EasyOS version 2.3.2, June 22, 2020

[Sage]

Many thanks for prompt response, Barry.
Sound: scsijon may not be correct. For the last few versions, the default sound has been set at 60%. This is far too low to hear the feeble output on diddy-little speakers in many laptops and certainly inbuilt speakers in monitors; better set @90% and turn it down if working? It is really helpful when testing kit to know whether sound is working at first boot, not least because sound settings cause so much angst amongst less well-informed users!
.isos always well appreciated. (Ha! my bad - didn't expect to look under the "legacy" listing. This is going to help at least one of my present colleagues trying to migrate but having fairly ancient kit. Will also save some unused CD-Rs languishing beside my desk from even more landfill. ).
As for Qsync, I like to have it running before completing the initial setup and why slow running of SM causes me grief - it's the first thing I want to check on defective machines, especially from rogue operators! And there' a lot of those, defecting (or wishing to...) from the devil's OS.

[Terry H]

@scsijon thanks for the response. Screen capture using screeny with the browser in Full screen mode is also distorted. Screen resolution is 1366 X 768. No problem for the very few screen captures I may do, i'll use TAS.

Interesting, it's not happening for me in pyro 1.2.7 with Screeny in both the inbuilt Seamonkey and installed Firefox (69.0.3). Is there anything in /tmp/xerrs.log after a capture that could help track it down? you could also try reducing your browser to 1024x768 (a standard size) and try capturing that, maybe there is a driver problem in widescreen.

Sorry, just saw this, I've been watching election results. It's late here now. I'll check some more tomorrow.

I checked xerrs.log, which gave me nothing to go on. I have no idea what the diff output shown below means. At present I don't consider it worth pursuing further.

Code: Select all

# diff /home/personal/xerrs.log /tmp/xerrs.log
61a62,63
> xwdtopnm: writing PPM file
> EXIT="Exit on timeout"
#
 

[illiapoplawski]

Running Pyro 1.2.5 had no issues however after upgrading to 1.2.7 the error in the attached image is given (written below).

"expr: warning: '^/dev/tty[0-90\+$': using '^' as the first character of a basic regular expression is not portable; it is ignored
/usr/bin/startx: line 169: /usr/bin/mcookie: No such file or directory
Couldn't create cookie"

I created a clean usb install and error persisted then reverted back to a clean pyro 1.2.5 and graphics work again. Running on an acer aspire e1-522-3884 with AMD Radeon HD 8240 graphics

[BarryK]

I notice that the Petget Package Manager ?isn't downloading to the right spot. It says it's downloading to /root, shouldn't it now be downloading to the $/file/downloads, especially when your working in a container. It use to download to /root/Downloads.

Some apps, such as SeaMonkey, default to download to /home/downloads

The PPM though, most people would choose to install packages, not just download them. When installed, they are kept in /audit/packages

Note, there is no /file folder, that was for Quirky only.

[BarryK]

In attempting to get a screen capture of the flashing icon change mentioned in previous post I attempted to get screenshots using screeny. I set to capture full screen to show the tray. When I viewed the output the images, they are all tilted and shown in monochrome. Screen capture of window only is OK, it is in colour and not distorted.

Dell laptop Core i5- 5200U processor.

Terry,

Try again with the browser set in full window size, it should be ok then, I wrote about this fix earlier but for referance, it seems that screeny doesn't like odd sizes, if you set it to one of the video drivers resolution matrix's it usually saves correctly, it's just if you try a non-matrix it fails. ie 1280x1024/960, 1680x1050, 1024x768 should be ok, setting it for a part screen will fail unless it's an acceptable resolution to your video card.

There was a permanent fix posted to the Puppy forum awhile back, but I haven't chased it up. Maybe I should just remove Screeny, as Take a Shot does a pretty good job.

Just did a search of the forum, yes, the distortion problem was identified a couple of years ago, and fixed:

http://murga-linux.com/puppy/viewtopic. ... 4&start=45

It is caused by the 'xwdtopnm' utility in the 'netpbm' package.

Quirky and Easy use a very old version of netpbm, as later versions were found to break something else, but I don't recall what.

OK, I will test the latest netpbm, and look at all scripts that use the netpbm utilities, see if any of them become broken.

If yes, then I could stay with the old netpbm and just use the latest xwdtopnm (with the libnetpbm library linked statically).

