(old)Puppy Linux Discussion Forum

Posted: **Mon 08 Oct 2007, 22:18**

Whenever the pup_xxx.sfs file cannot be found, due to either a "/" in psubdir, a nonexistant psubdir or no such file at the root level, the console output is

Searching for Puppy files in computer disk drives...
/bin/sh: can't access tty; job control turned off

But the init logic (line 365+) below implies that a more appropriate message was intended, but went arwy:

Code: Select all

if [ "$PUPSFS" = "" ];then
 echo -en "\\033[1;31m" >/dev/console #31=red
 echo -n "pup_${PUPPYVERSION}.sfs not found. Dropping out to initial-ramdisk console..."
 /bin/echo -e "\\033[0;39m" >/dev/console
 exec /bin/sh >/dev/console 2>&1
else

However, the line

pup_300.sfs not found. Dropping out to initial-ramdisk console...#

appears in the bootinit.log file. Furthermore, the "exec /bin/sh" line appears to be generating the "/bin/sh:..." error message, instead of restoring a color (I assume).

Correcting that logic to display the "not found" line would save people much frustration when they cannot get Puppy to boot (probably due to a typo in a menu.lst file).
Richard

Posted: **Mon 08 Oct 2007, 22:49**

The init script saves a copy of its puppy file locations held in /tmp for CD boots and installs with pup files in a psubdir, but creates no such files when puppy files are at the root level. So, when an empty /tmp directory is copied to /initrd/tmp it leaves residue from a previous boot, such as from CD (or the copy may not even be attempted).

To avoid misinformation to debuggers, /initrd/tmp should be left empty if the init /tmp directory is empty. Seeing the PUPPYFILES as being in hdc after a HD bootup is confusing..
Richard

Posted: **Tue 09 Oct 2007, 08:14**

rerwin wrote:Whenever the pup_xxx.sfs file cannot be found, due to either a "/" in psubdir, a nonexistant psubdir or no such file at the root level, the console output is
Searching for Puppy files in computer disk drives...
/bin/sh: can't access tty; job control turned off
But the init logic (line 365+) below implies that a more appropriate message was intended, but went arwy:
Code: Select all
if [ "$PUPSFS" = "" ];then
 echo -en "\\033[1;31m" >/dev/console #31=red
 echo -n "pup_${PUPPYVERSION}.sfs not found. Dropping out to initial-ramdisk console..."
 /bin/echo -e "\\033[0;39m" >/dev/console
 exec /bin/sh >/dev/console 2>&1
else
However, the line
pup_300.sfs not found. Dropping out to initial-ramdisk console...#
appears in the bootinit.log file. Furthermore, the "exec /bin/sh" line appears to be generating the "/bin/sh:..." error message, instead of restoring a color (I assume).

Correcting that logic to display the "not found" line would save people much frustration when they cannot get Puppy to boot (probably due to a typo in a menu.lst file).
Richard

Thanks for that. The missing line is now not missing.

Posted: **Tue 09 Oct 2007, 13:35**

Found a problem with my broadcom wireless card.

I opened a new thread:

http://www.murga-linux.com/puppy/viewtopic.php?t=22459

Posted: **Tue 09 Oct 2007, 20:22**

BarryK wrote:
Jesse wrote:Hello,

I have not been able to get my usb speakers to be retained as the working sound device on reboot. Puppy 3.0.
I've been looking into it, and it is not so straight forward, as on the surface and looking at how things interact, it should have been retained.
When I set the usb speakers as the sound device with the wizard they work great. But they are not a pci sound card, so there is no 'alias snd-card-X' line in /etc/modprobe.conf, only an 'alias sound-slot-0 snd-usb-audio', and the snd-cmipci is omitted from /etc/modprobe.conf.
Then when I reboot, the motherboards snd-cmipci driver gets loaded at 'alias snd-card-0' position in /etc/modprobe.conf, This is done by some automatic hardware detect script? (I have not found this one yet).
Motherboard BIOS does not let me disable the onboard sound card.
When the snd-cmipci gets loaded, the usb audio does not work.
When I blacklist the snd-cmipci it is still loaded.
When I manage to get everything into a non conflicting configuration, and reboot, and settings are retained, snd-usb-audio is still reported as 'default sound device' from "cat /proc/asound/cards", the usb-audio still does not work.

