Page 7 of 9

Posted: Fri 14 Dec 2007, 00:34
by PaulBx1
Even when the default connection is set, explicit use of Setup should take you back to the choosing all network options. Anyone else run into this?
Yes, I agree. I can't recall the details about when I ran into trouble, but I think it had to do with wanting to switch from one profile to another, or at least to looking at my profile information (e.g. the WPA password to set a router up). Yet in a working network, you never have access to that page any more.

Or something like that. Again, having trouble remembering the details.

Posted: Fri 14 Dec 2007, 04:14
by RobertB
I read through the thread, and noticed several references to applications taking a long time to start up. That's the problem I'm having, and it's also noted in this thread under "Beginners Help". In that thread, a new user of Puppy Linux started with 3.01, had troubles, and gave up on it. I'm not giving up, but I'm wondering if I need to go back to 2.17.

I did a full HD install of Puppy 3.01 (from a disk directly from Barry, yay!) on a new (to me) computer, freshly repartitioned (4 data partitions and a 1GB swap) and reformatted. It's a Celeron (don't know what vintage) with some 300MB of RAM. But from the beginning, it's behaved strangely.

After it completes the command-line bootup messages, there's a black screen with a mouse pointer for several minutes. You can move the mouse, but nothing happens. Then, a pixel lights up in the bottom corner of the screen. A long time later, the taskbar appears, but no desktop and nothing useful on the taskbar. After an eternity, the desktop finally appears.

Then, it's time to actually run something. A relatively tiny app, the very nice (except for the password thing) PupDial, takes a minute to load up. Opera takes 5 minutes sometimes. Once an application is loaded, it runs fine, but starting an application is a ridiculous chore.

One interesting behavior is that the system seems to figure something out after several boot cycles. Eventually, applications started coming up almost like normal. But then something glitched, and the black screen started staying at bootup. One time, my kids powered it down -- and that caused a file to get corrupted, requiring fsck before the system would boot (endless loop otherwise). After that, it was slow slow slow again, and then gradually seemed to get less slow.

I thought maybe my swap was causing the problem, so I axed it. But bootup is now back to the slowest -- it's like whenever anything about the system changes, bootup and starting applications is super slow.

I'm sure there's something different about my setup, and I'd be happy to dump whatever files might be helpful. It's just sad that the person in the thread I cited installed some other, inferior distro because he had these problems with the latest Puppy version.

Posted: Sun 16 Dec 2007, 19:30
by peppyy
I have been running 3,01, full hd install, on my Thinkpad T-22 for a little over a month now and have managed to work out most of the bugs, usually by changing or adding apps. I finally put a large drive inn and have room to play without having to keep a constant eye on thing.

My biggest problem right now is Mut. I have not been able to determine why it locks up the computer when I use my pcmcia smart media adapter. It has worked in all previous versions and P-mount works fine with it. I will just make the transition to P-mount and hide the desktop icon to avoid this in the future

I have a few minor bugs to work out still, most are with slackware compatibility and missing librarys on non-standard apps. I am very excited to have as many working allowing me to make this a great multi tasking machine.

Wifi wizard counts wrong number of digits

Posted: Tue 18 Dec 2007, 09:01
by nibl
The wifi wizard, inside the network wizard, counts one digit too few. When you enter an 8 digit password it counts 7. So it gives an error that the passw is too short, which is wrong.

I had to reboot into another OS, make the passw 9 digits long and then it worked in Puppy. This bug exists in older versions like 2.14. Odd that nobody noticed yet?

Where is the script for this so I could fix it?

Posted: Tue 18 Dec 2007, 10:51
by tempestuous
nibl,
I will take a guess that you are using one of the Ralink wifi drivers, and Puppy 3-retro?

ESSID truncation was a known problem with Ralink drivers when Puppy moved up to the 2.6.21 kernel for version 2.17.
http://www.murga-linux.com/puppy/viewtopic.php?t=20171

It turns out that Ralink wifi drivers (and maybe some others) are sensitive to the version of Wireless Extensions and Wireless-tools which, in turn, are sensitive to the kernel version.
The problem with Puppy 3.x is that the Wireless-tools were compiled for the 2.6.21.7 kernel ... and these should work fine ... but they are probably incompatible with the "retro" 2.6.18.1 kernel. Therefore, there will be wifi problems in Puppy3-retro.

The solution is not as easy as copying over the Wireless-tools from Puppy version 2.16 (the last one with the 2.6.18.1 kernel) because these will fail to work in Puppy 3.x with error message warning that the Wireless-tools were compiled with a different version of glibc.
The only way to fix this problem would be to recompile the Wireless-tools in Puppy3-retro ... and this may be difficult because the compiler will look for 2.6.18.1 kernel headers, which are not available in the Puppy3 devx package.

Simple answer; use the non-retro version of Puppy 3 for wifi support with Ralink. If the 2.6.21.7 kernel gives other hardware problems, go back to the Puppy 2 series.

Posted: Wed 19 Dec 2007, 13:39
by tempestuous
Oops, the information I just posted was correct, but it concerns errors relating to the ESSID, and I see that you were talking about problems with the passphrase.