[BarryK]

Running Pyro 1.2.5 had no issues however after upgrading to 1.2.7 the error in the attached image is given (written below).

"expr: warning: '^/dev/tty[0-90\+$': using '^' as the first character of a basic regular expression is not portable; it is ignored
/usr/bin/startx: line 169: /usr/bin/mcookie: No such file or directory
Couldn't create cookie"

I created a clean usb install and error persisted then reverted back to a clean pyro 1.2.5 and graphics work again. Running on an acer aspire e1-522-3884 with AMD Radeon HD 8240 graphics

That's interesting. Start X by running 'xwin', not 'startx':

Code: Select all

# xwin

It is interesting because back in the Quirky and Puppy days I had /usr/bin/startx as a symlink to xwin. However, now it is a script that some package has installed. Right, I need to get that restored to a symlink.

If X will not start and drops you back to the commandline, look in /var/log/Xorg.0.log see if anything obvious in there.

[illiapoplawski]

Running Pyro 1.2.5 had no issues however after upgrading to 1.2.7 the error in the attached image is given (written below).

"expr: warning: '^/dev/tty[0-90\+$': using '^' as the first character of a basic regular expression is not portable; it is ignored
/usr/bin/startx: line 169: /usr/bin/mcookie: No such file or directory
Couldn't create cookie"

I created a clean usb install and error persisted then reverted back to a clean pyro 1.2.5 and graphics work again. Running on an acer aspire e1-522-3884 with AMD Radeon HD 8240 graphics

That's interesting. Start X by running 'xwin', not 'startx':

Code: Select all

# xwin

It is interesting because back in the Quirky and Puppy days I had /usr/bin/startx as a symlink to xwin. However, now it is a script that some package has installed. Right, I need to get that restored to a symlink.

If X will not start and drops you back to the commandline, look in /var/log/Xorg.0.log see if anything obvious in there.

Tried starting with xwin and it blocks on a non blinking white cursor on a black screen that i can't move away from except for force restarting the laptop. When going into /var/log after restarting there are no Xorg logs to be found to help find the source of the issue. Tried with Buster 2.1.7 and i have no issues with graphics.

[gcav]

Hi all.. playing around with dosbox under EasyBuster from both downloaded and compiled dosboxes I get the following:

[/home]$ dosbox
DOSBox version 0.74-3
Copyright 2002-2019 DOSBox Team, published under GNU GPL.
---
**
GLib:ERROR:../../../glib/ghash.c:377:g_hash_table_lookup_node: assertion failed: (!g_atomic_ref_count_compare (&hash_table->ref_count, 0))
Aborted (core dumped)


Any hints before going down the rabbit hole?..

gcav

ps: as someone mentioned before the battery tray app is not updating.


--stack trace for the curious---
GNU DDD 3.3.12 (x86_64-pc-linux-gnu), by Dorothea LReading symbols from /usr/local/bin/dosbox...done.
[New LWP 7075]
[New LWP 7074]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Core was generated by `dosbox'.
Program terminated with signal SIGABRT, Aborted.
#0 0x00007fe8a5f127bb in raise () from /lib/libc.so.6
[Current thread is 1 (Thread 0x7fe8a45cd700 (LWP 7075))]
(gdb) where
#0 0x00007fe8a5f127bb in raise () from /lib/libc.so.6
#1 0x00007fe8a5efd535 in abort () from /lib/libc.so.6
#2 0x00007fe8a55b5dc3 in ?? () from /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0
#3 0x00007fe8a560f67a in g_assertion_message_expr () from /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0
#4 0x00007fe8a55d5cd6 in ?? () from /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0
#5 0x00007fe8a5ac0244 in pa_stream_unref () from /usr/lib/libpulse.so.0
#6 0x00007fe8a5abdb93 in pa_mainloop_dispatch () from /usr/lib/libpulse.so.0
#7 0x00007fe8a5abe0aa in pa_mainloop_run () from /usr/lib/libpulse.so.0
#8 0x00007fe8a5ac0e23 in ?? () from /usr/lib/libpulse.so.0
#9 0x00007fe8a688dfa3 in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0
#10 0x00007fe8a5fd44cf in clone () from /lib/libc.so.6
(gdb)

[scsijon]

Hi all.. playing around with dosbox under EasyBuster from both downloaded and compiled dosboxes I get the following:

[/home]$ dosbox
DOSBox version 0.74-3
Copyright 2002-2019 DOSBox Team, published under GNU GPL.
---
**
GLib:ERROR:../../../glib/ghash.c:377:g_hash_table_lookup_node: assertion failed: (!g_atomic_ref_count_compare (&hash_table->ref_count, 0))
Aborted (core dumped)


Any hints before going down the rabbit hole?..

gcav

ps: as someone mentioned before the battery tray app is not updating.


