generator, but this one is setup to always go in your tray instead of just a right-click, it acts more like a conventional menu in that respect, even supports tooltips (which is something i'd love to see in obmenu-generator).
So only downside is it always goes in tray, which is generally on right side of panel and usually includes clock, wicd icon, volume etc. Some folks might like it there though. You can switch the dock to the left side of the tray in jwmrc though, and put menutray there and you have artificially created a standard type setup similar to a Mate desktop, use pcmanfm or rox and have it run the desktop and the illusion is complete with desktop icons and all.
So that's kinda the concept i'm mulling over, making it 32 bit and ultra light and fast, but with all the convenience you'd expect from a more complex desktop environment and a very familiar set-up for folks switching older computers from windows xp to linux, but still with enough advanced useability and geekiness to make experienced users very happy too.
Hi Miyo,
Is it worth holding eg. apt-mark hold obmenu-generator package to stop the menu being stuffed up when upgrading???
Unfortunately that doesn't work. As Miyo mentioned the package isn't even in Vuu-do at least as far as Apt is concerned and the issue still occurs. And it's an outside package anyway, so there's no upgrade to it that goes on when you switch to ascii or ceres.
Trizen posted an update to the main obmenu-generator script like a week or two ago, it may possibly address the issue, haven't tested it that far yet, but in all other respects it works great so I guess we need to try it and see if ascii-upgrade still borks it.
UPDATE: Just checked and he posted yet another update to the script 3 days ago, I love seeing that, this is a great program and is being actively developed right now, here's the link to the raw script : https://github.com/trizen/obmenu-genera … -generator
3 out of 4 new isos up today, still need to upload the newest Openbox maximal, will do that tomorrow. Just minor fixes, if you already have Vuu-do installed from the last iso's I put up in early July then you probably don't need anything from these new ones, just some small fixes and conky changes and fixed a locale issue I neglected on the last run.
Then hopefully I can get started on a 32 bit version for older machines, probably gonna stick with Openbox on i386 and not do a Mate version, but I might do a JWM version in 32, been playing with it and what comes out may not be the JWM you are used to. ![]()
Guess what, in arch/manjaro/artix openrc where you can have the generator the menu does not work anywhere near the conky surface.
You shut conky off and it works everywhere.
Try that own-window-type setting in conkyrc, usually that's all I ever have to change if conky is behaving badly.
Good info to have, I use writer sometimes and wife uses it a lot, thanks for posting!
WOW! Thanks for checking into all that, I haven't messed with ascii myself as i'm sticking with the stable branch for project iso's for now, but it's awesome watching other people experiment with it! That's what it's made for after all. ![]()
As far as my configs for obmenu-generator go, I guess you guys read the notes, the .deb package from opensuse includes all the cpan stuff and dependencies, which obmenu-generator does not require to use (deleting things not used gives me endless joy, lol), but possibly it needed them on the upgrade?? Still curious as to exactly why that is, maybe a pathway was altered somewhere and a fresh install re-directs properly?
The conky deal is probably simple, I don't even know why that would change in ascii unless something in xorg has changed. Anyhoo, the minimum conky window size is set in conkyrc and is larger than what my current config requires to allow for expansion, so the area around the text is in fact a square window, you just can't see the borders because it's transparent.
There are also some settings regarding window type and how it interacts with various types of desktop, in mate it's different from openbox by necessity because the mate file manager is drawing the desktop, in openbox the background is on X. These are the settings that may need to be tweaked if something has changed in Openbox or xorg I think, these lines in .conkyrc are the ones:
background yes/no (fork conky to background, never even tried this setting)
minimum_size 300 (minimum size of window in pixels)
own_window ##yes/no (mostly always used "yes" for this setting and do so currently in all Vuu-do's)
own_window_type ## normal /override /desktop (Openbox uses "normal" setting usually. Mate requires the "override" setting or desktop shortcuts don't work properly.)
Conky is an amazing program, it will do an endless amount of things, but only if you ask it right, lol, i'm no conky expert so I try to keep it simple.
For me, i'm not going to even start trying to build an Ascii Vuu-do until the VUA's release a stable beta, but I encourage and applaud anyone else who does! ![]()
Okay...I've done some more investigating in regard to the post above regarding the menu in Vuu-do after an Ascii upgrade. In fact, I did a complete new install to start over from scratch.
I confirmed that it wasn't a mouse issue. Before I did the following, the menu would open once after logging in, but it was hit or miss as to whether it would open anywhere on the desktop thereafter.
I now have a fully-functional menu on Vuu-do in Ascii...with one caveat...
Here's what I did...
There still is a glitch with the menu having something to do with the area in and around the conky...
After downloading the .deb package...
1. I once again saved the config.pl and schema.pl.
2. I then deleted the obmenu-generator folder in ~/.config
3. I then deleted the obmenu-generator file in /usr/bin/ (fungus is correct; there is no "obmenu-generator package" in Vuu-do)
4. I then installed the obmenu-generator .deb package (it installed 10 dependencies; the screen blinked, so that was a good thing!)