Can someone think of what else i could try?

Jesse
Jesse, maybe these code areas would be helpful to examine:
/etc/rc.d/rc.local0, from line 368
/etc/rc.d/rc.modules, from line 76
/etc/rc.d/rc.modules2 <<<particularly this one.

Note, you can also blacklist any module.

Hi,

I managed to get the computer to boot with USB as the active sound device.
I modified the rc.modules and rc.modules2 so that they did not modify the alsa setup, there are comments in the alsa section of /etc/modprobe.conf file which says 'do not modify', lol.
Ok, so I think what might be happening is "hardware detection" by the rc.modules* scripts, which is what may get the sound card sorted on first boot. But, when I've already set my alsa setup, the scripts shouldn't be forcing me to have the module loaded.
Perhaps there needs to be some setting, with two user options: "automatic sound" and "no automatic, just selected by alsa wizard".

There is a bug in the rc.modules* scripts processing, as they can generate multipe configurations for the same X (card/slot) as they only look for the 'snd-card-' lines. This will undoubtably confuse alsa when the user has USB sound and another sound card which they are not using.

Even when there is no conflicting/double-up settings in /etc/modprobe.conf, and with all sound dirvers specified, the snd-usb-audio driver does not work for me. I have no idea why.
It only works when snd-usb-audio is the only sound card.

Note that blacklisting does not fix this, as the rc.modules* scripts still detect it and enter it with the usual descriptions "alias sound-slot-X ..." lines in /etc/modprobe.conf. i.e. two drivers specified for card 0.
Earlier I claimed that blacklisting snd-cmipci did not prevent it loading, I didn't check that was correct. I think I confused detecting with loading. The snd-cmipci is still detected when it is blacklisted, but presumably cannot be loaded.

Jesse

Posted: **Tue 09 Oct 2007, 20:22**

Whats going on with gxine ?

Run it, select Media -> TV
CSPAN1 : unable to resolve address
CSPAN2 : The xine engine failed to start. No input plugin was found. Maybe the file does not exist or cannot be accessed, or there is an error in the URL.
BBC NEWS24 : sound but no video
etc.

Posted: **Tue 09 Oct 2007, 23:58**

Was listening to music and Gxine froze. Wouldn't close. Had to kill it. Now it won't start back up. I'm going back to 2.17.1 until 3.0 is more stable. Too bad I wiped 2.17.1. It worked really well.

On a positive note. 3.0 does seem to boot faster than the 2.* series.

Posted: **Wed 10 Oct 2007, 02:02**

The problem with gxine freezing is a gxine problem (unless it's a kernel problem, or a driver problem, or...

). However the problem with not really being able to kill a hung gxine appears to be a jwm problem. I experience the same symptom when Pupscan hung or crashed, which it will do for me after pressing the "USB interfaces" button.

I don't know enough about jwm to chase it down though.

Posted: **Wed 10 Oct 2007, 03:24**

The following rather common website works for me using SeaMonkey under Puppy version 2.17 but not for version 3.00:

http://movies.yahoo.com/

There is more to the story than that, but I'd be interested in seeing if this is just my problem, or if there is really a bug.

Posted: **Wed 10 Oct 2007, 05:39**

sketchman wrote:Was listening to music and Gxine froze. Wouldn't close. Had to kill it. Now it won't start back up. I'm going back to 2.17.1 until 3.0 is more stable. Too bad I wiped 2.17.1. It worked really well.

On a positive note. 3.0 does seem to boot faster than the 2.* series.

Turn gxine visualization to none, then it wont freeze.

Posted: **Wed 10 Oct 2007, 06:20**

Just keep in mind the folly of treating symptoms.

Posted: **Thu 11 Oct 2007, 12:45**

Puppy 3 seems fast!!!! (but that perhaps because Im used to using the bigger 2.15ce).

