(old)Puppy Linux Discussion Forum

When I tested again in the eee, poweroff and reboot don't work, even in console. Halt freezes it. Could it be an X issue?

Another trick to add could be the activation of battery monitor when it's a laptop (or one of the baby laptops).

Hmm, ok. If you create a DRI PET package for Puppy, then it will need to overwrite the Xorg files currently in Dingo that are DRI-challenged.
I will need to look at those .conf and .patch files next time I rebuild with T2.

It must just be one or more of the libraries. I made a pet package of the Xorg stuff missing from dingo alpha3 using the Xorg I compiled and it works. It's here If anyone wants to try it:

http://myfreefilehosting.com/f/edf6fef129_20.04MB

After installing the pet package, if your testing with alpha3, the DRM kernel modules don't get loaded from the zdrv file automatically. I think I saw where Barry has fixed that. Anyway in Alpha3 you'll have to copy two files, the drm.ko file and the module for which ever card you have from the zdrv_393.sfs /lib/modules/2.6.24-rc4/kernel/drivers/char/drm to the same place in your file system. Then exit X and type:

# depmod -a
# xorgwizard

After restarting X, you can see if DRI is working by opening a terminal and typing:

# glxgears -info

You should see some gears runing in a box and you should see reported "GL_RENDERER = Mesa DRI ..."

Note: When I built with T2 I specified i686, so if you have something older than a Pentium 2 it might not work. If you have something older that a Pentium 2, you probably don't need 3D rendering anyway

Also, the devx file has an incomplete collection of Xorg include files. I posted the Xorg include files here:

http://myfreefilehosting.com/f/50d0e37588_1.28MB

capoverde wrote:But here too, after unmounting one SD card and inserting & mounting another, Rox shows false contents. However, with 3.01, just restarting the X server corrects this (same behavior from either PMount or MUT - could not try with MUT in Dingo).

There's a problem with Rox, in all puppies. You have to click the 'refresh' button in Rox (the two arrows in a circle).

kirk wrote:

Also, the devx file has an incomplete collection of Xorg include files.

That's deliberate, if Puppy is built with Xvesa only then apps will not have dependencies that are not there.

I'm thinking maybe it's time to move on from compiling for 'i486'. I thought that 'i686' is generic, for Pentium One also?

raffy wrote:When I tested again in the eee, poweroff and reboot don't work, even in console. Halt freezes it. Could it be an X issue?

Another trick to add could be the activation of battery monitor when it's a laptop (or one of the baby laptops).

Did it happen just once?
I have had lockup at shutdown three times so far, on my laptop. I'm using Dingo everyday, so it's a very occasional thing. It may be a conflict with the SMP kernel and Unionfs, as the Unionfs mail list mentions a possible problem that may cause hanging. They have fixed it in the most recent release.

BarryK wrote:I'm thinking maybe it's time to move on from compiling for 'i486'. I thought that 'i686' is generic, for Pentium One also?

I've been trying to get a handle on CPU optimizations for years now, and remain confused.
But I just found 2 useful links on the subject

http://linuxreviews.org/howtos/compiling/safe-cflags/

Replace -march with -mcpu to make the binary optimized for the cpu yet usable on i386.

and
http://gcc.gnu.org/ml/gcc/2005-02/msg00476.html

This is what the GCC manual says ...
-mcpu=cpu-type
Tune to cpu-type everything applicable about the generated code, except
for the ABI and the set of available instructions. The choices for cpu-type
are i386, i486, i586, i686, pentium, pentium-mmx, pentiumpro, pentium2,
pentium3, pentium4, prescott, nocona, k6, k6-2, k6-3, athlon, athlon-tbird,
athlon-4, athlon-xp, athlon-mp, winchip-c6, winchip2 and c3.

While picking a specific cpu-type will schedule things appropriately for
that particular chip, the compiler will not generate any code that does not
run on the i386 without the -march=cpu-type option being used. ***i586 is
equivalent to pentium and i686 is equivalent to pentiumpro***. k6 and athlon
are the AMD chips as opposed to the Intel ones.

-march=cpu-type
Generate instructions for the machine type cpu-type. The choices for
cpu-type are the same as for -mcpu. Moreover, specifying -march=cpu-type
implies -mcpu=cpu-type.

Some significant points I see:
"i586" and "pentium" are identical.
"i686" and "pentiumpro" are identical.
(PentiumPro was the first in a range of processors using the i686 instruction set)
http://en.wikipedia.org/wiki/I686

So "-mcpu=i686" would be OK for Pentium/PentiumMMX hardware.
but "-march=i686" would NOT be OK for Pentium/PentiumMMX hardware.

I though i686 was Pentium2, oh well. Of course there was never really a i586 or i686, they're all Pentium somethings. I wonder how many people are using Pentium 1s? I would think you could find something better in the trash. I have.

Noticed Dingo doesn't have a FTP server. I was having problems with pure-ftpd in 3.01 on my network. Turns out it needed the -H switch. Anyway I compiled Pure-ftpd in Dingo and made a nice simple script to launch it. Dingo would need to have a user named "ftp" with no password. And a home directory /root/ftp or some other place if you want to edit the script.

Also, Brad_chuck made a real nice script to download RSS feeds (podcasts). I made a couple of small additions to it. It's quite small only 25k compressed. It's here:

http://www.murga-linux.com/puppy/viewto ... 649#164649

[quote"BarryK"]There's a problem with Rox, in all puppies. You have to click the 'refresh' button in Rox (the two arrows in a circle).[/quote]

Why, sure! Pity me, I had missed this one; I did "refresh"... but in Pmount's or MUT's window. Refreshing in Rox, SD cards get read OK -- in Dingo Alpha3 & 4 as well.

