Wary Puppy 5.1.1 Final

[BarryK]

Dear Barry,

do you want me to try with less RAM(768MB, 512MB or 256MB+SWAP-partition)?

Cheers, FeodorF


Oh, I forgot one thing to mention (in common) about CD size.

http://en.wikipedia.org/wiki/CD-ROM

I have two very old spare drives; only being able to address 650MB(74min).

If you would have a drive like this installed with your SiS board, that could create md5sum errors to. (But only if the CD is filled up to the max with data.)

Thanks for doing that 'md5sum' test. That just eliminates that possibility. My PC with SiS chips that returns incorrect md5sums must have a hardware fault I think.

[BarryK]

What is the official method for LIveDVD Warry users to employ the DEVX SFS?

If there is no "official" method, what is a recommended method for employing any SFS on LIveDVD?

Thanks in advance.

I presume that you mean "multisession DVD"?
You posted this question at my blog and I have replied there.

As far as live-CD or live-DVD, they work the same. Loading an SFS file is well documented.

[BarryK]

A few of you have probably remember my saga with disappearing sound on my Toughbook CF-T2. Sound would always be detected on a new boot (no warysave) but disappear 5-10 reboots later. This never happens in Puppy 4.3.1 or wary 030, but happens in all warys starting with 051.

Something in Shinobar's latest post inspired me. To make a long story short, I began comparing warysave files that have sound with ones that don't and started selectively morphing one into the other till I found the difference that broke things.

The final solution was to delete the 'snd' directory in /dev at shutdown. For some strange reason, the sound card is detected differently some times, making the saved device nodes wrong.

I'm still poking around the alsa changes to try to find out what actually broke things between 030 and 051, but for now, wary is usable and I can finally move on from 4.3.1 and run bleeding-edge software one this circa-2001 machine.

zekebaby,
Thanks for figuring that out. I have posted about this at my blog:

http://bkhome.org/blog/?viewDetailed=02168

[tronkel]

Barry wrote:

I am so puzzled about this. I don't know how you can get a malformed last line.

/etc/rc.d/rc.sysinit generates /etc/modprobe.d/blacklist.conf at every bootup:

Code:
rm -f /etc/modprobe.d/blacklist*.conf
BLACKLISTVARS="`echo "$SKIPLIST" | tr '\-' '_' | tr ' ' '\n' | sed -e 's/^/blacklist /' | grep -v ' $'`"
echo "$BLACKLISTVARS" > /etc/modprobe.d/blacklist.conf


That grep -v ' $' should have removed that malformed line.

Well, the thing to check next is the SKIPLIST variable, you will find it in /etc/rc.d/MODULESCONFIG
-- would you mind posting the content of that variable?

Here is the current contents of the SKIPLIST variable as requested:

SKIPLIST="SKIPLIST=' nouveau arcfb cirrusfb cyber2000fb cyblafb hgafb neofb pm2fb s1d13xxxfb sstfb tdfxfb tridentfb vga16fb aty128fb atyfb radeonfb i810fb intelfb kyrofb nvidiafb rivafb savagefb sisfb btaudio es1371 i810_audio kahlua trident ub usbkbd usbmouse via82cxxx_audio rt2800usb rt2x00usb '"

This is the value shown with the pristine pup_save (as it originally was.). This seems to correctly show the 2 Ralink rtXXXX modules that I want to blacklist (at the end of the SKIPLIST line. Logical enough.

Seems OK to me now, but why would this get malformed in the first place - i.e. before the pup_save file was replaced with a fresh one? - Interesting! Hope this helps to lead to the solution. Let me know if you need anything else.
_________________

[BarryK]

Barry wrote:

I am so puzzled about this. I don't know how you can get a malformed last line.

/etc/rc.d/rc.sysinit generates /etc/modprobe.d/blacklist.conf at every bootup:

Code:
rm -f /etc/modprobe.d/blacklist*.conf
BLACKLISTVARS="`echo "$SKIPLIST" | tr '\-' '_' | tr ' ' '\n' | sed -e 's/^/blacklist /' | grep -v ' $'`"
echo "$BLACKLISTVARS" > /etc/modprobe.d/blacklist.conf


That grep -v ' $' should have removed that malformed line.

Well, the thing to check next is the SKIPLIST variable, you will find it in /etc/rc.d/MODULESCONFIG
-- would you mind posting the content of that variable?