The problem of WPA passphrases of 8 characters not being accepted has been reported before -
http://www.murga-linux.com/puppy/viewtopic.php?t=22503
and this is definitely a bug in the Network Wizard. Dougal has just reminded me that he thinks the problem can be fixed by modifying /usr/sbin/wag-profiles.sh at line 925 -

Code: Select all

if [ ${KEY_SIZE} -lt 9 ] || [ ${KEY_SIZE} -gt 64 ] ; then
change the "9" to "8".

It would be good if you could change your router setting back to 8-character WPA passphrase and test this modification. If it works, we will tell Barry.

Posted: Sun 23 Dec 2007, 23:09
by machare
If a pup_save file is encrypted the PC does not stop and ask for a password when it boots. Rather it displays a PASSWORD prompt in blue followed by red text. If any character is entered, and Enter pressed a message is displayed to say that the save file is corrupted and perhaps e2fsck -b 8193 should be run. The strength of the encryption makes no difference.

Further, if there is only one save file present, the PC does not stop during the boot process and give the user the option of loading withou a save file.

Posted: Mon 24 Dec 2007, 02:58
by nibl
tempestuous wrote:Dougal has just reminded me that he thinks the problem can be fixed by modifying /usr/sbin/wag-profiles.sh at line 925 -

Code: Select all

if [ ${KEY_SIZE} -lt 9 ] || [ ${KEY_SIZE} -gt 64 ] ; then
change the "9" to "8".
Yes, that fixes the WPA passphrase bug. Please pass on to Dougal and Barry.

Posted: Wed 26 Dec 2007, 00:41
by PaulBx1
If a pup_save file is encrypted the PC does not stop and ask for a password when it boots. Rather it displays a PASSWORD prompt in blue followed by red text. If any character is entered, and Enter pressed a message is displayed to say that the save file is corrupted and perhaps e2fsck -b 8193 should be run. The strength of the encryption makes no difference.
Try creating an encrypted file from an unencrypted one and tell us if it works any better, using this utility:
http://www.puppyos.net/forum/?1170535852
If it does work, that means there is something wrong with the way Puppy creates an encrypted file in the shutdown process. Disclaimer: I have not tested this utility in Puppy 3.01.
Further, if there is only one save file present, the PC does not stop during the boot process and give the user the option of loading withou a save file.
That's the way it's always been. Think about it, you don't really want to be bothered with this selection if there is only one pupsave. If you want the "0" option you type "puppy pfix=ram" at the beginning of boot.

Laotian instead of Latin American

Posted: Wed 26 Dec 2007, 15:53
by klu9
Bug: Keyboard: Select Latin American but get Laotian
when I select Latin American keyboard in the boot process, I actually get Laotian script when I type. I have to fix it by going into the Keyboard Wizard>Xorg keyboard selector.

Posted: Tue 01 Jan 2008, 13:47
by BarryK
nibl wrote:
tempestuous wrote:Dougal has just reminded me that he thinks the problem can be fixed by modifying /usr/sbin/wag-profiles.sh at line 925 -

Code: Select all

if [ ${KEY_SIZE} -lt 9 ] || [ ${KEY_SIZE} -gt 64 ] ; then
change the "9" to "8".
Yes, that fixes the WPA passphrase bug. Please pass on to Dougal and Barry.
Okay, I've implemented this is 'net_setup-3.94' PET package.

[SOLVED] 3.01 DRI graphics problem with Savage module

Posted: Tue 08 Jan 2008, 14:27
by iscove
I reported an issue with the xorg savage video driver after installing xorg_OPENGL-7.2.pet and xorg_DRI_MODULES-7.2.pet http://ww.murga-linux.com/puppy/viewtopic.php?t=24270

The issue is solved by updating to the xf86-video-savage module v2.1.3 contained in xorg 7.3
http://xorg.freedesktop.org/wiki/Releases/7.3

The attached files savage_drv.la and savage_drv.so were compiled on Puppy 3.01 from the savage driver package posted at http://lists.freedesktop.org/archives/x ... 00351.html

To install, drop them into /usr/X11R7/lib/xorg/modules/drivers/ replacing the previous version of each.

Same problem with Intel video driver

Posted: Tue 08 Jan 2008, 19:33
by rerwin
I have a similar problem with the screen blanking if I do ctrl-alt-bksp in 3.01 using xorg 7.2. Also, on shutdown or reboot the screen goes black (the beam shuts off completely) and remains black during the reboot, necessitating shutting off the PC manually. Puppy 4 has xorg 7.3 which does not exhibit the problem. So I conclude that compiling the xorg 7.3 version of the intel driver would cure things in puppy 3.01. But I have never compiled anything successfully and cannot locate the howto for compiling (if there is one).

Would you point me to instructions for compiling the driver? When I enter "./configure" while in the xf86-video-intel-2.1.1 directory, I get "configure: error: Package requirements...were not met" and "Package 'videoproto', required by 'XvMC', not found" and some suggestions. What did you do?

Thanks for any help.
Richard

Re: Same problem with Intel video driver

Posted: Tue 08 Jan 2008, 21:10
by iscove
Doing this was a painful learning exercise for me.