--stack trace for the curious---
GNU DDD 3.3.12 (x86_64-pc-linux-gnu), by Dorothea LReading symbols from /usr/local/bin/dosbox...done.
[New LWP 7075]
[New LWP 7074]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Core was generated by `dosbox'.
Program terminated with signal SIGABRT, Aborted.
#0 0x00007fe8a5f127bb in raise () from /lib/libc.so.6
[Current thread is 1 (Thread 0x7fe8a45cd700 (LWP 7075))]
(gdb) where
#0 0x00007fe8a5f127bb in raise () from /lib/libc.so.6
#1 0x00007fe8a5efd535 in abort () from /lib/libc.so.6
#2 0x00007fe8a55b5dc3 in ?? () from /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0
#3 0x00007fe8a560f67a in g_assertion_message_expr () from /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0
#4 0x00007fe8a55d5cd6 in ?? () from /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0
#5 0x00007fe8a5ac0244 in pa_stream_unref () from /usr/lib/libpulse.so.0
#6 0x00007fe8a5abdb93 in pa_mainloop_dispatch () from /usr/lib/libpulse.so.0
#7 0x00007fe8a5abe0aa in pa_mainloop_run () from /usr/lib/libpulse.so.0
#8 0x00007fe8a5ac0e23 in ?? () from /usr/lib/libpulse.so.0
#9 0x00007fe8a688dfa3 in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0
#10 0x00007fe8a5fd44cf in clone () from /lib/libc.so.6
(gdb)

There are a small number of error reports relating to this, most seem to blame libevent versions being mixed when building glib packages (something related to libevent versions being different, even when close numbers is mentioned in some of them). Mind you there are quite a few fault reports relating to the g_hash_table~ and all complain about glib / glibc version non-compatability for previously compiled packages. Even Ubuntu was caught with v17.10.

Doubt this helped, but your apparently not alone.

[tom22251]

Hi lp-dolittle,

apologies for not getting back sooner, I was (still is)
under the weather.
That external ext3 drive should be exelent, all you need
is a boot loader on it and setting the bootflag on, with gparted.
I recommend extlinux which is
what I use in the past 8-10 years. It is the simplest and
most reliable in my opinion.

So you already have the EASY_216 folder with the 3 files
inside. Now you need a BOOt folder where you install the
extlinux loader. Type in the terminal:

Code: Select all

extlinux --install /mnt/sda1/BOOT

I have only one USB drive (no harddrives) that's why sda1,
you of course should substitute sda1 with your own drives
name: sdb1 or sdc1, whatever.

That's it, you will have 1- or 2 files in the BOOT folder named ldlinux.sys
and some times an other one too, called ldlinux.c32.
Now the final thing: open a new text file and name it
extlinux.conf. This is where you write your configuration parameters.

Code: Select all

default 1
prompt 1
timeout 33333

say press 1 for Easy-216
say press 2 for Easy-215


label 1
kernel /EASY_216/vmlinuz
initrd /EASY_216/initrd
append  rw vga=792

label 2
kernel /EASY_215/vmlinuz
initrd /EASY_215/initrd
append  rw vga=792

explanation: timeout 33333 is the maximum you can
set and is more then an hour if I remember well.
I usually have sometime more then 10 different
puppy versions on the same drive, so I like to have plenty
of time to decide which one to run.

Now, regarding UEFI you indeed need 2 partitions,
where the boot partition has to be fat32, with the
boot flag on. Copy the EFI folder from Barry's easy-2.1.6-amd64.img file to this partition.
Open up EFI, inside there is BOOT, where you edit only
the file refind.conf :

Code: Select all

timeout 33333
scanfor manual

 menuentry "EASY-2.1.5" {
    volume LINUX
    loader /E-215/vmlinuz
    initrd /E-215/initrd
      icon /EFI/BOOT/icons/os_easy.png
    ostype  Linux
   options "rootwait rw vga=792"
}

 menuentry "EASY-2.1.6" {
    volume LINUX
    loader /E-216/vmlinuz
    initrd /E-216/initrd
      icon /EFI/BOOT/icons/os_easy.png
    ostype  Linux
   options "rootwait rw vga=792"
}

Thats all, it should boot nicely on a
UEFI machine, and in RAM only. You can unplug
the USB drive and plug it in again when you want
to save something, to the ext3 partition.