Here is the current contents of the SKIPLIST variable as requested:

SKIPLIST="SKIPLIST=' nouveau arcfb cirrusfb cyber2000fb cyblafb hgafb neofb pm2fb s1d13xxxfb sstfb tdfxfb tridentfb vga16fb aty128fb atyfb radeonfb i810fb intelfb kyrofb nvidiafb rivafb savagefb sisfb btaudio es1371 i810_audio kahlua trident ub usbkbd usbmouse via82cxxx_audio rt2800usb rt2x00usb '"

This is the value shown with the pristine pup_save (as it originally was.). This seems to correctly show the 2 Ralink rtXXXX modules that I want to blacklist (at the end of the SKIPLIST line. Logical enough.

Seems OK to me now, but why would this get malformed in the first place - i.e. before the pup_save file was replaced with a fresh one? - Interesting! Hope this helps to lead to the solution. Let me know if you need anything else.
_________________

Have you posted this exactly as it is in the file? It should be:

SKIPLIST=' nouveau arcfb cirrusfb cyber2000fb cyblafb hgafb neofb pm2fb s1d13xxxfb sstfb tdfxfb tridentfb vga16fb aty128fb atyfb radeonfb i810fb intelfb kyrofb nvidiafb rivafb savagefb sisfb btaudio es1371 i810_audio kahlua trident ub usbkbd usbmouse via82cxxx_audio rt2800usb rt2x00usb '

What you have posted is malformed.

[BarryK]

Barry,
To remove the residue from my alternative implementation of mode switching, please remove the following files from woof:
- /etc/udev/rules.d/40-usb_modeswitch.rules (an unneeded link)
- /lib/udev/usb_modeswitch.sh (obsolete)
- /usr/sbin/pusbmodeswitch
- /usr/sbin/pusbmodeswitch_adapt

In my testing of wary 5.1.1, I see that it does not check my USB flash drives for being ready. Puppy does not display the "usb-storage" text after "usb", in the "waiting for modules" line during bootup. When I substitute for:

Code: Select all

if [ "`cat /tmp/pup_event_backend/pup_event_module_devpath_log* | grep 'usb_storage'`" != "" ];then #see pup_event_backend_modprobe

the statement:

Code: Select all

if [ "`cat /sys/bus/usb/devices/*/modalias | grep -m 1 'ic08'`" != "" ];then #110225 usb-storage modalias interface fields are ic08isc06ip50, if needed.

The text always appears, implying that the check was made for the the USB devices being ready. This seems a more direct way of determining whether to check for the readiness of USB storage devices.

Thanks.
Richard

Thanks, I have deleted those files.

The test for usb_storage seems ok to me as it is. For most PUPMODEs, usb-storage.ko has been loaded in the initrd and the wait for scan already done. Hence you will not get the "usb-storage" message at bootup.

However, in the case of a full hd installation, no initrd, then when pup_event_backend_modprobe loads usb-storage.ko it also logs that to file, then that test in rc.sysinit will test as true and the wait for scan complete will occur.

[BarryK]

lxterminal-0.1.9-i686.pet is a terminal emulator with tabs and can be
configured with various font styles and sizes,transparent background,
etc.
It needs vte-0.26.2-i686.pet installed as well or it won't run.

Billtoo,
Thanks for that. I have announce them on my blog:

http://bkhome.org/blog/?viewDetailed=02169

[Billtoo]

lxterminal-0.1.9-i686.pet is a terminal emulator with tabs and can be
configured with various font styles and sizes,transparent background,
etc.
It needs vte-0.26.2-i686.pet installed as well or it won't run.

Billtoo,
Thanks for that. I have announce them on my blog:

http://bkhome.org/blog/?viewDetailed=02169

You're welcome, sorry that I didn't do a better job on the desktop file though.

[tronkel]

Barry wrote:

Have you posted this exactly as it is in the file? It should be:

Quote:
SKIPLIST=' nouveau arcfb cirrusfb cyber2000fb cyblafb hgafb neofb pm2fb s1d13xxxfb sstfb tdfxfb tridentfb vga16fb aty128fb atyfb radeonfb i810fb intelfb kyrofb nvidiafb rivafb savagefb sisfb btaudio es1371 i810_audio kahlua trident ub usbkbd usbmouse via82cxxx_audio rt2800usb rt2x00usb '


What you have posted is malformed.

Yes, that was exactly the contents of the SKIP variable as it appears on my system. So it does in fact seem malformed (contrary to what I previously thought). The contents of SKIP on my system do not match your version (the correct one). So something is amiss there.