I understand that seamonkey suite has been upgraded in puppy 3? Now each application can be started separately and it will use the core code in memory. This certaainly seems to be the case with the email and browser functions. However, when I click on html on the desktop (which seems to have re-appeared, better just on the menus I think) then I still get the message "please start composer from the window of the currently running mozilla".

I like the new menu setup, it is a lot clearer, and just one entry to install new programs. I noticed the cd/dvd setup wizard guesses which drives are which in previous puppies, perhaps on boot into ram this short script could be carried out automatically without the users needing to run the wizard (puppy just works!). Glad to see Gemgame is back in! And turma and gtkfind to go with pfind.

the shutdown menu particularly is a lot clearer for new users, just one small fault, the "change window manager" entry currently does nothing. It Should bring up a message to say "You currently only have only one window manager, jwm, installed". Ideally it could take you to a screen where you could download additional window managers (perhaps that can be in the CE edition).

PupcTorrent on the internet menu is missing an icon associated with it.

The new NETSURF Browser is supurb and a great improvement on Dillo. I do notice that both browsers do not have any puppy bookmarks preloaded any more. In Ecopup, Ive recieved good feedback that many of the most popular sites are already bookmarked (not just puppy ones). Since puppy is for "linux newbs" perhaps some bookmarks to general linux pages (open source philosophy, ways of doing things, advantages etc) would be appropriate.

Although I like the idea of a separate wallpaper for puppy running in RAM to running from the desktop, the wallpaper with the instructions is white and is not attractive for first time users to get a good first impression. A white desktop background is kinda harsh on the eyes. I think a pop up box was a good idea on previous puppies (although youve probably taken it out as your not using dillo any more). If you set the screen resolution to anything other than 1024x768 then the pupzip and trash icons obscure the writing (perhaps they can be included in the icons in the top left corner?).

Other than that, I think a whole lot more users would be won over initially to explore puppy further if a different wallpaper and a colour was used for the taskbar. Im afraid users still think that grey means old (win 98/95). Puppy 1.05 had the best out of the box look.

Sorry for mixing suggestions in with the bug reports. The 3 series in a major improvement with its additional features. I look forward to working with it.

Posted: **Thu 11 Oct 2007, 22:38**

I was running IceWM on Puppy 3.00 and I decided to switch to JWM; it worked. Then I tried to switch back to IceWM with what seemed to be the appropriate menu item in JWM; nothing happened. I tried the "Restart X Server" item in JWM and I got a black screen that just stayed there, with no evidence that the X server was being restarted. I had to reboot Puppy with Ctrl+Alt+Del to get back to IceWM! Can anything be done about this?

David McClamrock

Posted: **Fri 12 Oct 2007, 00:38**

hi if you take a fresh copy of puppy 3.0 and Dougal remaster it, and boot up with xvesa on, the screen size is over large and withe xvesa wizard does not work at all, you can't change the screen size, This has been really frustrating for me, i tried everything to fix it and then big bass mentioned just try to remaster it without changing anything to the puppy 3.0, so all i did was did the network wizard, logged into chat, and then i logged out of chat, and remastered it and the problem was there, please if you could find a patch, that would be great, and if you please message me where the link is, thanks ever so much for your time, ttuuxxx

Posted: **Fri 12 Oct 2007, 00:44**

sketchman wrote:Was listening to music and Gxine froze. Wouldn't close. Had to kill it. Now it won't start back up. I'm going back to 2.17.1 until 3.0 is more stable. Too bad I wiped 2.17.1. It worked really well.

On a positive note. 3.0 does seem to boot faster than the 2.* series.

Sketchman you might want to try Fire Hydrant, its larger 192Mb, but its based on puppy 3.0 but i removed Gxine and replaced it with Xine, also installed XMMS with a nice sony theme, so more then likely you won't have that problem anymore, and yes its also really fast, i made it to replace my 2.15ce, you can get it here at www.ttuuxxx.com

In the future i'll have other releases, plus it has the PPoE fix already embedded into it

Posted: **Fri 12 Oct 2007, 09:32**

I have had three different problems with USB flash drive installations, but I'm not sure whether they are different problems or different symptoms of a single problem.