Seeing that, I tried calling Rox from the console with the -x option, and it worked -- that is, it refreshed the directory. Knowing my low skills in programming, I assume calling Rox that way from PMount or MUT is too obvious and trivial a solution to be good?

Other stuff:
- NetSurf seems slower in Dingo than in Puppy 3.01 - and quite a bit slower than Dillo in previous Puppies. Will try and take precise test times.

- About PostScript-related issues: I frequently use a tiny music notation program (abcm2ps, it's in Sourceforge) to convert a coded text file into a perfect musical score. Of course, in order to see the result one needs a PS viewer: in this respect Ghostview was better than EPdf or other PDF/PS viewers, as it has a "file-watching" option (though requiring manual activation in Puppy).

Lack of this feature requires closing and reopening the output file at every change -- quite annoying and time-consuming. No other file-watch-capable PS viewer around? Or maybe a quick-and-dirty shell script to get the same effect?

Seeing that, I tried calling Rox from the console with the -x option, and it worked -- that is, it refreshed the directory. Knowing my low skills in programming, I assume calling Rox that way from PMount or MUT is too obvious and trivial a solution to be good?

I somehow overlooked the '-x' option! I think when I saw the docs, it stated that -x is for rescanning a "file", so I didn't think any further on it. But, you're saying it works for directories -- well '-x' is going into Pmount right now!

This seems to work:
rox -x /mnt/sdb1 -d /mnt/sdb1

Wonderful, Rox won't have that problem in Puppy any more. Thanks for putting my two cents to good work!

(As for NetSurf's speed, it is actually normal on stored HTML pages; I must have hit a heavy-load moment on the remote server in my first test).

Just downloaded & mastered Dingo Alpha4 SeaMonkey.

This time mhwaveedit hung on the very first attempt to load an mp3 file (Xine played it perfectly a moment before), with high CPU load; the window could be closed with the "kill" command from the window bar menu, but the load on the CPU wouldn't drop.

After rebooting, mhwaveedit decoded and played an OGG Vorbis file perfectly, but hung again on an mp3.

If it were a problem with the Lame decoder, Xine should also suffer from it, so it's probably a bug in mhwaveedit -- a pity, it's a nice and fast program.

Also, the bell offered by the GTK Xset can't be heard though the audio works OK -- it's the first time I try it and actually don't need it, but was just curious.

tempestuous wrote:

Replace -march with -mcpu to make the binary optimized for the cpu yet usable on i386.

Ha! I always wondered about this -- I was afraid compiling for 686 would make it use MMX/SSD and thus be incompatible with AMD processors...

Barry, I've tested the shutting down of alpha4 twice on the eee, always failing - no poweroff or reboot working.

Now, with the Jan 5 version, it can't find pup_394.sfs, either in the SD card or in partition 2 of the internal drive. I will test again using USB keydrive.

Note that eee's internal storage is seen in Puppy 2.13 as /dev/hdc and the SD card as /dev/sda.

Alpha4 has no firewall problem anymore -- for me at least.

Also tested it on an old K6-II 400Mhz -- as Kirk says, better stuff can be easily found in the trash nowadays, but it gives an idea of how well Puppy does on old hardware -- running fine in general; boot time is impressively short even here.
Of course mhwaveedit has the usual problem opening mp3s; surfing with GTKMoz (wrongly referred to as NetSurf in my former posts, sorry) is surprisingly smooth and fast -- seemed even better than with Alpha3 on the Celeron 2800: obviously there were server problems at the test time.

Seems a really very performing Puppy.

There is a later version of mhwaveedit, I'll upgrade for alpha5.

raffy wrote:Barry, I've tested the shutting down of alpha4 twice on the eee, always failing - no poweroff or reboot working.

Now, with the Jan 5 version...

Will someone please post a link to the Jan 5 version? I haven't seen it and I'm getting confused. I don't know which thread to post my bugs to.

Well, the 'jan 5' version has some unfornate new bugs, that I have now fixed. So wait for alpha5 that I'll announce on the blog soon.

In the blog comments I usually post a link to a forum thread for further discussion.

This is a very marginal issue, and was present in previous Puppies too, but can become annoying:

if for some reason the mouse isn't working, in some instances it's not easy to select menu items from the keyboard (TAB, arrows, space and Enter) since the active item is not always clearly highlighted.

E.g., try changing the screen resolution this way in Dingo's XVesa wizard window: the chosen value can be clearly seen, as a dot appears in the corresponding button, but then neither the <CHANGE> nor the <OK> button get highlighted or focused to show which is the active one, so the job becomes a hit-or-miss game or guesswork (only the script name inside the text box goes in and out of focus by pressing TAB repeatedly).

don't know where to post the dingo 5 stuff.. so I do it here.

Issue with external PCMCIA CD-Rom Drive (sony vaio pIII 64MB RAM, subnotebook):

Until 1.07, the live CD booted with no problem
Until 3.01, and dingo alpha 2 the CD booted with boot-parameter "ide1=0x180"

Now it doesn't boot at all. The error is, that after loading the initrd (this still works), the system can't find the CD ROM...

Is there a new parameter available - or could be the PCMCIA stuff is missing in the kernel....?

Please help