4.1 Alpha 5

[exProphecy]

Tried 4.05 and everything that I commonly use on my Averatec 3200 is working well.
- Network settings are saved and network connection is automatic on every reboot.
- Gxine worked for my favorite online video channels.
- Abiword worked.
- PPLOG worked.
- Gnumeric worked.
- Saves few seconds during boot compared to earlier versions.

The only thing I notice (as always with Puppy > 4 using this kernel) is that the network (RT2500 module) is slower. In Dingo its much faster. This is puzzling and may prevent me from moving to these versions.

I just had a question about your network connection. I'm assuming it's wireless. Is your wifi turned on automatically when you boot Puppy? Or do you boot up Puppy THEN turn it on with a switch or plug it in etc?

[Caneri]

Hi Adam aka pigshed,

Hooray!...that did the trick. I opened the ".pmountengine" and deleted the entry for mut2 and voila pmount is back...thanks.

Now the only problem is pmount doesn't show swap partitions..grrr..lol. I'll just use another install to disable the swap I want to get rid of. Thanks again.

I always liked mut because it allows a swapon/swapoff in the gui....but console will also do the trick.

Best,
Eric

[Obi]

A Newbie question: is there a need on feedback or errata (to avoid posting the same stuff) for Puppy Alpha 4.04. I am currently online with it with no crashes so far ( I keep my fingers crossed:). Anyway I would like to contribute if possible for testing, but I do not currently have a burning CD-drive to burn 4.05 ISO.

[linuxcbon]

Obi now it's time to test alpha5.

[paulh177]

Lin'n'Win frugal.

Installed mplayer from the Pet manager.
Every time I run it two blank windows open and X locks solid -- no mouse or keyboard.
ctrl-alt-bkspc to kill X then restart gets X going again OK.

paul

[Caneri]

Hi All,

I may have a problem with psip in alpha5....if the other testers see that the gizmo operator is not able to be heard maybe a full test run is needed inside alpha5.

Smokey and I had a problem but one call does not make a good test run.

This may also be my local ISP problem..so please give psip a test run from various countries to see if it's just me or there is another issue.

Thanks,
Eric

[rerwin]

In attempting a frugal install on my TP i1720, the installer reported a successful installation, but the three Puppy files had zero length. Running from the command line showed each file encountered an I/O error on the CD(RW):

# puppyinstaller
sda "ATA TOSHIBA MK4313MA, size 4.030 GiB"
sda
ls: cannot access /mnt/sda1/pup_[0-9][0-9][0-9].sfs: No such file or directory
sda1: ntfs, size 3.907 GiB
ls: cannot access /mnt/sda1/pup_[0-9][0-9][0-9].sfs: No such file or directory
cp: reading `/mnt/sr0/vmlinuz': Input/output error
cp: reading `/mnt/sr0/initrd.gz': Input/output error
cp: reading `/mnt/sr0/pup_405.sfs': Input/output error
/usr/sbin/puppyinstaller: line 869: 30532 Terminated yaf-splash -font "8x16" -outline 0 -margin 4 -bg orange -text "Please wait, copying files to $DESTPART, in folder $NEWPSUBDIR..."
#

The messages log contained:

