dpup-482beta testing

[Béèm]

If wbar pops-up over everything when the top of the screen (or what ever side wbar is configured to) I vote for wbar.

I remember this handy functionality in the muppies of MU.

[2byte]

gposil,

I saw your pics of the new control panel in another thread and noticed that you have a screensaver icon but nothing for the mouse and keyboard. Assuming you are keeping the screensaver from beta4 (I hope) a nice addition would be an icon for the mouse and keyboard. This could be a link to pupX with the screensaver tab removed. I removed the tab in my beta4 just by deleting all lines with 'screensaver' from the pupX script.

Just a suggestion.

[Béèm]

@Guy (and others who are interested)

I appreciate the efforts you make on this issue.
The extra 256MB could make a difference.
(Do you have a Belgian keyb as well? )

You will be more stumped.

I use the Lin'N'Win method, you booted from CD.
I don't do that any more, but for this occasion I created a CD and booted from it.
Config of keyb, X, locale, time zone and network went well and I rebooted.
This time I got the option to create a save file.
After the reboot from the CD I selected that newly created save file.
But instead of the desktop, I got the cmd prompt.
When typing xorgwizard I realized that my machine wasn't in the Belgian keyboard but the US one.
So apparently all the config done before wasn't retained.
So now I am stumped as well.

From the cmd prompt I could save the bootxxxx.log files, in case that might be interesting to look at.

Happy reading.

[2byte]

Béèm,

Since you use the Lin'N'Win method, have you tried defragging your Windows hard drive before creating the save file? It's a long shot but you never know.

[gposil]

gposil,

I saw your pics of the new control panel in another thread and noticed that you have a screensaver icon but nothing for the mouse and keyboard. Assuming you are keeping the screensaver from beta4 (I hope) a nice addition would be an icon for the mouse and keyboard. This could be a link to pupX with the screensaver tab removed. I removed the tab in my beta4 just by deleting all lines with 'screensaver' from the pupX script.

Just a suggestion.

Yes 2byte...already in beta5...

[vtpup]

Béèm,

Since you use the Lin'N'Win method, have you tried defragging your Windows hard drive before creating the save file? It's a long shot but you never know.

Driving back from Boston today, my mind wandered to the supposedly off limits area of Béèm's savefile problem again. No stop, don't think about it, I heard a voice say. Concentrate on driving. But no, why do other R40's not have the problem? What's different?

Can it be LFAT? That's been stable for a long time. But it probably is different from gposil's test, isn't it?

No, enough, I thought. Just stop at the next gas station and get a Snickers....

[gposil]

Back to Béèm's problem...

From Béèm's logs....
Two errors immediately standout...kernel log reports corrupt sfs file and sysinit reports azerty Belgian keyboard map load error...

Just one of these errors is enough to ruin your day, let alone both of them...the more serious one is the squashfs error and could well be causing the second...

I tried with Belgian keyboard on The R40 here and it loaded ok...so it leads me to believe we have a cascading error based on the filesystem problem...as far as I am aware, the cause of this can be media related, partition related or I/O related...but I am not sufficiently expert in these matters to venture a concrete opinion, other than vtpup's musing about LFAT...

I'll continue investigating this...Béèm

[gposil]

Béèm,

I have done some more investigation, both in Debian and Ubuntu.
The

SQUASHFS error: zlib_inflate error, data probably corrupt

error, is universally associated with hardware failure not OS or software...it appears that roughly 50% of problems are RAM related, about 40% media related, and the rest evenly spread between SATA/PATA controller failure and imminent Drive failure....

I will keep looking at it, but I would certainly investigate the Hardware...

[vtpup]

Well, I don't know if it's true of the R40, but the T30 and 600e have definite ram socket solder problems. They flex and crack the solder joints after awhile. I know --- I soldered the 400 joints on my T30 to make both memory bays functional again -- it was an ebay machine. Same thing on the 600E in one of the bays.

Symptoms are intermittent memory problems at first -- often temporarily cured by users inserting cardboard or even guitar picks between the ram module and the cover plate. Eventually this trick doesn't work with one bay. Then they get a bigger ram for the other bay, and remove the ram from the bad one. When that second one starts acting flakey they put it on Ebay.

[01micko]

"Bring back the Paw and Bones!"

