Next Puppy 5.0.2 29 September 2011

[stu90]

I think (not sure) that JWM still has full screen bug - interesting thing i noticed when reading about the latest version of openbox is you can now have menu icons much like JWM.

in my opinion JWM has NO BUG but YOU, yes, yourself, are responsible for some data necessary to work well (s. at your .jwmrc - I'm sorry to say it, but the puppy linux .jwmrc with under directory is a pollution of JWM! install Debian and experiment in Debian with «JWM only» but you have to uninstall all other window managers and it is in Debian not always really easy...)

possible that it is the same with Window Maker (I did experiment with Window Maker 10 years ago!)

please look also at the size of your window manager:

- JWM for i386 comes with a size UNDER 100 kbyte!
(size of the full Debian package!!!)

in the Basic Linux distribution (2 FD with each only 1450 kbyte )

JWM comes in those 2 old floppy discs of ridiculous size directly with different background pictures!!! JWN manages directly background pictures, fonts, menus, bars (more than 1 and also dynamic hidden bars!) and icons for the menu and can auto-start foreign programs like font upsetting, skype and other goodies!

and is fast!

and as only very little dependencies...

and the user can manage all that in a short text format document always named .jwmrc (logged in /etc/jwmrc or /etc/jwm/jwmrc or ~/.jwmrc depending of the distribution).

Launch Firefox or Seamonkey and press F11 (full screen) and then try to get out of full screen mode.

I never use this short cut and I did try now and I constat the same but Firefox and Seamonkey belong to absolute the same browser's family. How is responsible?

- Did you check the keyboard parameters of your .jwmrc?

- Is JWM or Mozilla responsible (under the point of view that JWM give the ability to influence the short cuts and keyboard comportment? see the last lines of you .jwmrc! they are very important!)

Your prealable answer was: JWM

are you certain it is so and it is not a matter of keyboard management or mozilla?

You can add opera to the list of browsers that doesn't exit full screen as well - this problem has been well documented on the puppy forum, one would imagine if it was a simple case of an edit to the .jwmrc file some one around here would have worked it out and posted a fix for it.

For the benefit of others the JWM version is:
JWM vsvn-500

[Bert]

JWM is a special jewel. But we need to ask Joe what his plans are. After 3 years of silence, he just released a 'new' version, the day before yesterday.

20110925
Given that it's been over four years since the last version of JWM was created, I decided it was time to wrap up all the changes since then into a new version. JWM v2.1.0 is now available! Although development has been kind of slow over the last four years, there are a lot of changes since the previous version

I'm just a Puppy user, wondering if there's still a future for JWM, if taken up by a new Joe or a new JWM working group??

PS It took me a couple of years to understand why Barry stubbornly kept choosing JWM

[ldolse]

Tried Next on some real hardware instead of VMware - found some show-stopper level issues, will work on gathering more data about it tomorrow, but wanted to outline the issues in case the cause was obvious:

• Seems to be hardware specific, my board is an Intel Atom board, the D945GSEJT, graphics controller is the 82945GSE, not sure what kernel module this needs, will provide that later. Previous testing was on VMware, which was working fine.
• The biggest issue is you can't boot Puppy after creating the initial save file. Even though there is only one save file on the boot flash, for some reason Next thinks there are two save files, and pops up the menu to choose one. That would only be a minor annoyance except that the keyboard doesn't work after the reboot, so you can't get past the prompt to select a save file.
• Not quite as bad but still not good is that X won't launch - not sure the cause.
• Least problematic is that the latest realtek drivers aren't in the kernel, need to be compiled - code is here. It's a fairly popular chipset, so would be good to include.

The good news (or maybe it's the bad news), is that it's NOT the remaster scripts. The base woof-built dpup I made exhibits exactly the same problems as Next. So I don't know if the problem is specific to Woof or to the dpup recipe, or to the kernel.

I'll gather more log data, and I'm also going to try a few different Puppy variants to see if I can find a common thread.

I've had this board working perfectly with Puppy 4.2, so hopefully this won't be too difficult to solve.

edit:
Just tried with Pemasus' Dpup Exprimo 5.x.6, and it's exhibiting most of the same problems with this board - the only difference is that the keyboard does work upon reboot - so you are still forced to choose a save file even though only one exists, but at least you are able to choose it. Xwindows errors are exactly the same. I've attached a couple log files and dmesg/ddcprobe output.

edit2:
Just tried Wary 5.1.4.1, and it's working fine - none of the errors I saw for either dpup occurred. I'm thinking that probably eliminates Woof and just leaves the basic dpup recipe or the newer kernel as the potential culprits, but not entirely certain. Note xerrs.log had many of the same errors in Wary, so those may be a red herring. Lsmod shows that the only graphics driver loaded on Wary is intel_agp.