[Sage]

I have two very old spare drives; only being able to address 650MB(74min).

Independent of the md5 issue, I have not found any drive old or new or any 'Linux' utility (Brasero, Xfburn, etc) that will reliably blank and/or re-burn 650Mb CD-RW s. Traxdata samples seem especially prone. Most of these discs have already been used successfully under earlier, much earlier, versions of a range of Linux distros, probably including Puppy back as far as series 1. In many cases, these can be fixed by running Nero under W98 ! Some of them were successfully restored using a trial Linux version of Nero, but the trial period ran out.

Progress: One step forward, two steps....?

[rodin.s]

I like Wary-511, it's great but mplayer and ffmpeg don't work on i586. On the other hand there was good (for i586) mplayer/ffmpeg in Wary-0.3. Is it possible to install somehow mplayer/ffmpeg from Wary-0.3 into Wary-5.1.1? Is there PETs maybe.
Solved: downloaded ffmpeg-2010-06-14.tar.bz from Barry's site, unpacked and made pet (without source files). After installing this package into Wary5.1.1 both mplayer and pmusic work on i586.

[gcmartin]

Wary has many good things going for it. I have found one annoying problem that I MAY have reported earlier.

When I create a doc in Abiword and try to drag the screen's contents to some forums open in SeaMonkey, I am getting incomplete drag and drops. This is particularly annoying but, I am unsure about how to trap to provide additional useful information for resolution.

There also have been issues resulting from my drag to Geany and trying to save it. This gives me refusal messages from Geany about the characters in the file(s).

One character that comes to mind is the quote character (") where some of the websites just flat won't allow the drag-and-drop transfer.

I don't know if this is a clue, but on a completely different problem, here, it MAY HINT as what's wrong in WARY and ABIWORD. If this helps, please address this in WOOF and future WARYs.

Sometimes, we cut back so much in efforts to make things small, that we negatively affect the behavior of applications within the OS. i understand.

Hope this helps

[mcewanw]

When I create a doc in Abiword and try to drag the screen's contents to some forums open in SeaMonkey, I am getting incomplete drag and drops. This is particularly annoying but, I am unsure about how to trap to provide additional useful information for resolution.

There also have been issues resulting from my drag to Geany and trying to save it. This gives me refusal messages from Geany about the characters in the file(s).

One character that comes to mind is the quote character (") where some of the websites just flat won't allow the drag-and-drop transfer.

Maybe some or all of the problems you are documenting is just something to do with the code page/Encodings the application is set up to use. For example, by default Geany (as seen in Edit menu > Preferences > Files) is installed to use Encodings: Western (ISO-8859-1). I have a feeling (but could well be wrong) that I've read somewhere that changing that to Unicode (UTF-8) is necessary for some character sets to be recognised correctly. Hopefully someone who knows more about such matters will comment.

[Shep]

Maybe some or all of the problems you are documenting is just something to do with the code page/Encodings the application is set up to use. For example, by default Geany (as seen in Edit menu > Preferences > Files) is installed to use Encodings: Western (ISO-8859-1). I have a feeling (but could well be wrong) that I've read somewhere that changing that to Unicode (UTF-8) is necessary for some character sets to be recognised correctly. Hopefully someone who knows more about such matters will comment.

The encoding setting in geany is the bane of my life. (Well, not quite, but you know what I mean.)

Often I copy something from a web page into geany, edit it, then try to save to a file and it refuses because of encoding conflicts. So I have to tediously go through the document, editing out weird characters and typing in approximations. For example, I have to replace sloping apostrophes with the vertical one on my keyboard; replace a longer hyphen with the one on my keyboard, and replace a character consisting of 3 adjacent dots with 3 characters each comprising a single dot, and so on. This is a real PITA.

I have tried various encodings, via GEANY > Document > Set Encodings but when I find one that works it doesn't seem to stick for next time I use geany. Yes, I think UTF-8 might be the universal one to aim for.

I just tried your suggestion: Edit menu > Preferences > Files so will see if that sticks. I got rid of the bulky icons while I was there, too, don't need them taking up some of the editor window.

[Shep]

Upgrade for USB 3G modems!

If you have wary 5.1.1 (or 5.1 and you have applied the service_pack_wary-5.1-to-5.1.1.pet), I have created a Service Pack that upgrades usb-modeswitch to version 1.1.7. See my blog post:

http://bkhome.org/blog/?viewDetailed=02166

If you have a USB 3G modem, I would appreciate if you could test this Service Pack. I am particularly interested to get feedback on coldplugging versus hotplugging.