*drumroll*

Here are ttuuxxx icewm themes from 4.2, they all work, I just replaced the menu button on some ... because this is not 4.2! Uses standard dpup button, like the default one in Jwm.

*Round of applause.*

Thanks ttuuxxx.

ttuuxxx-small-ice-themes.pet (50k... 5 themes )

Cheers

[01micko]

While I was at it I updated the Icewm package with all the 4.2 stuff such that you can point and click to switch between Jwm and IceWM. It should work fine, please report if there is any problem, I did test on a pfix-ram install and it worked ok. There is the possibility of timing issues when Icewm starts up. You may see a little window in the top left corner, if so then some sleep needs to be added.

icewm-1.3.5-me1-i486.pet (645k)

Cheers

[Béèm]

Some new elements in the diagnostic.

Just to remind my setup.
I use Lin'N'Win
On sda1 is my windows installation
I created a dpup482b3 and dpup482b4 directories in which are the corresponding beta3 and beta4 files.
I put all save files on sda2 (/mnt/home normally)
From now on I booted from CD only.
At first boot from the CD I was able to create a save file, which explained I did put on sda2
At the reboot selecting this save file with the results know and of which I already uploaded the boot log files.
With upup 476 I checked the created save file and gor errors.

Code: Select all

# fsck.ext2 dpupsave.2fs
e2fsck 1.41.4 (27-Jan-2009)
dpupsave.2fs was not cleanly unmounted, check forced.
Pass 1: Checking inodes, blocks, and sizes
Deleted inode 846 has zero dtime.  Fix<y>? yes

Deleted inode 847 has zero dtime.  Fix<y>? yes

Deleted inode 848 has zero dtime.  Fix<y>? yes

Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 3A: Optimizing directories
Pass 4: Checking reference counts
Pass 5: Checking group summary information
Block bitmap differences:  -(5634--5637) -(5641--5685) -(5687--5694) -5698 -(5713--5714) -(5726--5727) -(5745--5748)
Fix<y>? yes

Free blocks count wrong for group #0 (7013, counted=7079).
Fix<y>? yes

Free blocks count wrong (25820, counted=25886).
Fix<y>? yes

Inode bitmap differences:  -(846--848)
Fix<y>? yes

Free inodes count wrong for group #0 (1197, counted=1200).
Fix<y>? yes

Free inodes count wrong (6786, counted=6789).
Fix<y>? yes


dpupsave.2fs: ***** FILE SYSTEM WAS MODIFIED *****
dpupsave.2fs: 1403/8192 files (0.4% non-contiguous), 6882/32768 blocks
# 

But even with the errors corrected, this save file is unusable.

So I rebooted from CD with pfix=ram
This time I creates a save file on sda1 in stead of sda2
To my surprise the save file is stored in dpup482d3. Don't understand the logic behind this.

Anyway I rebooted from the CD and selected this newly saved save file.
This time I got X started correctly and I can work.

But my way of operating isn't the normal one as now my /mnt/home is sda1, the ntfs windows partition.
I don't like that. I leave the ntfs partition unmounted unless really needed.

Maybe these new elements will give an idea to someone.

EDIT

Well I did one more test.
Instead of booting from the CD I booted from the beta4 entry in the menu.lst and selected that save file on sda1. (dpup482d3)
Apart from the network settings which aren't retained, all seem to work.

[Jim1911]

While I was at it I updated the Icewm package with all the 4.2 stuff such that you can point and click to switch between Jwm and IceWM. It should work fine, please report if there is any problem. . . .

Works fine with one exception. After switching from jwm to icewm, icewm works fine. However, when switching back to jwm, jwm works fine, but the shutdown portion of the shutdown menu will not work, advanced section still works.

Cheers,
Jim

[musher0]

(Please forgive me if this is a repeat from the PM messenger, which does not appear to work properly.)

While I was at it I updated the Icewm package ...
Cheers

Hello, 01micko

I hate to mention this... This will not work in your ~/Startup/startice script :

Line 7: if [ "$CURRENTWM" = "ice" ]; then

'Cause... there is no window manager called "ice" in your package...
Probably an oversight, eh?

Did you mean rather:
Line 7: if [ "$CURRENTWM" = "icewm -o "$CURRENTWM" = "icewm-session" ]; then

In case someone wants to use the session feature ?