[pemasu]

puppy_next_5.0.0.sfs/lib/modules/2.6.39/kernel/drivers/net/r8169.ko

So...that module is included in Iguleders and in my kernels.
Your problem with that gigabyte ethernet lies somewhere else. Maybe compiling that driver from your link will fix it.
You seem to have failure to load 6 kernel driver modules, according to your logs.

WARNING: Error inserting hwmon (/lib/modules/2.6.39/kernel/drivers/hwmon/hwmon.ko): No such device
WARNING: Error inserting thermal_sys (/lib/modules/2.6.39/kernel/drivers/thermal/thermal_sys.ko): No such device
WARNING: Error inserting video (/lib/modules/2.6.39/kernel/drivers/acpi/video.ko): No such device
WARNING: Error inserting drm (/lib/modules/2.6.39/kernel/drivers/gpu/drm/drm.ko): No such device
WARNING: Error inserting drm_kms_helper (/lib/modules/2.6.39/kernel/drivers/gpu/drm/drm_kms_helper.ko): No such device
FATAL: Error inserting i915 (/lib/modules/2.6.39/kernel/drivers/gpu/drm/i915/i915.ko): No such device

[pacer106]

Here is a pdf I put together with instructions on how to have emelfm2 install .pets with a double click on the .pet.

Just remove the fake .gz from the file.

[ldolse]

puppy_next_5.0.0.sfs/lib/modules/2.6.39/kernel/drivers/net/r8169.ko

So...that module is included in Iguleders and in my kernels.
Your problem with that gigabyte ethernet lies somewhere else. Maybe compiling that driver from your link will fix it.
You seem to have failure to load 6 kernel driver modules, according to your logs.

WARNING: Error inserting hwmon (/lib/modules/2.6.39/kernel/drivers/hwmon/hwmon.ko): No such device
WARNING: Error inserting thermal_sys (/lib/modules/2.6.39/kernel/drivers/thermal/thermal_sys.ko): No such device
WARNING: Error inserting video (/lib/modules/2.6.39/kernel/drivers/acpi/video.ko): No such device
WARNING: Error inserting drm (/lib/modules/2.6.39/kernel/drivers/gpu/drm/drm.ko): No such device
WARNING: Error inserting drm_kms_helper (/lib/modules/2.6.39/kernel/drivers/gpu/drm/drm_kms_helper.ko): No such device
FATAL: Error inserting i915 (/lib/modules/2.6.39/kernel/drivers/gpu/drm/i915/i915.ko): No such device

Those were the errrors that I mentioned occurring with Wary as well - it was late last night my time so I didn't get chance to pull the logs from Wary and add them here. I think there must be some hard coded attempts to load specific modules in the init scripts. I'll upload some of the other log data shortly.

[ldolse]

puppy_next_5.0.0.sfs/lib/modules/2.6.39/kernel/drivers/net/r8169.ko

So...that module is included in Iguleders and in my kernels.
Your problem with that gigabyte ethernet lies somewhere else. Maybe compiling that driver from your link will fix it.

Regarding this module, r8169.ko is not the right module for the current generation of Realtek cards, you need r8168.ko. The PCI ID is the same for both cards, so it will try to load r8169.ko and fail. That said - I think you must have r8168.ko in your kernel modules because 5.X.6 did work in terms of networking. I think it's just Next that's missing this. The source code for this r8168.ko isn't in the main kernel tree, you need to download it separately.

Anyway the networking isn't a concern for me as I know exactly what the problem is in that case - and how to resolve it, just mentioning it for completeness.

[pemasu]

Anyway. Exprimo does not have r8168.ko module. I have only 8169.ko. Interesting if it worked with your eth0.

[ldolse]

Anyway. Exprimo does not have r8168.ko module. I have only 8169.ko. Interesting if it worked with your eth0.

I'll double-check it when I boot back in to Exprimo to get some more logs. Trying a variety of different puplets to see if I can find a common thread, lupu 5.25 also worked flawlessly on the same hardware, and it had loaded the i915 driver as well. Will also try building from woof without stripping any modules and switching out rc.shutdown/rc.sysinit to older versions to see if that helps.

Is there a non-dpup puplet that's using 2.6.39+? That might help isolate the problem.

[pemasu]

Polarpup-005 has 2.6.39.3 kernel. It is Ubuntu Lucid based puplet.

[ldolse]

Making some progress, but still can't reboot.