Until now I have been using the 5.1 CD + your zzz .pet. This has demanded hot booting.

Today I applied your 5.1 to 5.1.1 service pet, and also your 5.1.1 to 20110303 service pet.
I have switched off and find that upon switching on the USB modem is found if already plugged in (coldplugging). I have also switched off and find that upon switching on with the USB modem not plugged in, it is found if then plugged in (hotplugging). K3571-Z

This will make the start-up procedure just that much simpler. Thank you for your continued work with puppy, Barry.

Note that if your 3G modem uses the 'hso.ko' kernel module, use Wary with the 2.6.32.28 kernel. I think that hso.ko is broken in the 2.6.31.14 kernel (?)

I have no idea what my modem uses.

[BarryK]

Until now I have been using the 5.1 CD + your zzz .pet. This has demanded hot booting.

Wary 5.1 already has zzz, the pet is not needed. On the otherhand, if you mean Lucid Puppy 5.1 then it does need the zzz pet.

[gabe]

This is on a fresh installation Wary 5.1.1 on my Inspiron 2400. It's slow as molasses. I didn't do anything except install vlc and firefox -- and I'm pretty sure this problem existed well before then, too.

"I'm running kernel 2.6.31.14" <-- How? [edit: from the blog post, I found this: http://distro.ibiblio.org/quirky/wary-5 ... oc-ide.iso -- what's the IDE for? Is 2.6.31.14 useful for anything other than modem drivers? Is it lighter? Please advise!]

It's unusably slow, yet I'm barely using it enough to type this response. The text is lagging by up to 5 seconds after I type, the hard drive seems to be constantly swapping as well. This is one of the most common desktop systems in the world -- at least a few years ago. Inspiron 2400, 256MB RAM. The first [edit: it appears second] screenshot was a clean boot, the second [first on the post] is with one other program running (in addition to the terminal with top) -- firefox displaying only this simple forum. Typing this very response.

Any suggestions? Normal Puppy 5.11 and 5.2 fail utterly with a crash sometime after use -- I blamed the kernel. (memtest ran for hours without an error, as did XP).

[Billtoo]

xfe-1.32.2-i686.pet (X File Explorer) is a filemanager similar to midnight
commander, it also needs fox-1.6.43-i686.pet installed or it won't run.

[gabe]

The i810 driver is now being used and things seem just a tad better, but trying to "upgrade xorg" from the wizard leads me to install a package called "root" and it obviously fails. (To reproduce: Menu -> Setup -> Wizard Wizard [lol, the name itself is a bug] -> upgrade X drivers -> Xorg7.5).

So it looks like upgrading X is out of the question? And you say in the first post not to use XVesa? Just making sure. (Also, the new xorg could have been causing the random crashes in the past when I used normal puppy 5.11-5.2, so my "wariness" in this regard is justified, no?).

Edit: From the blog post, I found this: wary-511-k2.6.31.14-uniproc-ide.iso -- what does IDE stand for? Is 2.6.31.14 useful for anything other than modem drivers? Is it lighter? Would it for any reason be advisable on Intel 8xx hardware? Is there a pretty decent chance it could address the issue I've been having? Will it eat my pets and siblings alive? Please advise!

[starhawk]

Not 100% certain that this is a bug, but I get a very consistent kernel panic on Wary 511 on my hardware... it loads vmlinuz, unpacks init.gz... and splat. Says something about something not syncing correctly... but it sinks all right. Wouldn't say it's the right /type/ of sinking (syncing), but it /does/ sink!

Wary 500 works fine, but neither version (standard or uniproc) of Wary 511 seems to like my system.

System details:
HP Pavilion s7220n Slimline Desktop, Celeron M CPU @ 1.8GHz, 1GB of RAM, 200GB SATA HDD + IDE DVD Burner. Except for the RAM, everything is stock hardware.

-----

@Gabe: IDE, also called PATA, is the "bus" (wires and protocol) between, say, a hard drive and the mainboard of a computer. IDE has since been made obsolete by a new style of connecting things called SATA. In this case, however, it also refers to the driver-set that Puppy uses in order to communicate with a drive of /any/ kind (confusingly enough, this includes other buses, like SCSI and USB). The "ide" driver-set uses a different set of names for the drives that it sees.

In other words (or if you want to say tl;dr) -- it's something very technical that you probably don't need to worry about. BTW, uniproc is short for "single processor" (uni=one, like unicycle).