Chihuahua Alpha6 plus Lassie_Office

[prehistoric]

@MU, tronkel,

Good call there! I've removed the execs from wmreboot and wmpoweroff on the problem machine and done four reboots and a shutdown without losing the console messages. Now that I know, it seems obvious that the order in which processes were killed made a difference when this was another process. The non-determinism goes all the way back to start up. Removing the exec serializes the reboot/shutdown operations.

We still have non-determinism at start up, and so does Muppy 008.3. (Based on 3.01 w/ patches.) This is significant when HotPup starts up, and sometimes results in that pile of icons in the upper left corner. I've also noticed that blinky and freemem tray applet get in a tangle sometimes, because both disappear from the tray and a tiny window is stuck in the upper left corner. (If I seem to be hung up on these problems, blame it on years with simulators and real-time control.)

I've just made the changes for reboot/shutdown on this laptop. If the problem is solved there, we're out of the woods. Cue "Heilige Dankegesang".

[slvrldy17]

Have been mucking about off and on with various flavors of Puppy for quite some time now and I have to agree that some problems are on going. Strangely enough while I was working with Gray's 3.01 based NOP Puppy the Xorg problem wasn't really all that annoying - the needed tweaks for Xorg seemed to carry over between boots - what the difference between the versions is that makes this so I don't know.

Just tried to reproduce this on a machine where I've had to adjust the monitor because the modeline generated by xvidtune was ignored. I used NOP r6 and checked that the modeline was inserted in xorg.conf. It's there, but still gets ignored. What version of NOP are you using? Could you have adjusted the monitor and forgotten? Could this be another non-deterministic problem that only goes right once in a blue moon?

@gerry,

Not a bad idea, I've used the wrong pup_save on occasion.

Prehistoric,

Excuse the delay in getting back to you - I just woke up to go to work on an overnight shift. If memory serves the version of NOP that I was using was NOP 3.01r3 or r4 - I had it set up to boot from a 4Gb compact flash card in an IDE adapter and it is still sitting here - If it will help I'll put it back in the machine, boot it and confirm my memory. I can also copy off the output of dmseg and send that. Let me know - I'll be working midnights this week and it may take a day or two to do - sleep and work permitting - but is no problem from my perspective.

[prehistoric]

@rerwin,

I've tried your suggestion about the Option "PreferredMode" with the string matching the resolution named in that modeline. No luck at this location on getting xvidtune changes accepted when X restarts. Also tried this under Muppy 8.3, in case their changes in Xorg setup had solved problems. No luck with that, either. Changes made by xvidtune are temporary.

@slvrldy17,

If, as I suspect, your first attempt to reproduce the success you had before fails, could I impose on you to try several times. If you get it to work on any try it will be better than I am doing, and that will be an important clue.

Anyone, can you find a change between NOP 3.01 r3 and r6 which would affect this?

[jonyo]

Edit the xorg.conf file in /etc/X11 to uncomment this line in the "Device" section.

Code: Select all

	#Option     "SWcursor"           	# [<bool>]

Changing it to:

Code: Select all

	Option     "SWcursor"           	"True"

Don't see the above in xorg.conf or conf0.

Point of interest. Pup auto loads rt2570 & the interface in the net wiz shows rausb0. At the same time, rutilt shows rt2500 wlan & rt2500 is the win driver.

If I load my profile in the net wiz, then click use this profile, 'pute locks up solid though the mouse cursor still moves but doesn't work on anything. Only way out is the power button. I've had many lockups with rt2570 in diff vers of pup when doing things which I'd think are normal.

I get a connection by doing a scan in the net wiz first & that picks up my network. Only then can I use rutilt, click on my profile there & then get connected.

There have been many reports of rt2570 locking up 'putes in the 3xx series & I think even before. I had a boot crash (disk error) with xp one time too after a rt2570 lockup.

Bottom line is I can use rt2570, but it requires a specific sequence (gui wise) & doing it diff can cause a lockup.

There are no issues when I use the the win driver.

[tronkel]

@preheistoric

I've just added exec to the .icewm startup script. /root/.icewm/startup

#!/bin/bash
# this script is executed by starticewm when IceWM starts



killall icesound
icesound &

killall icewmtray
exec icewmtray &
sleep 1

killall freememapplet-tray
sleep 2 && /root/my-applications/freememapplet_tray_puppy215ce/freememapplet-tray &

killall blinky
sleep 3 && /root/my-applications/blinky-0.8-icewm/src/blinky -bg "#4f5ff3" &



# cludge repair for root menu corruption
fixmenus &&

This seems to repair the randomness of the startup order of Blinky and Freememapplet. The little frame at the top LHS seems to have also gone. Seems to be OK on this system here. Please try this on your laptop as well.

The irony of it all. Remove 2 showstoppers by taking exec out of one script and putting it into another .... lol

Now Firefox has just inexplicably crashed on me.
Never mind ... can't have everything I suppose

[prehistoric]

@Jonyo,

You must not have rerun xorgwizard to create a new xorg.conf. Do that, then choose the option to edit xorg.conf after you test.