Jul 30 22:35:18 (none) user.notice kernel: scsi 1:0:0:0: CD-ROM SANYO CRD-S372B 1.24 PQ: 0 ANSI: 5
Jul 30 22:35:18 (none) user.warn kernel: sr0: scsi3-mmc drive: 8x/24x cd/rw xa/form2 cdda tray
Jul 30 22:35:18 (none) user.info kernel: Uniform CD-ROM driver Revision: 3.20
Jul 30 22:35:18 (none) user.debug kernel: sr 1:0:0:0: Attached scsi CD-ROM sr0
. . .
Jul 30 14:42:07 (none) user.info kernel: sr 1:0:0:0: [sr0] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE,SUGGEST_OK
Jul 30 14:42:07 (none) user.info kernel: sr 1:0:0:0: [sr0] Sense Key : Medium Error [current]
Jul 30 14:42:07 (none) user.info kernel: sr 1:0:0:0: [sr0] Add. Sense: No seek complete
Jul 30 14:42:07 (none) user.err kernel: end_request: I/O error, dev sr0, sector 183552
Jul 30 14:42:07 (none) user.err kernel: Buffer I/O error on device sr0, logical block 45888
Jul 30 14:42:07 (none) user.err kernel: Buffer I/O error on device sr0, logical block 45889
. . .
Jul 30 14:42:09 (none) user.err kernel: Buffer I/O error on device sr0, logical block 45888
Jul 30 14:42:12 (none) user.info kernel: sr 1:0:0:0: [sr0] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE,SUGGEST_OK
Jul 30 14:42:12 (none) user.info kernel: sr 1:0:0:0: [sr0] Sense Key : Medium Error [current]
Jul 30 14:42:12 (none) user.info kernel: sr 1:0:0:0: [sr0] Add. Sense: No seek complete
Jul 30 14:42:12 (none) user.err kernel: end_request: I/O error, dev sr0, sector 183556
Jul 30 14:42:12 (none) user.err kernel: Buffer I/O error on device sr0, logical block 45889
Jul 30 14:42:14 (none) user.info kernel: sr 1:0:0:0: [sr0] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE,SUGGEST_OK
. . . (and more of the same)

I subsequently installed manually onto the TP, copying the files from the same CD. On another PC, the install succeeded. Go figure. (I doubt it is relevant, but the install was to an NTFS partition on the laptop, ext2 on the PC.)

Although the CDROM drive is definitely suspect, I feel that the installer should detect the problem and inform the user appropriately -- at least that the install failed (and possibly clean up the "mess"). Otherwise, this behavior undermines Puppy's credibility.
Richard

Observation: While using a diskette to transfer the evidence, I noticed that the floppy drive icon never showed it as umounted, whether I right-clicked it and unmounted, or used Pmount. Pmount also showed it as mounted, until I clicked the "unmnt" button, but the icon retained its green ("mounted") dot.

[dogone]

Petget in both alpha 4 and 5 appears to have difficulty connecting to source sites and downloading. It appears as if the sites are not responding but very often none do. I do not recall any such problem using Petget under v4.0 and earlier Puppies.

Looking at /usr/sbin/petget, I see that Barry's been experimenting with alternate site addresses. I tried configuring Petget to use those commented out. Petget is then better able to connect to the site but downloads often fail shortly after they begin.

My net connection here in Arizona is working perfectly otherwise so something very fishy is going on,. Anyone having similar problems?

[Fishy]

I have just put together a new P5K ASUS based system with a Pentium dual core processor. I have tried Alpha4 and now 5 with the install stopping at the same point i.e. kernel mount.

puppy pfix=ram was used as well as letting it try the NOP 4.0 pup save file.

Dingo 4.0 DVD was tried and installed, puppy pfix=ram so it must be something different in the Alpha. Yes I know that there is a lot different in this new version but as somebody who uses Puppy because "it works" I'd like to help out. Sorry, but this is the only testing that I can do.

Cheers

[KJ]

@ Caneri

For extended drive information, you might try zigbert's Pdrive 0.27 discussed in the following thread.

http://www.murga-linux.com/puppy/viewtopic.php?t=21200

I just tested it with pup405 and it works great but I don't have a swap partition available ATT. I did use it with pup404 on two different kits a couple days ago and all drives were IDed including the swap.

KJ

[KJ]

'Set global font size' doesn't work for me in pup404 and now in pup405. Are others have the same problem? My tired eyes need some extra font size these days.

KJ

[BarryK]

Lobster,
In your 'number one' computer which has non-functional USB mouse, does it accept a PS/2 mouse? If so, can you do some tests?...

1.
# dmesg
run this in a terminal, can you see anything that looks like an error message to do with USB? No need to post the out here.

2.
/tmp/bootsysinit.log
Open that in Geany, again, anything that looks like an error message about USB?

3.
Do other devices work in the same USB socket, like a Flash pen drive? Does a USB keyboard work?

4.
Open the BootManager (System menu), choose "Click here to blacklist a module", and move 'usbkbd' and 'usbmouse' from right to left pane, then move 'usbhid' from left to right pane. Then reboot, then test your USB mouse. (this of course needs a pup_save file, so the change is remembered)

[rerwin]

Regarding:

wvdialconf bug workaround
Rerwin found another strange bug when using 'wvdialconf' to probe for modems. The 'Modem' entry in /etc/wvdial.conf can get set to '/dev/modem', but /dev/modem is a symlink to 'modem', that is, to itself. Peculiar situation.