Besides, why not use icewm's own ~/.icewm/startup script ?

[musher0]

Also, you might not need these two lines, since icewm has its own monitor for the Internet line (ethernet).

# killall blinky
# sleep 4 && /root/my-applications/blinky-0.8-icewm/src/blinky -bg "#000000" &

Besides, blinky is less useful, it just blinks if there's ethernet activity, whereas the icewm ethernet monitor shows the actual net activity in colours, in a little graph, at the top and the bottom of its little window.
The icewm one gives more specific info, too... ( Sorry to be such a party-pooper! )

[2byte]

gposil,

This one is too easy

wbar takes a long time to load on my system. In the beginning of testing I actually had an application running before wbar appeared and it captured the application window as part of the wbar background. Anyway I just renamed /root/Startup/wstart to /root/Startup/0wstart (wstart with a zero) and presto, it loads much sooner.
.

[musher0]

As to Jim's problem, 01micko, would you not need to introduce this at the very beginning of the startice script and then in the exit script ?

CURRENTWM=""

This clears the name of the wm. Obviously, if the exit script thinks you're still in icewm and you're really in jwm, the jwm things won't work!

(I'm laughing at the script, here, not at any person!)

So the exit script needs a way to distinguish which wm is used.

Maybe a plain " restartwm jwm " or " restartwm icewm-session " (not as variables) will do the trick in the exit script ?

I haven't dissected the exit script, mind you, (I'm at work) so I don't really know. This is from memory, from my work on wmx.

Just a thought.

P.S. Maybe look at aragon's exit script for inspiration? (Puppy 4.2 testing thread I think. Again, no time to check, I'm at work...)

[vtpup]

Béèm,

Memtest86+

Overnight.

If you don't already have it on a liveCD somewhere,

http://www.ultimatebootcd.com/

[2byte]

gposil,

A thought about secure browsing in dpup. I just noticed that Firefox cannot be run as spot. It's missing a couple of gnome directories in /root/spot/, at least. With all the recent interest in secure on-line banking and secure browsing in Puppy, should this be addressed?
.

[mikeslr]

First gposil, fantastic job!! Well done! It the fastest puplet I've used. And to all the testers, great job. I'm only sorry I couldn't join you so far. I downloaded the latest iso, beta 4, but desiring to know what problems to expect and not wanting to repeat problems already covered, it's taken me several days to read thru the thread --which often kept growing as fast as I could get around to reading it. I'm not certain how much help I can be, being only little more than a noob, but with two fully functional systems to play with I'll do what I can.

The two systems: Desktop, emachine (Gateway? from Bestbuy) 2 gigs Ram, AMD Athlon 64 X2 3800+ / 2.2 GHz, built-in NVIDIA GeForce 6150LE.
Laptop is a Thinkpad T42, 1 Gig Ram, otherwise factory specs.

On the Desktop, booting from CD, using 1024x768 xorg, display is off-centered to the right and missing the top of available screen by about 1/8 inch. I manually adjusted using controls on my CRT. Rebooting into XP, the adjusted offset was not retained; that is XP's nVidia driver centered screen properly. I haven't yet tried an nVidia pet.

On above and following computer, Network wizard finally connected to my wireless network. I had to run it several times before it did. But I've grown used to that as its been a recurrent problem on all Puppies since 4.12, and occasionally with PCLinuxos 2009, and Pure-dyne. (All other distros fail completely, being unable to handle restricted/shared WEP). The Desktop uses a D-Link PCMCIA WNA 1330 Adapter. This is not your problem, but as I've noted others on the Forum having to run Network wizard a couple of times for it to "find" a live network, perhaps someone with more credibility than I should suggest that a warning be included in the wizard's instructions so that newbies don't abandon the effort prematurely.

Update:Subsequent efforts to connect thru Desktop D-Link have been unsuccessful. Network wizard sometimes didn't see any network, including an open network in my vicinity. Both pwireless scanner and network wizard (when it did see network) did not connect. pwireless scanner did not save profile. When a static IP was set in the network wizard, success was reported, but Firefox --though not reporting 'server not found'-- did not load webpage. There are two things every operating system must be able to do: generate wordprocessing documents and access the Internet. If Puppy doesn't soon get its wireless capabilities in order, I'm afraid my passion for it may wain.

