Version 14.2 has been released

Global announcements
Scooby
Site Admin
Posts: 826
Joined: 09 Sep 2013, 16:52

Re: Version 14.2 has been released

Postby Scooby » 28 Dec 2013, 14:05

ncmprhnsbl wrote:pidof is covered by busybox,
dont know if this is the correct way but a quick fix might be to put a symlink in /usr/bin pointing to busybox


or I could install sysvinit-tools
pacman -S sysvinit-tools

end get pidof

My question was more along the line "shouldnt it be present on alphaos from the start"

As I said it seems to work anyway but who knows?

*edit*
tried to install sysvinit-tools but some part was already there to conflict so I tried your suggestion.

Put

Code: Select all

#!/bin/bash
busybox pidof "$@"


in /usr/bin/pidof
and chmod it

chmod 755 /usr/bin/pidof

seems like frisbee is running a little smoother now

Scooby
Site Admin
Posts: 826
Joined: 09 Sep 2013, 16:52

Re: Version 14.2 has been released

Postby Scooby » 28 Dec 2013, 18:34

Sim, as you know i run a slightly modded version to
be able to boot directly from iso

When I load modules during startup in /modules dir
and then deactivate them They get removed from union and
unmounted but still occupys loop device.

I can free it with losetup -d /dev/loop3

When I load and unload with bundle loop device is freed
automatically.

Is this because in init loop devices get setup using losetup
and not by mount?

*edit*

When I check livekitlib losetup is not used?



Do you have the same behaviour?

Maybe bundle have to check if loop device is really freed?

Scooby
Site Admin
Posts: 826
Joined: 09 Sep 2013, 16:52

Re: Version 14.2 has been released

Postby Scooby » 28 Dec 2013, 18:55

simargl wrote:GNOME Goal: Header Bars

Since 3.10, many applications have switched to use header bars rather than the traditional combination of titlebar, toolbar and menu bar. This has the advantage of reducing the amount of screen space used by window chrome, keeps the window close button available, and allows dynamic controls that change according to the view in the window.



from http://alphaos.tuxfamily.org/forum/viewtopic.php?f=9&t=810&p=1643&hilit=headerbars#p1643

I dont think this goal is reached in alphaos!

Look at pic below! looks like headerbars use even more screen space!

Image

simargl
Site Admin
Posts: 466
Joined: 16 May 2013, 10:54
Contact:

Re: Version 14.2 has been released

Postby simargl » 28 Dec 2013, 19:05

Scooby wrote:Do you have the same behaviour?

Maybe bundle have to check if loop device is really freed?


I've checked, and that's right it is same here

loop device is freed (after bundle -d) when module is loaded (activated) with bundle -a, but not if it's one of modules loaded on boot from init and livekitlib. I would never think of checking that. Is this really important? Increase RAM use or something.

Scooby wrote:looks like headerbars use even more screen space!

yes but they can hold buttons inside, and drop down menus in that case they really save some space,
And gtk3 theme define how will program look in any window manager. I like new Gtk widgets, beside header bar especially stack switcher and search bar all useful in Emendo.

Scooby
Site Admin
Posts: 826
Joined: 09 Sep 2013, 16:52

Re: Version 14.2 has been released

Postby Scooby » 29 Dec 2013, 14:24

Solution

use in bundle-->bundle_deactivate()


Code: Select all

umount -d /mnt/live/memory/bundles/"$BUNDLE" 2>/dev/null


-d, --detach-loop
In case the unmounted device was a loop device, also free this
loop device.



Also sim you should really add pidof command, frisbee is a 100% realiable with it

simargl wrote: Is this really important? Increase RAM use or something.


I'm not sure?
it is a resource-leak and I think it is considered
good coding standard to free loop device as well

It wouldnt be to hard to check from bundle though

Something like


Code: Select all

bundle -d grub.sb
...
LOOPDEVICE=`losetup | grep -w grub.sb | cut -d ' ' -f1`
if [ $LOOPDEVICE ]; then
   losetup -d $LOOPDEVICE
fi


But maybe this really is a infrequent problem?
Maybe users just removes from modules and restarts?

But what is the cause, is it busybox mount?
*edit*
Tested with busybox umount in bundle.
This frees loop device as well

simargl
Site Admin
Posts: 466
Joined: 16 May 2013, 10:54
Contact:

Re: Version 14.2 has been released

Postby simargl » 29 Dec 2013, 18:32

Done. Indeed umount -d fixes that minor problem, removal of loop device

I replaced busybox with toybox http://www.landley.net/toybox/about.html in core bundle (not in initramfs)

commands used from toybox
1) init
2) fstype
3) pidof
4) reboot
5) poweroff

that's it. 1-3 will be symlinks pointing to toybox, 4 and 5 will be used only from rc.shutdown (no symlinks)

toybox size is 140KB, busybox is 1.7MB

Toybox binary is in package alphaos-tools, compiled with default config + init enabled

Scooby
Site Admin
Posts: 826
Joined: 09 Sep 2013, 16:52

Re: Version 14.2 has been released

Postby Scooby » 29 Dec 2013, 21:16

simargl wrote:2) fstype


Is this the same as guess_fstype?

if so Will you link to fstype or guess_fstype


Peace on the Taeni + headerbars issue.
I've installed lxterminal and am content with it.

simargl
Site Admin
Posts: 466
Joined: 16 May 2013, 10:54
Contact:

Re: Version 14.2 has been released

Postby simargl » 30 Dec 2013, 17:48

guess_fstype is removed, fstype from toybox is replacement. I think that command is used only from spacefm to send notification about filesystem type of mounted partition.

Scooby
Site Admin
Posts: 826
Joined: 09 Sep 2013, 16:52

Re: Version 14.2 has been released

Postby Scooby » 30 Dec 2013, 19:46

simargl wrote: I think that command is used only from spacefm to send notification about filesystem type of mounted partition.


There and in my personal scripts.

I guess they'll turn up so I can fix them


Return to “Announcements”

Who is online

Users browsing this forum: No registered users and 2 guests

cron