When I saw this item, I presumed Barry used my solution to wvdialconf's setting of "/dev/modem" in the wvdial config file. I did not view it as a bug. But what modemprobe does with that information is where the bug exists.

Placing /dev/modem in wvdial.conf is quite legitimate for wvdial's operation. Avoiding its use is apparently a Puppy design decision. Modemprobe assumes that. But that results in making /dev/modem a softlink to itself. (EDIT:) 4A5 does avoid (/EDIT) doing that if the config entry is "/dev/modem", since what it already points to is what was probed successfully!

(EDIT:) As indicated in the modemprobe commentary, making the creation of the softlink conditional avoids the need to protect Puppy in rc.sysinit (/EDIT) from he "peculiar situation", and fulfills the (presumed) intent of the author to always probe the current device (as indicated by /dev/modem) before any others. I infer this intent from the messages issued by wvdialconf:

Port Scan<*1>: Scanning ttySM0 first, /dev/modem is a link to it.
ttySM0<Info>: No such device or address
Port Scan<*1>: SM0
ttyS0<*1>: ATQ0 V1 E1 -- failed with 2400 baud, next try: 4800 baud
. . .

Coincidentally, I found some other clues.
1. Part of a 2006 posting by linmodem guru Marvin Stodolsky:

wvdialconf only looks at /dev/ttyS* ports, and in some releases
follows a /dev/modem symbolic link.

We now see that it appears to take anything tty*.
2. Part of a 2003 posting by (the same) MarvS regarding the intel modem device /dev/536ep, part of a solution contains:

# ln -sf /dev/536ep /dev/modem
Install the package wvdial and run
# wvdialconf

implying that /dev/modem would be used to get to /dev536ep, which would not otherwise be probed.

This tells me that we should just work with it and not fight it, as 4.0.5 appears to do. It is really a "feature" to test the current device first.
EDIT: However, the other part of the workaround is "what to do with the 'modem =' line in wvdial.conf."

If wvdialconf sets "/dev/modem" in that line, modemprobe restores it to its earlier state, which suffices if the modem has not been changed out. If the modem is changed, the user is instructed to click the "Erase" button and reboot. The new modem is detected and connects, but probing it results in "no modem", while the log shows the new one as being detected. The solution is to, instead, update that line to the target of /dev/modem.

Since I wrote the initial post, I have successfully tested 3 modems: ess, pctel & intel536. For the latter I had to force loading of its driver. The others were probed through /dev/modem, but the Intel was not detected until I added a "ttyS" link (ttyS_536ep); even then it was not detected through /dev/modem, but by using that link. So the risk is that probing the Intel modems might "lose" them, even though the modem is functional.
/EDIT
Thanks.
Richard

[eprv]

Barry, xfs is detected after modprobe xfs, here is the output

# probepart
/dev/sda1|xfs|8209152
/dev/sdb1|swap|1012032
/dev/sdc1|vfat|995966

here is the pmountdlg file:

<window title="Pmount Puppy Drive Mounter" icon-name="gtk-harddisk">
<vbox>
<text use-markup="true">
<label>"<b>DO NOT REMOVE MOUNTED MEDIA</b>"</label></text>

<notebook labels="floppy|usbdrv">
<vbox><frame ATA HYPERSTONE FLASH></frame><frame ATA TRANSCEND ><hbox><text><label>sda1 xfs 3.9G</label></text><button><label>MOUNT</label><action type="exit">DOMOUNT_/dev/sda1_xfs</action></button></hbox></frame></vbox><vbox><frame SanDisk Cruzer Mini ><hbox><button><input file>/usr/local/lib/X11/mini-icons/mini-filemgr.xpm</input><action>rox -d /mnt/home &</action></button><text><label>sdc1 vfat 486M</label></text><text><label>330M free</label></text><button><label>UNMNT</label><visible>disabled</visible><action type="exit">UNMOUNT_/dev/sdc1_vfat_/mnt/home</action></button></hbox></frame></vbox>
</notebook>