Like others using Thinkpads, the default setting of "numlock on" creates keyboard problems. Commenting out appropriate line in .xinitrc appears to solve them. Gposil mentioned somewhere along the way that among the startup wizard buttons would be one to set "numlock" to "off" and that help files would be accessible for assisting in the determination of choices. I strongly recommend that the help file for the "numlock" button identify computer models known to require the "off" setting. This should include all Thinkpads (and perhaps other brands) which under Windows use a Ctl-key combination with keys on the right side of the keyboard as a substitute for the lack of a numeric keypad.

[Personal note to gposil. I know you were kidding when you suggested that owners of Thinkpads replace them with Toshibas. There are those of us, who having owned a Thinkpad --known for their reliability, durability and comfortable and dependable keyboards-- are far more likely to replace an operating system than give up this trustworthy piece of hardware. I got "into Linux." When I became exasperated by the demands Xp placed on my Thinkpad 600e. When the 10 year old 600e's cmos battery and keyboard finally became erratic, it was replaced with the Thinkpad T42 whose cost was only slightly more than the repair of the 600e would have been. Four years ago, the T42 was about $1,300. Purchased on ebay, (albeit w/o an operating system I wasn't going to use anyway) its cost was $100s less than a new netbook: money I may apply to a Terabyte USB drive, more than negating any advantage a netbook would have provided. Recommendation to developers: if it won't run on a Thinkpad, you've eliminated a substantial portion of your potential market.]

As noted by others, Rox defaults to the last used folder rather than "Home."

Abiword 2.8.0 does not have a built-in thesaurus. mcewanw was kind enough to direct me to the Artha.pet created by sikpuppy. I'm please to report that it's fully functional under dPup. Its available here: http://cid-da097326680e08e6.skydrive.li ... 0.9.1a.pet.

If your computer has limited RAM, please note this is a "terminate & Stay resident" application so that it may be called up from any program using a hot-key combination. EDITED: Errata deleted. Per Artha thread, Right-clicking its taskbar icon enables closing the app completely.

Mu's dual pane file manager, Muppy-filer, is also fully functional.

Dissatisfied with Puppy's labyrinthian menu, the limited functionality of Wbar, and prefering an almost icon free desktop, about the first thing I install is icewm windows manager: this because neither the current version of Rox nor jwm enable the creation of taskbar toolbars for quick access to favorite apps such as I outline at http://www.murga-linux.com/puppy/viewtopic.php?t=41923. So, before 01micko announced the availability of the ICEWM-1.3.5 pet, I installed the EZPup-4.0 pet from http://www.puppylinux.ca/tpp/EZPup/ as well as the EZicewm-themes.pet I had acquired somewhere. (I save to disk pets and sfs' I commonly use). These appear to work well, with only one annoyance. The themes pet overwrote dPup's default wallpaper, named "default," which I really liked. To recover it, I had to boot my other computer and copy it. I suggest that "backgrounds" should include a second copy with a unique name.

Booting with a clean save file, I installed 01micko's ICEWM-1.3.5 pet. This worked for a few minutes. Unfortunately, I may have misunderstood ttuuxxx's comment as I also installed his themes pet Version 3.2. Note the Version. When I attempted to change themes, my system crashed and hung and I had to hard boot. I'll try again with a clean save file and see what themes are installable.

The problem with either wbar or using a quicklaunch toolbar under icewm is installing the apps you want is very labor intensive. This is because there is neither a central location for executables nor for icons. Before the text file referring to them can be written, a search for each required item must be conducted (unless, for the icewm toolbar, the executable is in the "path"). It occurred to me that Puppy Package Manager has to read and retain that information in order to be able to install and uninstall apps. So I wonder how difficult it would be to include a "subroutine" in ppm which, when an app was being installed, would write to a text file, saved in say, "My documents" folder, the name of the application, the full path of its executable and, at least, the path to the folder containing its icons? Then someone more cleaver than I might be able to modify 01micko's wbar configuration tool to use that file when the population of wbar is to be changed; or to write/rewrite the toolbar file when populating an icewm quicklaunch toolbar.

I look forward to beta5 and hope that I can be of some assistance in vetting it. I'll note that I have a Dell Latitude Cpx I can also do some testing on, albeit its semi-functional keyboard may raise more questions than provide answers.

Mike Kessler