I just re-built the base pup from Woof, but this time I didn't choose any options Woof offers to strip/remove/optimize anything. I'm not sure if that had anything to do with the xwindows solution, some further reading on the i915 driver offered some further clues.
From the Debian article on Kernel Mode Setting:

945GM and 82865G

• In my system (HP/Compaq nx7300 with 945GM) enabling of KMS when loading the i915 module caused blank screen (seems not to be related to incorrect mode, rather the system got suspended or hibernated) when booting the system. Problem could be cured by changing of /etc/modprobe.d/i915-kms.conf to options i915 modeset=0. What's interesting, booting with acpi=off also cured the problem. The same problem occured in a desktop system with 82865G
• After last (ca. 20th October 2010) update of my debian/testing systems, the xserver-xorg-video-intel stopped to work with i915 modeset set to 0. Therefore the workaround mentioned in the previous point was not applicable any more. Setting of i915 modeset to 1 still caused blank screen, as soon as the i915 module is loaded. After more investigations I've found, that the problem is related to framebuffer console. If KMS is switched on, the framebuffer console support must be selected. In my systems (self compiled kernel without initial ramdisk) I had to to set CONFIG_FRAMEBUFFER_CONSOLE=y, CONFIG_FRAMEBUFFER_CONSOLE_ROTATION=y . Maybe with initial ramdisk, compiling it as module would be sufficient, but I've not checked it. Unfortunately the kernel "make menuconfig" doesn't check if the framebuffer console support is selected when CONFIG_DRM_I915_KMS is set to "y" (LKML, 601558, 600802).
• [Another report:] On my 945GM, CONFIG_FRAMEBUFFER_CONSOLE must be built-in, and not built as a module (this will require building-in AGP / DRM / Intel / Consoleframebuffer) Kernel Bugzilla

I didn't know what those settings meant the first time I ran Woof, and just assumed the default settings were the most conservative. This time I changed the i915 driver option to '0'.

Sounds like some modules may need to be built into the kernel to enable KMS on the i915.

Anyway when I booted this build a lot of things behaved differently:

• Console switched to full monitor resolution immediately after rc.sysinit (before X loaded)
• X loaded successfully
• A error inside of rc.shutdown regarding failing to find a floppy drive that I've seen in many puplets disappeared.

Unfortunately it didn't fix the issue with finding two save files where only one existed and the keyboard failing to work - need to keep digging.

[Iguleder]

Firefox 7 is out; doing a clean build of 24092011, this time with the latest Firefox and other small fixes.

The testing build was good, so if it's stable I'll declare this as the next version.

[ldolse]

I opened a different thread under bugs for the issue with the double savefiles/failing keyboard, as further investigation seems to indicate this is primarily related to Woof and how it handles the latest kernels.

Thread is here:
http://www.murga-linux.com/puppy/viewtopic.php?t=72087

Looks like perhaps I need to modify the initrd-tree to include different kernel modules...

[ldolse]

After further research and some feedback from Pemasu I'm starting to think this problem at least somewhat Next specific. Pemasu inspired me to start hunting through Barry's blog, and I stumbled on this article:

Reason cannot boot USB when no HD

And coincidentally I have no HD in my system... I double-checked the package specs and busybox version, and it turns out Squeeze is using a newer version of Busybox. Based on Barry's blog sometime after 1.4.2 busybox changed it's grep function to be incompatible with Puppy's boot scripts.

Looks like another package that needs to be added to the custom package list for Next.

That probably doesn't solve the keyboard not working problem...

[Iguleder]

Isolate the problem.

Try another kernel, the same kernel on another Puppy, a different X version, etc'. Don't blame a specific package before you have facts to share.

[ldolse]

Working on it, apologize for jumping the gun, got excited when I read the post, but need to test it and verify.

[Iguleder]

Next Puppy 29092011 is being built right now, with two important changes:
- A dhcpcd bug fix - for some reason the build script got messed up so some files were missing
- Transmission 1.93 - the 2.x series is much bigger and heavier, while 1.9x is lighter and small

[darkcity]

added to big Puppy list
http://puppylinux.org/wikka/PuppyVersion

[Iguleder]

Thanks, darkcity!

____

Uploading Next Puppy 5.0.2 29092011, with all the new features since alpha 3. It's simply amazing how much better it is

EDIT: still uploading, it's extremely slow ...

I decided to make small Abiword and Gnumeric packages, too - I want to see how small I can get them.

Abiword's core dependencies (wv and libgsf) are ready and extremely tiny - now I need to get Abiword to compile. Gnumeric (and goffice, its dependency) will follow, hopefully

I'm doing this in protest against ICPUG's comment in Barry's blog, which claims Next Puppy isn't aimed at content creation because it lacks these. This is incorrect, because things like Google Docs exist (and there are more reasons why they're out, like their limited usefulness in a world dominated by LibreOffice).

[stu90]

I don't recall ever using Abiword and Gnumeric.