Quirky 1.3 feedback and bug reports

[BarryK]

The burner, which is also used to boot the live CD is an external USB device connected to the main PC via a USB hub.

The extra message isn't really needed. I know from my own experience with an external USB CD drive that you would have been getting the 5-sec pause before, but without the "pausing" message.

[BarryK]

when i unsquashed the puppy.sfs file, then resquashed, all from another running system, the file puppy.sfs is then no longer seen by the initrd during bootup. It drops out right away. Is there a trick to this for the new signature system?

After recreating puppy.sfs, do this:

echo -n 'w081100929084800' >> puppy.sfs

...or, whatever the id-string is in Quirky 1.3 (I'm running Wary right now)

The id-string is found in /etc/DISTRO_SPECS:

DISTRO_IDSTRING='w081100929084800'

This is only needed if the simplified filenames are chosen when building Puppy. For the next Wary I will use the traditional naming, so it will be 'wary_090.sfs' and will not need the id-string.

The 'init' script looks for the correct id-string on the end of puppy.sfs because there is no identification in its filename.


EDIT: I see some of the guys have already answered the question!

[BarryK]

double post

[BarryK]

Noticed In Quirky/Wary that pmount has an incorrect translation variable at lines 72 & 75-> $m_34 it should refer to m_31="KILL"
line 72-> m_35="If you press the "$m_34" button, Puppy will attempt to
kill the offending programs for you. Only do this as a
last resort. "
line 75-> m_36="Firstly you should try to close the programs
manually, so do not press "$m_34"."

pmount still contains 2 buttons that are not taken into account, lines 419-420, it's an easy fix
the OK variable is already present (m_38), just missing the CANCEL for it to be complete.
There are spare variable numbers 40-49, which i think pertain to the relevant part of the script,
so I added a new translation variable : m_40="CANCEL" #your choice of nber here
then changed lines 419-420 in pmount
<button ok></button> ... <button><input file stock="gtk-ok"></input><label>$m_38</label></button>
<button cancel></button> ... <button><input file stock="gtk-cancel"></input><label>$m_40</label></button>

Thanks for that, I have fixed pmount.

[Proudog]

I found this little bug when a try /usr/sbin/filemnt. When yaf-splash is invoked with the parameter option -font "xxx" it's reported a "false" output, and not is screening the message.

Edited
Line error be changed in gtkdialog-splash:

Code: Select all

#100917 if old yaf-splash installed, and '-font' param, execute old yaf-splash.

if [ -f /usr/X11R7/bin/yaf-splash ];then #100917
-  [ "`echo "$*" | grep ' \-font '`" != "" ] && exec /usr/X11R7/bin/yaf-splash $@ #BK
+ [ "`echo "$*" | grep '\-font'`" != "" ] && exec /usr/X11R7/bin/yaf-splash $@ #Bug-fix
fi

It's a simple sintaxis on spaces error.

Note: I like gtkdialog yaf-splash, is a UTF-8 complaint.

[zygo]

Using grep -E -o. GNU's grep 2.5.4 (in Q130) will not match 'literal.* literal' near the end of a single line bigger than about 150kB. GNU's grep 2.5.3 (copied from P431 & tested in Q130) and the one in BusyBox v1.16.2 (in Q130) will match. However, a more efficient way to search is with {1,99} in place of * which works in all 3 versions of grep.

[linuxcbon]

Why still using oss since alsa replaced it ?

[zygo]

I know that /dev/sda is an unknown device but why the 19 seconds delay?

Code: Select all

# date && mount /dev/sda /mnt/sda || date
Mon Oct 18 11:03:15 GMT 2010
mount: /dev/sda: unknown device
Mon Oct 18 11:03:34 GMT 2010

This delay does not always happen in Quirky 130 but I never noticed it in Puppy 431 or earlier Puppies.

[Béèm]

This is only needed if the simplified filenames are chosen when building Puppy. For the next Wary I will use the traditional naming, so it will be 'wary_090.sfs' and will not need the id-string.

All warry's before where wary-0X0.sfs. Why the underscore now?

[PANZERKOPF]

After recreating puppy.sfs, do this:
echo -n 'w081100929084800' >> puppy.sfs

As I found, Quirky's kernel image has same "not native tail" (16 bytes long).
P.S.
I have one question about Quirky kernels.
There are three pakages in Quirky's repository:
linux_kernel-2.6.33.2-tickless-smp-q1.pet
linux_kernel-2.6.33.2-tickless-smp-1-q1.pet
linux_kernel-2.6.33.2-tickless-smp-2-q1.pet
What is the difference between?
Also, I found "modprobe fbcon" in rc.sysinit. May be better way is to compile
kernel with enabled framebuffer? With this kernel we can boot directly in high resolution mode using vga=??? boot option.