@rerwin,

Check message below, where Jonyo confirms that others have been seeing this cursor problem for some time without reporting it.

@tronkel,

I've applied the hack and will check the effects in a moment. (Hmmm, conservation of execs, sounds like some company reorganizations I've been through.)

If you think I'm causing you trouble, you should see the screenshot of the latest bug I found in Muppy 008.3.

Added: So far, so good. I've restarted X a few times and it seems repeatable. I'll try the hack on that other problem machine.

Still good. I posted an erroneous report of a problem, then remembered I forgot to sync between the edit and restarting X.

Hotpup needs more debugging, but we now have some important basic GUI behavior under control.

BTW: Does calling me "preheistoric" suggest I'm a criminal mastermind?

[jonyo]

@Jonyo, You must not have rerun xorgwizard to create a new xorg.conf.

Yes, that's it. When I ran xorg, didn't make a save file at that point. I'll go over it again. thx
Btw, this prob goes back to ~ 2.15 & though I posted on it at that time, never resolved the issue, or heard of your fix 'til now. At that time, the cursor'd go away, or disappear at random, after working for a bit..sometimes days..

[slvrldy17]

Have been mucking about off and on with various flavors of Puppy for quite some time now and I have to agree that some problems are on going. Strangely enough while I was working with Gray's 3.01 based NOP Puppy the Xorg problem wasn't really all that annoying - the needed tweaks for Xorg seemed to carry over between boots - what the difference between the versions is that makes this so I don't know.

Just tried to reproduce this on a machine where I've had to adjust the monitor because the modeline generated by xvidtune was ignored. I used NOP r6 and checked that the modeline was inserted in xorg.conf. It's there, but still gets ignored. What version of NOP are you using? Could you have adjusted the monitor and forgotten? Could this be another non-deterministic problem that only goes right once in a blue moon?

@gerry,

Not a bad idea, I've used the wrong pup_save on occasion.

Prehistoric,

Excuse the delay in getting back to you - I just woke up to go to work on an overnight shift. If memory serves the version of NOP that I was using was NOP 3.01r3 or r4 - I had it set up to boot from a 4Gb compact flash card in an IDE adapter and it is still sitting here - If it will help I'll put it back in the machine, boot it and confirm my memory. I can also copy off the output of dmseg and send that. Let me know - I'll be working midnights this week and it may take a day or two to do - sleep and work permitting - but is no problem from my perspective.

Prehistoric.
Just woke up and decided to try the compact flash card/IDE adapter combo back in the test rig. Result is as before - video settings including refresh rate are there and usable. I suspect that their being carried over may have to do with the fact that I'm working with Puppy installed to and booted from the CF card rather than booting from a live CD with a save file only on the hard drive or other media. If you want to look at the output of dmseg or the Xorg.conf files let me know - I'll email or PM them to you rather than clutter this thread up by posting them here - I'll keep it set up this way for a couple of days.

[prehistoric]

@slvrldy17,

Yes! Please send me a PM with those attached, or post a message in this thread with gzipped text files attached, so I can download them. You should be able to drag and drop them on Pupzip. If that doesn't work, you can open a terminal and type, e.g.

Code: Select all

gzip xorg.conf

to convert the file to a gzipped file (with extension gz). This should be acceptable as an attachment. That way you don't have to clutter up this thread with all the text.

[jonyo]

Negative on you tube.
"Hello, you either have JavaScript turned off or an old version of Adobe's Flash Player. Get the latest Flash player."

[prehistoric]

@Jonyo,

Should have made this clear, Flash player is not yet built into alpha 6. I have been installing from pets for testing. The puppylinux.ca/bugs directory has updated versions of many Puppy packages, including a Flash player 9.0.124.pet. On slow machines with limited memory you may do better with Flash 7, as used in Muppy 008.3.

@tronkel,

Now, I can't even go to bed without tripping over a concurrency problem. When I finished last night, the "Saving RAM to 'pup_save" " message was displayed when I did shutdown. This morning I found the laptop still on, and still shutting down. (Fortunately, it was on external power.) Looks like a classic deadlock.

Added: I've had HotPup disabled on my main installation of alpha 6 while I was chasing other concurrent behavior. After I found an assumption that all mounted drives had partitions in some HotPup scripts (in /etc/hotpup) I started checking for other places where missing partitions might cause problems. Pmount handles USB drives formatted in superfloppy mode correctly. However, if I try to mount a pup-save.2fs file inside that device I get an error.

Apparently, there is also an assumption about partition numbers in the code which mounts these files when you click on them. If the file is in /mnt/sda1, it mounts without errors. If the file is in /mnt/sda, I get an error message.

[jonyo]

Got Flash player 9.0.124.pet from the Canuck site, installed & good to go. Installed the latest seamonky from the official site to usr/lib & good to go.