5. I then placed the "old" config.pl and schema.pl files in the newly created obmenu-generator folder in ~/.config (overwriting the new schema.pl file).
6. I then ran the command...
obmenu-generator -p -iTHE CAVEAT...That gave me a working Vuu-do menu again...HOWEVER...if I click in any area in (or closely around) the conky, the menu doesn't show up. Imagine drawing a rectangle about around the conky and the rectangle extends about 3/4" past the conky...the menu won't open anywhere inside of that rectangle. I would guess that probably has to do with the conky's "window" configuration, but I didn't take time to look.
However, it opens anywhere else on the desktop or any empty space on the panel.
With all of that said, this is not an official Vuu-do Openbox fix.
This is what worked for me. Greenjeans may figure out something better. I was just having fun trying to see if I could fix it. LOLOLOL!
...but I'm running out of time to try and research it any longer.
ABSOLUTELY AWESOME DISTRO PROJECT!!!!
* ath9k is free software and therefore comes baked-in with Devuan's default kernel as well as with the more strict linux-libre kernel.
If i'm not mistaken, it is on later kernels like ascii's 4.9, but not in the stable 3.16 kernel that's currently used in Devuan 1.0, so if you're running jessie you still need the atheros wi-fi firmware package.
I got suspend to work great in both Mate and Openbox, in my case system always had the capability, it was just some programs asking it wrong and others refusing to try.
I installed Obsession, it's just a simple shutdown program for Openbox, but it works great in Mate too, it has a gui but you can also use the cli portion by invoking various obsession-exit commands, I tied them to keyboard shortcuts to make it super easy, so to suspend on my machine I just hit alt+s .
FYI the Mate power manager doesn't do much of anything really, the xfce4 power manager works in any DE and still includes suspend and hibernate functions which the Mate PM does not.
@Fungus: Thank you!!! It really is countless hours of work, if I was better at all this it would probably go quicker, I really need to document better too.
Miyolinux figured out the ascii menu bug, weird, still don't know what made it burp like that in the first place, but it's good to have the fix so thanks Miyo!
I seem to recall the last time I did the ascii upgrade test that a newer version of Perl was installed, maybe something about that caused it to stop working, and was reconfigured automatically when the package was re-installed?
Just rolled up 4 fresh iso's , fixed some small stuff and added to release notes, I was chainsawing and chopping wood all day, so it was a good day to let the machine squash files while I worked. Should be able to upload at some point this weekend, wife is having minor surgery today so probably won't be online long today.
I never learned to type, so it's two fingers-hunt-and-peck and slow going, but bit by bit I am trying to document the whole process and basically make the Vuu-do release notes kind of a how-to reference, noting locations of files, various shortcuts and hacks etc. to try and encourage users to start experimenting themselves, this is in keeping with the Devuan/Refracta/Miyolinux philosophy that I share, making it easy for folks to "make it your own"!
Crowz also has JWM (64bit only)
+1. Crowz is nice!
Been playing with JWM myself, I think a nice light 32 bit Devuan version would be great for older machines.
Wow, so all it took was re-installing the package? that is weird....I had noted it stopped working, figured maybe it was missing a dependency but hadn't had time to mess with it.
Thanks for the fix!
Did you try the new main executable? Don't think it's in that .deb package, it's still version .66, new obmenu-generator.pl is version .99 I believe.
greenjeans wrote:golinux wrote:Relax and enjoy the stability that 'just works'.
You know, it's funny, because probably nobody that ever posts on this forum will ever do that . . .
Well, I do! I like stable and boring and only try new things when absolutely necessary.
I'm posting from the first Devuan netinstall I did last year, this is my test-mule partition, the original 32 bit beta I first tried, dang thing has never crapped out on me, been through a gig of updates with it, have tested close to 300 programs with it (it has at least 100 on it now), probably several thousand packages overall, and chunks of leftover stuff I deleted. I have hacked it close to death many times, but still it never fails and it still runs pretty fast, I use it every day.
I REALLY like that kind of stability and toughness.
Many candles lit from one flame . . .
That is a beautiful and spot-on analogy ma'am. ![]()
Relax and enjoy the stability that 'just works'.
You know, it's funny, because probably nobody that ever posts on this forum will ever do that, lol, while on the other hand, every system I install here locally for people who aren't geeks and just want to use their computer, will do just that, and religiously.
Seriously, when I see them in public after months of use, I always ask if they have ever updated and they never have, nor have most even installed any more programs over and above what I installed for them originally.
I think sometimes linux enthusiasts forget that not everyone is into computers like they are, 99% of computer users worldwide just want something simple and stable to use daily, not a hotrod project with constant breakage.
1. Obmenu-generator (though all of its dependencies came from Devuan)
Hey before I forget, Trizen pushed an update to the main obmenu-generator perl script a couple days ago, says "minor tweaks" but there's a fair bit of extra code, it works but have not tested it thoroughly yet.