imppup / indripup 64bit ... alpha

[Gnimmelf]

here is something a little strange - on an 8 GB ram computer i managed to get imppup running at about 326 mb. then i created a partition with 16 gb of linux-swap (overkill i know - but theres plenty of harddisk space )
and then the ram usage dropped as low as 296 mb with desktop runnig and internet on. i turned of al the "unnecessary" conky, windows effect etc. the question is - 8 GB of ram should be more than sufficient, how come that running ram drops when creating a swap partition?

kindly gnimmelf

[bigpup]

Your brain will freeze over, trying to worry about stuff you have no control over.
The memory controller, in the Linux kernel, is constantly automatically adjusting memory allocation and how it is used.

Turing off some of Conky, is maybe the reason you see less memory use.

Also, stuff is done in the background, that sometimes is using memory and sometimes is not.

A memory problem, is when you are using way more memory, than it should.
Way more!

Swap is good to always have available.
But with 8GB of memory, it will probably never get used.

[Gnimmelf]

i was just curius about it the measurements in ram and the adding of ram-swab was done after disabling conky and stuff!! so the reason must lie elsewhere?

kindly Gnimmelf

[dimkr]

RAM consumption is surprisingly good.

Even with absolutely huge 5.15.x kernel (nearly identical to the Ubuntu one), latest woof-CE, GTK+ 3 everywhere, ConnMan, Blueman and other heavy stuff, the RAM consumption of a fresh jammy64 woof-CE build is very similar to that of a fresh fossa64 build:

bionic64.png (38.91 KiB) Viewed 2391 times

fossa64.png (30.7 KiB) Viewed 2391 times

jammy64.png (37.53 KiB) Viewed 2391 times

Considering the 2 year gap between 20.04 and 22.04 and all the extras in this build, this is really good. (And I don't think bionic64 would still be much ligher, if built with necessities like PulseAudio and Bluetooth support.)

[Grey]

RAM consumption

By the way, when using different kernels, LXTask shows a different amount of memory (total, second digit) on the same computer. The video card is discrete, it does not take away parts of the RAM. Who knows the link to the explanation document. Or even better, tell us "in your own words"

[bigpup]

Your brain will freeze over, trying to worry about stuff you have little control over.
The memory controller, in the Linux kernel, is constantly automatically adjusting memory allocation and how it is used.

As the Linux kernel is developed. All the stuff in it is tweaked, changed, added to, etc......
So, no two Linux kernels, will do stuff, exactly the same way.

About the biggest control of memory usage, is what program(s) you have running or loaded in memory.
But even what memory, that program is using, at any one point in time, is controlled by the memory controller, in the Linux kernel.
That kernel controller is constantly adjusting for overall system memory demand.

Usually a high usage of memory, that seems to not be normally required, is usually a program with a memory leak or very badly coded.

A memory leak is a type of resource leak, that occurs when a computer program, incorrectly manages memory allocations, in a way that memory, which is no longer needed, is not released.
A memory leak may also happen when an object is stored in memory, but cannot be accessed by the running code.

[dimkr]

Your brain will freeze over, trying to worry about stuff you have little control over.

I disagree, the kernel is configurable, and it's not just the kernel. Small things like woof-CE's "ram-saver" (https://github.com/puppylinux-woof-CE/woof-CE/pull/2435), zram, mksquashfs options, swappiness and even lightweight GTK+ 3 themes can make a huge difference.

The memory controller, in the Linux kernel, is constantly automatically adjusting memory allocation and how it is used.

The kernel's memory management is not some mysterious deity. Its behavior is documented, most concepts (page size, swap, memory compaction, idle page tracking, etc') don't change and there's a (finite) list of changes between kernel versions.

[bigpup]

dimkr,

Thanks for all the work you are doing with Woof-CE and Puppy Linux!

Take your coding hat off and get into the world of us normal users

You are just adding details to what I was giving as general information.

A normal user of Linux and Linux software is not normally going to get deep into the weeds of internal configuration of memory management.

They are only going to affect memory usage by what programs they are using and how Puppy is installed.

With Puppy Linux, you have to be general in info or be very specific, to a specific version, of Puppy Linux.
All that updating and changes in Woof-CE build system, is only in it, until someone makes a new Puppy version, using the latest Woof-CE.
Anything new or changed in Woof-CE, added after a Puppy version is released, is not going to be in that specific Puppy version.

This is kind of high jacking this topic.

If you want to start a new topic about memory control options and the latest improvements to Woof-CE.
Yes do that.
I would be very interested in what is being done and improved.
I am sure others would also want to know.

[dimkr]

Anything new or changed in Woof-CE, added after a Puppy version is released, is not going to be in that specific Puppy version.

Don't worry, the vast majority of these new things is included both in @666philb's and @Grey's builds. woof-CE is not dead, and this effort is not wasted. The "official" 22.04 based Puppy will be a huge improvement over its predecessors.

[bigpup]

dimkr,

How about providing us a topic on what changes, improvements, etc.. have been put into the Woof_CE build system.

Seems like it would be a good topic to have posted in the Announcement section of the forum.

Unless you closely follow the work in Woof-CE, you have no knowledge about these changes.

I am sure people that develop Puppy versions would find this useful info.

I would do it, but half the stuff you do at Woof-CE, I find very hard to understand

Well, maybe this is all that is needed.
viewtopic.php?t=5534

[Grey]

The alpha version of ASCII (okay, ANSI) in terminal
Yes, it is necessary to make the jellyfish not round , remove half of the circle:

jammypup_terminal.gif (78.63 KiB) Viewed 3948 times

[Grey]

Jellyfish was offended that I depicted her incorrectly and caused a short circuit of the wiring

I have electricity in half of the rooms, but not in the other half. There is no electricity where the computer is. I temporarily solved the problem using a bunch of extension cords (I found 11 pieces ). I have to do something tomorrow. Basically, I was going to replace the electricity meter anyway... and here's the reason.