<hbox>

<button><label>preferences</label>
<action type="exit">SETPREFS</action>
</button>
<button><input file stock="gtk-refresh"></input><label>REFRESH</label>
<action type="exit">REFRESH</action>
</button>
<button><input file stock="gtk-quit"></input><label>QUIT</label></button>
</hbox>
</vbox>
</window>

[CHLee]

1. The Freememapplet disappear after frugal install(have pup_save.2fs), it have worked when pfix=ram.
2. pmedia=ideflash and pmedia=ataflash still no work.
3. The atl2 Ethernet driver was well work for me, thanks!
Thank Barry very much! CHLee

[Lobster]

Lobster,
In your 'number one' computer which has non-functional USB mouse, does it accept a PS/2 mouse? If so, can you do some tests?...

No ps2 does not work - well 5 - 10 seconds and then freezes
I also tried USB similar problem
The disconcerting thing is after the mouse goes
the keyboard freezes (usually within 2 minutes)
and if I press ctrl + alt + backspace in time to get a prompt
even the prompt freezes

This is why I suspect the Linux kernel
However no one else is experiencing or reporting this
So not sure how best to test
maybe . . . if I go directly into command line puppy from boot options
never done that - I believe it is an option?

I can try a serial mouse (again) as I have not attempted that with Alpha 5

m m m . . . let me try some of that stuff at least . . .

OK - update:

USB mouse is working - sort of - basically it does not respond very well - you have to move it slowly - all in one movement
and very difficult - I seem to remember this problem before
and it was a question of assigning different numbers to the mouse sensitivity - somehow?
Keyboard works for a lot longer 10-15 minutes before freezing

I was able to use the serial mouse
sometimes for only 5-10 seconds
but restarting xwin and it then worked for 2-5 minutes
then usually back to 5-10 seconds

update 2:
Now that I have a badly behaved but working USB mouse
I will do some research on sensitivity and see if changing the settings
alters things. Also will do some of the tests Barry recommends.
Does anyone know where mouse configuration info is kept in Puppy?
What are your settings if using a working USB mouse?

[John Doe]

I've also had some random catastrophic freezes with various 4.1 alphas where it would have been nice to log kernel messages to a separate drive/partition.

Can anyone give some thoughts on how to do such error logging?

[pigshed]

Re-boot works fine. Shutdown causes machine to stop at "system halted" as with other 4.1's. This doesn't happen on any other puppy versions/puppys I've tried. (Panasonic CF-71 laptop, PIII).

[Sage]

There are always HW & FW (BIOS) idiosyncrasies. Apart from swapping around the HW, there are two alternative approaches that sometimes assist diagnosis.
There are utilities (mainly DOS/'doze) that permit alteration of the entire range of BIOS functions, ie not just those provided by AMI/Award/Phoenix at the DEL function. Care is needed when adjusting some of these, although mostly complete reset is possible by jumper (test first with an innocuous parameter!).
Alternatively, it sometimes helps to install one of the distros that still uses the full complement of options via the old and loathed (Slack) textual installer, eg Absolute or an heavily customised installer like VASM in Vector. In this way, a wide range of, esp., mouse drivers can be tested.
Notwithstanding, remember that some Logitech mice behave in utterly atypical ways except when their proprietary drivers are loaded into 'doze. Experience suggests that the cheapest mice are most reliable, notable Genius.
There is no excuse for not trying half-a-dozen different PS/2 mice which can be liberated from the local tip. I discovered some companies do not allow their secretaries to open up mice to clean the rollers "in case they get electrocuted" !! Oh yes, ignorance amongst senior executives abounds, at least on this side of the Atlantic, where progress towards the boardroom is still mediated by father's influence and Oxbridge college. But then, fear of the H&SE and litigation is equally prevalent - after all, companies might face a £m legal bill if an employee breaks a fingernail when scraping the rollers....

[boscobearbank]

Desktop (connecting wirelessly via my cheap Atheros-based wireless card that chokes on ath5k but works fine with ath_pci) and my laptop (connecting wirelessly via the rt2500pci module) both seem very pleased with alpha 5.

My complements to the chef and his helpers