Petget mgr's not working for me ..~ failed dload scenario something. Booted into Lighthouse & works fine.

Can't get Opera going, neither the official 0.23 pet, nor the slackware setup I usually use. Went to lighthouse & setup ok.

Had a few lockups (solid) when I attempted restart x server a few times (..other times worked fine) in ..somehow dealing with "True" in xorg/conf but the cursor works now. Have to sort out refresh rates (I think). Get some faint but noticeable upward horizontal pulsing mainly on the desktop. Guess I should try to load a nvidia driver.

I have to click on ~ 8 things (described in prior post) on each boot to get the rt2570 wifi up & running, which humms then, very quick. Any way to fix? (tried all I know..) I'd like to boot straight into wifi. Failing that, I have much better luck with the win driver.

Tried to follow & mod tempestuous's how to here, changing eth1 to rausb0 but got an error.

Code: Select all

rm /etc/dhcpc/*.pid
dhcpcd -t 30 -h puppypc -d eth1

[prehistoric]

@Jonyo,

I am using the Opera 9.27 pet from Canada. So far, I haven't run Flash player under that, because I'm still chasing concurrency problems with the regular ISO and Firefox.

I've mostly been operating with HotPup disabled, because of problems with lock ups, etc. I've also applied MU's changes to wmreboot and wmpoweroff, plus tronkel's changes to the wm startup script shown above.

I agree with you about wlan configuration, particularly with rt25xx drivers. New drivers should be out soon and the wireless wizard is due for an overhaul.

I'll need to check petget, I thought I had tested it, but that may have only been with files already downloaded.

[jonyo]

9.27 Opera A-ok.
Firefox3 Beta 5 doesn't seem to save passwords.

[slvrldy17]

@slvrldy17,

Yes! Please send me a PM with those attached, or post a message in this thread with gzipped text files attached, so I can download them. You should be able to drag and drop them on Pupzip. If that doesn't work, you can open a terminal and type, e.g.

Code: Select all

gzip xorg.conf

to convert the file to a gzipped file (with extension gz). This should be acceptable as an attachment. That way you don't have to clutter up this thread with all the text.

Sent you a PM with the dmesg and Xorg.conf files we talked about - hope it works - let me know if it made it through - I'm off for a nap!

[alienjeff]

And this thread goes silent as Puppy 4.00SC/RC and Final are released. The "finite masses" flock to the official release, thus putting a serious crimp on testing of this CE, variant or version du jour -- lending credence to the supposition that there are too many versions and too few testers.

Remember to register your Puppy v4.00 Final online to secure your Puppy Genuine Advantage benefits, including product news, security updates and SP1 auto-install.

Coming soon: Puppy v4.00.1 codename Whoops

[tronkel]

Alienjeff wrote:

The "finite masses" flock to the official release, thus putting a serious crimp on testing of this CE, variant or version du jour -- lending credence to the supposition that there are too many versions and too few testers.

lol ... yes.

I once read somewhere that Microsoft pay people to sit in usability workshops and try to spot where the faults are. I think they're wasting their money on this. Even paying out good money, they still can't get it right.

Testing is the most important activity of the user community. If you want that perfect OS tool, you should help to test it. OK, you don't get paid - but you get to use it for free. Money can't buy everything.

[prehistoric]

9.27 Opera A-ok.
Firefox3 Beta 5 doesn't seem to save passwords.

Just tested this on my installation. I don't have trouble. Could you have accidentally set preferrences to ignore passwords?

@AJ,

Since I haven't been involved with Dingo, it has not affected my testing. However, non-deterministic problems which show up while playing videos have required a lot of testing time. We still have occasional lock-ups and 3b5 and/or Flash 9 probably still has slow memory leaks. I've also lost confidence in what freemem tray applet tells me.

@tronkel,

I've removed another exec in /usr/X11/restartwm to stop screen lockups. Other than that code hasn't changed. I, on the other hand, am worn to a frazzle. It doesn't seem possible to spend hours watching a variety of video from the Internet without getting a seriously warped view of relationships between people. Yes, some things are funny, once. (Here's an example.) I am now planning to spend several days reading books, to cool my overheated imagination. Someone else can test video playback.

prehistoric

[ttuuxxx]

naaaaaa when it comes to dingo the new kernel is faulty, And the latest main version has a faulty clipboard, most of the time and just frustrating to play with. I spent maybe 3hrs with it and thats it, the cdrw is erased, and I'll be loyal to the 3.0 series since I like slackware so much. Why would I want an operating system that can only do 1/8 of what a previous version can. Thats just crazy
I'm not to worried about people flocking to 4.0 and not testing 3.2.
2.15ce was the biggest release I think of any puppy. People are still downloading it.
ttuuxxx

[prehistoric]

@alienjeff, etc.

Now that you're posting bugs found in Puppy 4.00, try to remember to post a note here when you find something working there which has not be satisfactory in Chihuahua. Having working examples to compare makes debugging a great deal easier. When we have people testing multiple versions, we might as well get all the advantages.