1) Puppy 3.0 final installed on a 1Gb flash drive with a 256Mb heavy encryption save file. Couldn't install EZpup-3.0 - too little space to expand the archive.

I tried to expand the save file by an additional 512Mb, but then Puppy wouldn't boot.

Re-installed into second setup.

2) Puppy 3.0 final installed on 1Gb flash drive as an HD install using the whole drive.

Same problem - insufficient space to expand EZpup-3.0 .PET archive.

Installed into third setup: a 2Gb flash drive with a 512 Mb save-file with heavy encryption.

Same problem. Tried to expand the save file by an additional 512Mb and re-boot.

3) Puppy 3.0 final on 2Gb flash drive with heavy encryption save file and a queued request to expand the save-file at next re-boot.

Re-booted OK but ignored the expand request. Still won't install EZpup-3.0.

SOLVED

=====

I don't actually know WHY it worked when previous methods didn't, but installation via MU's PSI application (aka "Puppy Software Installer") worked perfectly.

Also; it appears that not expanding encrypted save files is a feature not a bug! I read somewhere that Barry plans to change that, so it may appear in 3.02 or 3.1.

Posted: **Fri 12 Oct 2007, 10:12**

Just a quickie : gxine wont play ogg files - demuxer not found.

replaced the dmx file with one from 2.17 & it works -obviously the version in Puppy 3.0 is f****d.

Posted: **Fri 12 Oct 2007, 18:34**

Regarding my /bin/sh:can't access tty;job control turned off problems, I fixed it specifying the partition grub had to work on in ALL lines in menu.lst

My entry now looks like this:

title Puppy
rootverify (hd0,1)
kernel (hd0,1)/vmlinuz root=/dev/ram0 PMEDIA=satahd
initrd (hd0,1)/initrd.gz

I am now typing this from Puppy 3...

HTH
RickyVaio

Posted: **Sat 13 Oct 2007, 08:15**

When running Puppy 3.00 as a LiveCD I continue to find these problems, even after the PPPOE-fix. (My internet connects through a DSL modem, no router)
Roaring Penguin starts and connects, but Firefox doesn't get access to the internet, neither does Thunderbird.
Same problem occurs in Firehydrant and TMXprism3BETA.
TMXprism3 doesn't even start Thunderbird, neither from the menu, nor a changed desktop icon, nor the console. Also why does TMX use the old mut again? The new one in 2.17 was much more pleasant to look at.

None of this occured in any version of 2.17 I have tested, not Puppy, not TMX, NOP, etc.
Could you please address these issues? Guess this is the 3rd time I have posted this.
Otherwise I am very pleased with Puppy, the 2.17 versions that is. You are doing an excellent job.

Posted: **Sat 13 Oct 2007, 21:39**

seems to have been a problem with this particular usb pendrive (although it worked with previous versions). Using a USB pendrive with only 1 partition solved the problem.

barker wrote:Hi I've used puppy's universal installer several times before, and never had problems- but with 3.0 it does not seem to work.
however my pendrive is formatted (fat16, ext2, ext3, fat32, one partition, two partitions, bootable flag, no flag - formatted in windows, formatted with linux, partitioned with gparted, or fdisk, or cfdisk (you see I've tried a few options Wink) it is only recognised as 'superfloppy formatted with ext3 - no files are copied across (although the operation takes some time and uses both the CD-drive and hard disk (swap?)). No options to install a boot loader or make the disk bootable appear afterwards.

Has anyone else had this problem?
Installing on ATA hard-drive no problem.

barker

(old)Puppy Linux Discussion Forum

Puppy 3.00 bug reports here

Misleading init message for missing pup_xxx.sfs

init script housekeeping nit

Re: Misleading init message for missing pup_xxx.sfs

A few problems but overall pretty good!

JWM in 3.00: Change WM and Restart X don't work right

XVESA Broken

Problem(s) with USB Flash Drives

gxine

FIX!

Puppy 3.00 bugs

Re: universal installer failure on usb flash [solved]