I encountered the same message, and more after this one was
taken care of. This is the drill:

get the required packages or files from the xorg site, searching from
http://xorg.freedesktop.org/wiki/
if searching there doesn't locate a file, try google for the requested file plus xorg in your search string

e.g. configure stalls and asks for videoproto

download the videoproto package from the xorg site and unzip into a folder

in the console, cd to the folder

the packages are generally processed using the sequence
./autogen.sh
then
./configure
then
make

the needed item may be already generated after just autogen, or after
configure - look in the same folder

place the needed item in the folder generally indicated in the configure
error message

run configure again

solve the next configure issue as above, etc, until configure exits
without an explicit error message.

make will be the final step for generating the driver files
in my case, I had to struggle with an error reported in parsing
/usr/X11R7/include/xorg/xcb.h

it was a syntax error in two lines of xcb.h:
I corrected the line

const xcb_query_extension_reply_t *xcb_get_extension_data(xcb_connection_t *c, xcb_extension_t *ext);

to 2 separate lines:

const xcb_query_extension_reply_t; declaration
int *xcb_get_extension_data(xcb_connection_t *c, xcb_extension_t *ext); prototype

after make has executed without error, you will be looking for
yourdriver.so and yourdriver.la
in my case, they were located in mydriverpkg/src/.libs
note the .libs folder is hidden

USB Wifi doesnt work unless booting from live CD

Posted: Sat 12 Jan 2008, 17:06
by Quadron
Hi all,

I'm new to Puppy linux, and a novice to linux in general. I usualy get a lot done via man pages and searching forums and so on, but now I've found something I really don't get.

When I boot my thinclient hardware via an IDE connectd CF card I can't get my USB wifi adapter to work. Loading modules manualy or using auto load USB, even using modprobe in the console doesn't work.
The oddest thing about it is that when I attacht a cd drive and boot from the live CD, Puppy finds the addaptor and automatically loads modules for it.
I have also tried booting without the adaptor and inserting it later, and running from the live CD everything is ok. Back to the CF and nothing...
I also tried booting from a USB flashdrive with the same result as the CF.
I haven't tried an harddrive yet, but the whole idea was that the sistem would use no moving parts te reduce noise and energy consumption.

Any Ideas anyone?
Other than this issue Puppylinux is great. Keep up the good work!

Cheers,
Maurits

Posted: Wed 23 Jan 2008, 01:19
by crabbypup
sorry, i had posted my bug report in the wrong section. here is a link to it:

http://www.murga-linux.com/puppy/viewtopic.php?t=25633

sorry for that. i'm kinda noob to the forum. but i have been with puppy for a while and really like it alot.

Posted: Wed 23 Jan 2008, 01:27
by floborg
crabbypup wrote:sorry, i had posted my bug report in the wrong section. here is a link to it:

http://www.murga-linux.com/puppy/viewtopic.php?t=25633

sorry for that. i'm kinda noob to the forum. but i have been with puppy for a while and really like it alot.
You actually got responses - seems like that's the better method. I doubt this thread is looked at much since it's become apparent that 3.0x or whatever will never be released.

Posted: Wed 23 Jan 2008, 22:53
by Pizzasgood
I thought Barry was still going to maintain 3.xx, but was just working on 4.xx for the time being.

3.01 Wifi wireless issues - with laptop

Posted: Thu 24 Jan 2008, 18:08
by prit1
Hello,

I have an Averatec 3200 series laptop with an Ralink wireless card (I guess it is 2500). The puppy versions - 2.14 through 2.17 easily recognized the card and showed the network interface as RA0 and I could enter the required details and get connected.

In Puppy version 3.01, the interface is recognized as RA0 and I can enter the details, but there is no connection.

Any help is greatly appreciated.

Thanks!

Posted: Thu 24 Jan 2008, 19:30
by RobertB
I posted earlier about how it would take as long as 5 minutes for applications to start up in 3.01:
RobertB wrote:I read through the thread, and noticed several references to applications taking a long time to start up. That's the problem I'm having, and it's also noted in this thread under "Beginners Help". In that thread, a new user of Puppy Linux started with 3.01, had troubles, and gave up on it. I'm not giving up, but I'm wondering if I need to go back to 2.17.

I did a full HD install of Puppy 3.01 (from a disk directly from Barry, yay!) on a new (to me) computer, freshly repartitioned (4 data partitions and a 1GB swap) and reformatted. It's a Celeron (don't know what vintage) with some 300MB of RAM. But from the beginning, it's behaved strangely.
My ex-wife got a new Windows box (obviously, I had no say in the matter), so I got the memory from her old box. Replacing the 128MB stick with a 512MB stick appears to have solved the problem with applications starting up slowly.

Great for me, but Puppy's claim to fame is bringing back old equipment, even ones with very little RAM. And the programs ran fine in low memory; it was just starting them up that was a problem.

But as was noted before, 3.01 looks to be something of a lost puppy, with all the current development occurring in the 4.x direction. I won't worry about it any more -- just wanted to post the update for future reference.