Let me know if you need more explanation.
Be well,

Tom
[/b]

[lp-dolittle]

@ tom22251

Hi Tom,

thank you for the additional support!

It will take me some time to fully digest the detailed instructions.

In a first attempt, I will try to figure out your 'starting configuration'. Running EasyOS from a usb-stick which is prepared according to Barry's instructions and then configuring a second usb-stick according to your method?

As far as I can see, this second usb-stick cannot simply be a sdx-drive which has only one ext3 partition (labelled LINUX). I need to ponder on when and how you create the UEFI-aware fat boot-partition.

Don't feel obliged to respond quickly. I will let you know the result of my attempts.

kind regards and thanks again

Jürg

[Sage]

Wow! Am I happy. These .iso s on CD-R are oodles faster and with write-once media, ultra secure. Goodbye USB sticks and all that fiddle-faddle with two partitions & co. Muchas Garcias, Barry.

[scsijon]

Wow! Am I happy. These .iso s on CD-R are oodles faster and with write-once media, ultra secure. Goodbye USB sticks and all that fiddle-faddle with two partitions & co. Muchas Garcias, Barry.

If your cd/dvd drive is a dual layer type, as a lot are nowadays, you can also get double layer dvd's that are write-once on the main level and rewrite on the second level, V'b'n have them so they are not the cheapest out, but I haven't had any from the 20pack I got a couple of years ago fail yet. I just formatted the second layer as it's own partition and mapped it across in /etc/fstab. That's 4.5g ro +4.5g rw of space to play with. I did rebuild my iso to have the devx, firefox, buiulding stuff like woofq and a couple of my favorite games in the iso so I wasn't wasting too much space in the first partition. Much better than dual layer rewrite discs for cost too.

[belham2]

Hi Barry,

Hope you're well and in good health & spirits.

Updated to 2.1.7 Buster EasyOS from 2.1.6, went very smooth/easy. All seems good.

Found one little thing, not real important, but if you get a chance maybe could look at it: shouldn't urxvt control (if you set it up & apply it like I did in below pic)..shouldn't it control what rxvt will look like when opening rxvt up?

Here's a pic showing where I can't change rxvt's appearance at all (or am I incorrectly forgetting what urxvt controls??):

[Sage]

Many thanks for those interesting comments, scijon.
Yes, I do, and use, all of those things. I usually use DVD-RW for testing the OS and overwrite them when the next version arrives. These are not what I need and use for HW testing, though. I have a stack of CD-R which have been gifted over the years. These are perfect for machine testing and repair - I never want to save or write to them - quite the opposite - I want them closed/finito! Of course, one can select the option to leave them 'open' so that additional tracks can be added - not good for my purposes. Furthermore, the compact size of Puppies and their various manifestations provide 'light' in the sense of small, size but fully operational OS. [BK's original Puppy was ~29Mb - then John came along with a 50Mb 'credit card' size version, ideal for the top/shirt pocket. Sadly those media are no longer available thanks to the entire field being caught up in the WIntel bloatware race initiated by Redmond many years ago.]

[tom22251]

Hi Jürg,

Let'say you have an internal drive with 3 partitions.
They are called sda1, sda2, sda3 and we want to
install EASY-2.1.6 on an external USB drive.

1) UEFI
Take 1 anything USB: memorystick or external drive just
large enough to fit 3-or 4 times the easy.img's size.
8GB is more than plenty. This will be the sdb drive.

2) GParted
Make a fat32 partition, set the bootflag on.
Give it a label like LINUX or EASY or anything.This will
make it easier to write the configuration files, using
this instead a UUID number. This drive is now sdb1.
Make a second partition ext3, this will be sdb2.

3) Make a new folder in the fat32 partition with
a simple name like E-126.
From now on you'll do everything manually. That is clicking
on the easy.img file and drag easy.sfs, initrd and vmlinuz
into this E-126 folder.
You also drag the EFI folder onto the fat32 drive, where
you'll have only these 2 folders: EFI and E-216.

