Puppeee 4.3X

[01micko]

Here's a fixed eee-control for the 701SD.

Works a treat, thanks

[jemimah]

Here is xpdf and evince if you're having trouble with epdfview. Evince needs the updated libpoppler that I've posted here but xpdf is standalone. Evince and Epdfview both render with poppler so you'll probably get similar results. Xpdf has an ugly interface, but it's stable and seems to work on more pdfs.

[zicozico]

Hi Jemimah,

I'm running Beta 4 right now. Pretty fast booting and running in general.

One thing though, how can I get hibernation working from /dev/sdb2 (USBFLASH)?

It seems that suspending to /dev/sdb2 swap partition (4GB) is working OK, but resuming from it not because puppy always boots up normally without looking to the hibernated image at all!

This is my grub config:
title Puppeee B4
rootnoverify (hd1,0)
kernel /puppeeeb4/vmlinuz resume=/dev/sdb2 pmedia=usbflash pfix=ram nosmp root=/dev/ram0 psubdir=puppeeeb4
initrd /puppeeeb4/initrd.gz
#end puppy B4

Any suggestions?

Regards,
Zicozico

[flatmo]

Jemimah,
Thanks for the new PDF program. I tried overclocking yesterday and there was no in Chrome (Youtube). Nice to see the system suspending after closing the lid. After several suspending cycles my computer crashed. I can' get Xorg or Xvesa to work...

[flatmo]

I meant there was no noticeable difference overclocking.

[jemimah]

Zicozico, I'm not sure if resuming from usb can work. Probably the kernel checks for the resume device before loading usb. I'll see if I can get to work, but your bootloader looks correct to me.

Flatmo, mixing overclocking and suspend might be a bad idea. Sometimes suspend crashes anyway - but it seems more stable to me than it used to be.

I don't know about the chrome/youtube problem - still works fine for me - even full screen. If opera works better, go for it.

Usually typing 'xwin' on the console will bring Xorg back after a crash. If you boot with 'pfix=fsck' it will also fix any filesystem corruption.

This is beta testing - so it's best to backup your save file frequently.


----


I now have a working kernel patch for the 701SD - it was trickier to figure out than I thought it would be. 4.1 should be out next week with the fix.

[jemimah]

I seem to have gotten unionfs to work. So hopefully your save file will be safe in the next release.

[zicozico]

Hi Jemimah,

Please note that my /dev/sdb1 actually is the 32GB SDCARD in my 701 (is accessed via USBFLASH)
The reason I hibernate to that drive is because my 701 was upgraded to 2G of RAM; hence I need 2G of swap space for hibernation which I find difficult to free up on /dev/sda because that internal drive is only 4G in total.

I really appreciate all your efforts!!

By the way great that you got unionfs working again!

One other thing: I guess that if the savefile gets corrupted it might also appear like hibernation is not working as files that were in use are not reachable anymore for instance. Just to keep in mind that corrupt savefiles can trigger all kinds of unrelated erratic behavior....

Regards,
Zicozico

[zicozico]

Hi All,

Due to the fact that I (and many others) have savefiles that are in use for a very long time and probably have been upgraded several times, these savefiles become more full over time. This is likely due to the fact that they might contain copies of files that in fact are also available in the (latest) distribution SFS. Hence unnecessary.

Now I was thinking, wouldn't it be nice to have a tool that simply parses through the whole directory tree on your save file, and for each file existing in both the savefile as well as the distribution file do the following:
1) is the size equal then byte compare the files
2) is the file exactly the same -> delete it from the savefile
3) is the file different, then leave it alone

Anything wrong with this theory?

Regards,
zicozico

[hokal]

Jemimah,

because I had problems upgrading to 4.0, (remember I lost my boot drive) I now did a new frugal installation. Everything went fine. But this morning I booted again .....and ... yes, lost my boot drive again. I still can boot from that drive, but have no access. Shutting gives again xt4-fs problem.

No problem with version 3.6

There must be something wrong with that kernel for sd drives. My boot drive (sdb1) is 4GB sd and my data drive (sda1) 32GB. I'm pretty sure, that it works for hds. The boot drive is gone. Also not to be seen with fdsik -l. Funny thing is, when I start Gparted , I'm offered sdc or sdb, depending whether I boot with stick or internal drive. But when I push "okay", -no (boot) drives detected.