[zygo]

I can't remaster Q130. I applied these 2 bug fixes:

Bug 1: the remaster script terminates with "your Puppy is too old"
my (apparent) fix is to edit /usr/sbin/remasterpup2 line 236
change $DISTRO_SPECS to $DISTRO_IDSTRING

Bug 2: DaveS (after BK) describes the need to append the $DISTRO_IDSTRING value to the puppy.sfs http://www.murga-linux.com/puppy/viewtopic.php?t=60720

I don't change the new root or .etc when offered but I do allow console customisations and edit:
- isolinux.cfg to add a pfix=ram boot option
- boot.msg to add indicate it's remastered and
- puppy.sfs as in Bug 2.

Attempt 1
At the badly drafted "make zdrv.sfs inside?" dialogue I click "no" (thinking it starts outside so why change it). Before the dialogue offering to change /tmp/root there's an orange window which starts with an error and then a progress bar which completes. I let the remaster script make the cd-rw when it offers. The new cd boots but stops saying it can't find puppy.sfs. Game over.

Attempt 2
At the "make zdrv.sfs inside?" dialogue I click "yes". The error in the orange window is something like "cannot stat exclude puppy.sfs ignoring... parallel mksquashfs". I let the remaster script make the iso when it offers and use the iso burning utility to make the new cd. This cd boots much further but will not start X. Very hard to see what's wrong but maybe the intel graphics driver is missing. This runs so smoothly at first boot but not a dodo now. Where is this in the original iso? It may be just a matter of adding it to the sfs before burning.

Edit: added "exclude" to "cannot stat" error.

[zygo]

Attempt 3

As attempt 1 but when it offers you chance to alter /tmp/root and /tmp/etc
- delete everything within
- copy everything from the live version
- delete .XLOADED

Success.

Note I got these extra 3 steps from the forum and used them on Puppy 431 maybe earlier.

Barry, you do ask in one of the dialogues to tell you if something is not copied that should be. The answer seems to be to copy all but the console customization files and .XLOADED

[fluxit]

I was under the impression that "puppy pfix=ram" should leave no non-swap partitions mounted when the desktop appears. Twice in a row today, I have booted with pfix=ram only to find that "sda1," my ntfs partition is mounted as /initrd/mnt/dev_ro2 and cannot be unmounted as it is "in use."
Puppy was not using my pupsave as far as I could tell.

I normally run from the CD with no boot options, and my pupsave on sda1.

Another issue that has come up(it was not doing this initially) is that JWM will no longer "Exit to prompt," "Reboot computer," or "Power-off computer" without Ctrl-Alt-Backspace to help it along.

I have changed many of JWM's configuration options and have fallen afoul of the "JWM configuration" empty file bug(which then went away on its own,) but it seems to be working properly otherwise currently.

One of the options I changed was to sliding-focus, which I didn't really want, but it keeps me from having to double click on the taskbar to get it to do things that it should do with one click like switching virtual-desktops or minimize/maximize. It took me a while to figure out that I was double-clicking and not merely falling victim to a failing mouse button(my mouse is fine.)

[tlchost]

Attempt 3

Would it be possible to post the fixed remaster script?

Thanks

[PANZERKOPF]

Fragment of code (alternate_driver_func)

Code: Select all

*) #100430 get rid of all drivers except chosen one.
   REPLACEDDRV=""
   mv -f /usr/lib/xorg/modules/drivers/* /usr/lib/xorg/modules/drivers-backup/
   [ -h /usr/lib/xorg/modules/drivers-backup/i810_drv.so ] && mv -f /usr/lib/xorg/modules/drivers-backup/i810_drv.so /var/local/xorg/ #100902
   [ -f /usr/lib/xorg/modules/drivers-backup/$SUBSTDRV ] && cp -f /usr/lib/xorg/modules/drivers-backup/$SUBSTDRV /usr/lib/xorg/modules/drivers/
   [ -f /usr/lib/xorg/modules/drivers-alternate/$SUBSTDRV ] && cp -f /usr/lib/xorg/modules/drivers-alternate/$SUBSTDRV /usr/lib/xorg/modules/drivers/

I think that must be:

Code: Select all

*) #100430 get rid of all drivers except chosen one.
   REPLACEDDRV=""
   mv -f /usr/lib/xorg/modules/drivers/* /usr/lib/xorg/modules/drivers-backup/
   [ -h /usr/lib/xorg/modules/drivers-backup/i810_drv.so ] && mv -f /usr/lib/xorg/modules/drivers-backup/i810_drv.so /var/local/xorg/ #100902
   [ -f /usr/lib/xorg/modules/drivers-backup/${SUBSTDRV}_drv.so ] && cp -f /usr/lib/xorg/modules/drivers-backup/${SUBSTDRV}_drv.so /usr/lib/xorg/modules/drivers/
   [ -f /usr/lib/xorg/modules/drivers-alternate/${SUBSTDRV}_drv.so ] && cp -f /usr/lib/xorg/modules/drivers-alternate/${SUBSTDRV}_drv.so /usr/lib/xorg/modules/drivers/

A value of this variable is the name of driver but we need to copy a file of this driver.