4) editing the configuration file inside EFI => BOOT
named: refind.conf (it's a simple text file)
Erase everything and paste in just this text:

Code: Select all

 menuentry "EASY-2.1.6" {
    volume LINUX
    loader /E-216/vmlinuz
    initrd /E-216/initrd
      icon /EFI/BOOT/icons/os_easy.png
    ostype  Linux
   options "rootwait rw vga=792"
}

Leave the other files untouched,
like: BOOTIA32.EFI and BOOTX64.EFI
Without them it won't boot.
You'll also see the icons folder where you
can put new logos, 128x128 pixels. I have
about 8-10 different easy logos, for different
easy versions. But it works just as well without
logos.

Now you'll ask what about the second partition ext3.
That is for saving, unsquashing, editing a remaster etc.
You cannot do that in fat32.

So wat I do is I set up a folder with a similar name
inthis case E-216 and after clicking an easy.sfs, drag
everything here. Basically what you have is the whole
Easy-2.1.6 unsquashed ready for editing. Add-remove whatever
you like, and resquashing the folder E-216:

Code: Select all

aplay /usr/share/audio/bark.au           
cd /mnt/sdb2/
mksquashfs E-216 easy.sfs
sync
aplay /usr/share/audio/2barks.au
exit

Now you can replace the original easy.sfs in the fat32 patition.

Inside /mnt/sdb2/E-216 create an other folder called SAVE
or whatever other name you like.
This is the folder for saving any changes you want during the
live session when drives sdb1 and sdb2 are unplugged.
Replug before you want to save and mount sdb2:

Code: Select all

mount /dev/sdb2 /mnt/sdb2
aplay /usr/share/audio/bark.au           
cd /mnt/sdb2/E-216/
cp -a /mnt/wkg/.session/* SAVE
sync
aplay /usr/share/audio/2barks.au
cd
umount /mnt/sdb2
exit

End when you start Easy anew, of course will
boot into the original, pristine version.
So to apply the changes you run this little
script "updt.sh":

Code: Select all

mount /dev/sdb2 /mnt/sdb2
aplay /usr/share/audio/bark.au           
cd /mnt/sdb2/E-216/
cp -a SAVE/* /
sync
aplay /usr/share/audio/2barks.au
cd
umount /mnt/sdb2
restartwm
exit

That's all for now.
Let me know if it works.
This EFI business worked even with
my sister's MacBook Pro, which everybody
says is wery difficult.In fact that's where
I experimented the stuff.

Oh, and by the way, you can have just one
ext3 partition but then you need the other
bootloader "extlinux", and each Easy Version
has its own folder, with the SAVE folder
and unsqushed folder inside, next to easy.sfs,
vmlinuz and initrd files. So basically the
whole drive will have one BOOT folder and
the different Easy folders like E-126, E-127,
E-215, E-217, E-1.1 etc

Bye.
Tom

[Terry H]

am I incorrectly forgetting what urxvt controls??

Hi belham2,

I did similar test to you to check. rxvt is not changed following changes made in urxvt control, but the appearance when using urxvt is changed.

[tom22251]

In my last post I left out 2 lines.
The correct code is:

Code: Select all

timeout 33333
scanfor manual

 menuentry "EASY-2.1.6" {
    volume LINUX
    loader /E-216/vmlinuz
    initrd /E-216/initrd
    icon /EFI/BOOT/icons/os_easy.png
    ostype  Linux
    options "rootwait rw vga=792"
}

sorry

[belham2]

am I incorrectly forgetting what urxvt controls??

Hi belham2,

I did similar test to you to check. rxvt is not changed following changes made in urxvt control, but the appearance when using urxvt is changed.

Oh man .......is this a sign that old age is starting to get me?

How did I brain-fart to such an extent that I forget to open urxvt and stubbornly bang on xrvt wondering what's going on?

Apologies all.

Apologies Barry.

[lp-dolittle]

@ tom22251

Hi Tom,

your ongoing support proves to be very helpful. Despite being slow on the uptake, amply lacking basic knowledge (I'm just a layman who is willing to learn more about IT-technology), I begin to understand the details.

I will follow your instructions, step by step, as soon as possible; and you will get feedback.

If I'm right, you actually are speaking of 2 variants, the 'fat-partition plus save-partition' method and the 'ext3-partition only' alternative. Firstly, I have in mind to focus on the former which seems to work without an "extlinux" boot loader.

thank you very much, your help is highly appreciated!!

kind regards

Jürg