Can someone test and confirm this with an EEEPC 901 sd drives ???

Also Eee-control FAN "Apply Threshold Changes" will not work.

[jemimah]

Zicozico,

I believe that's what pfix=clean does - well, it also deletes files that you've changed as well, so certain settings do get lost. What you can do, is make a copy of your save file (not while it's mounted), and rename it. Then you can click on it to mount it, and delete unwanted files.

I think most of the time, unless you do things like install packages that are already installed, the files in your save file will be different than the ones in the sfs.

Or you can create an empty save file, install everything you want, and set everything up properly. Then make a backup of that save file - and reuse it when you want to start fresh. If you know what you're doing, you could then convert that save file into a personal sfs or pet package.

--
Hokal, I'll take a look at it. I did a few things to speed up the init script that I may revert and see it that fixes it. 4.1 should be released by Tuesday or Wendnesday.

[jemimah]

Hokal, could you also open a terminal, type dmesg, and send me the output?

[zicozico]

@hokal

I recognize the problem with disappearing drives, although also in earlier versions.
Usually I can trigger it when I boot with an USBkey attached. Then sdc1 and sdb1 might get swapped after which sdb1 (my SDCARD) usually is gone on next boot.

I can get it back by issuing a mknod command (google mknod /dev/sdb1 if necessary) in the /dev directory.

I think your problem is simular to mine.

@jemimah
thanks for your suggestions. Will do a major cleanup once the savefile corruption is fixed.

[prehistoric]

Last night I saw a new Eee-PC a friend bought at HHGregg, the 1001P. This looks good and has a keyboard layout without that short right shift key. I'll report on a test once I get an SD card set up for it.

It also came with W7 starter. Since I was (almost) there when it came out of the box, and saw it start up once it was charged, I noticed a worrisome feature: there are now two end user agreements. One is essentially the familiar EULA. The other is an on-going "support agreement".

What I saw when that went by was that it applied whenever "you use your machine". My friend didn't get it to boot off the flash drive, which threw him into a boot sequence which demanded you agree to these terms before you do anything else. There was a long delay while it installed W7 from a factory restore partition on the hard drive.

Does anyone have a copy of the text you have to agree to if you use a machine with W7 installed?

[hokal]

Jemimah,

here is my dmesg-file. rgds hokal
attention it's a txt file.

[BJF]

Jemimah: A little abberation noted when I started up this morning...
Although the icon(s) lit up, nothing was called. Save queued but didn't, and it wouldn't reboot or close. I did Cntl-Alt-Back and when xorgwizard did nothing I did a hard stop. It works again now with X intact, something I haven't managed before.
And right from new, no combination of fiddling with the icon manager so far has changed the fact that a new USB item plugged in appears on the desktop jambed up against the left margin over the top of sda1 even though sdb1 stays put. In previous Beta's sdc1 would line up to the right and behave nicely. What am I missing?

Thanks.
John.

(edit) BTW, it's a Celeron with the correct processor install.

[flatmo]

Hi,
I'm having some weird issues when testing the installation of Open Office.
I tried installing three different versions of OO and I always get the sane result: all programs in OO work fine except for the writer (I'm loading the sfs at boot).
When I click in OO Writer puppy freezes and I have to remove the battery and unplug it to get it out of that state. I don't know if perhaps there is a conflict with Abiword but is strange that the only OO program that does not work is writer.

[jemimah]

BJF, if you open the event manager, there is an option to delete and redraw the drive icons. Or you can just delete them all and restart X and it should redraw them correctly.

Flatmo, it looks like a kernel bug. I tested OpenOffice writer on the 2.6.31 kernel and it works fine, but it definitely kills 2.6.33. I'll grab a copy of 2.6.33.1 today and rebuild and see if the problem persists. I imagine this one will get fixed quickly since the script kiddies will have a heyday with it otherwise.

[jemimah]

Yep, Openoffice Writer does work under 2.6.33.1 so I'll use that for 4.1 Beta.

[zicozico]

Hi Jemimah,

Will the upcoming beta4.1 have the save file corruption fixed?
(then I'll switch over to it, else I'll wait)

Furthermore, will hibernation work from /dev/sdb1 (internal SDCARD accessible via USBFLASH)?

Please let me know if I need to test something.

Kind Regards and always appreciating your hard work.

Zicozico