[rockedge]

@Grey Don't overheat those extension cords!!

[Grey]

@Grey Don't overheat those extension cords!!

The service responsible for the repair of electrics did not send anyone, although they promised by phone. Even for money (they charge quite a lot) they didn't come I understand that military operations are not very far away, but still.

Therefore, I called a friend, a 70-year-old specialist. Together we changed half of the wiring. The wiring is aluminum, so the cable was bought the same in cross-section and material (so that nothing oxidized). In principle, it is possible to change to copper, but sometime later. Now it remains (next week) to change the electricity meter, add a new outlet and change the chandelier in the corridor to an LED lamp.

Meanwhile, Ubuntu 22.04 came out (I hope Phil is doing something)

[Grey]

I changed the "recipe" a little, built Jammy on the final Ubuntu libraries. It starts, and there is no system tray at the bottom

[Grey]

@666philb Hi. For the second week I've been forgetting to ask Have you been influenced by the release of Vanilla Upup in terms of enthusiasm and desire to release Jammypup as soon as possible?

As for me, I'm at a crossroads. The problem is the same. Scrolling between tabs with the mouse wheel does not work in either Jammy or Vanilla.

I have seditious and impious thoughts. I can use Fossa for as long as I can... and then switch to something with KDE in which everything works. A traitor of faith ?! Mmmm... That's something like that.

[Gnimmelf]

hi there
what is happening to Imppup?
is there soon a Beta or a final?
does quickpet come in at the end?

kindly Gnimmelf

[LateAdopter]

Ubuntu Impish went EOL on 14th July.
https://fridge.ubuntu.com/2022/07/19/ub ... y-14-2022/

The packages have moved to old-releases.ubuntu.com

If a new build is released it would be Jammy.

[mistfire]

hi there
what is happening to Imppup?
is there soon a Beta or a final?
does quickpup come in at the end?

kindly Gnimmelf

What do you mean QuickPup was just moved to 64-bit. It was still in beta due to xfce's strange behavior since the devs port it to GTK3

[Gnimmelf]

hi there
what is happening to Imppup?
is there soon a Beta or a final?
does quickpup come in at the end?

kindly Gnimmelf

What do you mean QuickPup was just moved to 64-bit. It was still in beta due to xfce's strange behavior since the devs port it to GTK3

didnt mean the distro - I ment the funktion quickpet sorry

[bigpup]

Quickpet is a feature program that 666philb has been putting in his versions of Puppy Linux.

I would say it will most likely be in any Puppy version he produces.

[Grey]

I've put together a version of Jammy for myself, 9.7c , but it's obviously too early to post it.
1. llvmpipe 13.0.1 is used (QuickSetup shows nouveau). Although I don't care, I immediately install the Nvidia driver. In Qemu, virgl is used as expected.
2. Sometimes logout_gui stops working under some conditions. Starts abruptly scolding the 114 line and the symbol ; (semicolon). It seems that the 111 line becomes longer and changes the number. Perhaps here we may need the help of some scripting guru.
3. PPM works fine. However, there is no text in the window of the difference between installed and found versions. Or it is written in white on a white background

But Grub is cute. This is an experimental Puppy anti-gravity frame, inside which everything hovers above the ground. See, there are veins of copper wire under the folds on the eye-coil ?
And on the sole-base it will even be possible to indicate the version or something else

jammypup_grub.gif (68.85 KiB) Viewed 2206 times

[thinkpadfreak]

Sometimes logout_gui stops working under some conditions.

What is the version of jwmdesk in 9.7c?
In 9.7b jwmdesk 3.3 modified logout_gui and eventually broke it. So I installed jwmdesk 3.5.4.

[Clarity]

Hello @Grey
I Like your GRUB2 Menu screen for your distro offering

Would your share your QEMU stanza w/"virgl"? I would like to replicate for tests.

Thanks.

[Grey]

jwmdesk 3.3

Of course 3.3 I didn't suspect him. The last time jwmdesk drank my blood was when jwmrc-tray was changed (then a colleague from Japan advised me to make a separate settings file, thanks again, @MochiMoppel ).

So I installed jwmdesk 3.5.4.

It sounds too simple. If it worked, then one less problem It remains to solve the issue with the third point (PPM), but I don't have any theories yet, plus in practice I was lazy.

[Grey]

It remains to solve the issue with the third point (PPM), but I don't have any theories yet, plus in practice I was lazy.

Well, of course I know that the desired window (Version differences) is called by the show_installed_version_diffs.sh script. And it, in turn, is from ui_Classic. But so far there are no theories anyway

[Grey]

Something like that (picture). The difference window pops up, but there is no data. In addition, there is no icon to the left of the name of the found program. I think it's an /usr/share/icons/hicolor/scalable/status/complete.svg (which links to /usr/share/pixmaps/puppy/dialog-complete.svg). But some change in the script did not help. Who knows for sure - tell me.

And yes, jwmdesk was really to blame for the failure of the exit window

ppm_bugz.gif (21.94 KiB) Viewed 1972 times

[jrl]

Right up front, as a newcomer, I want to emphasize that I greatly appreciate all the effort expended by developers.
Accordingly, everything I write should be treated as constructive ... without the slightest tinge of complaint.

- - -

I recently downloaded the jammypup64-9.8.iso version of jammypup64 at

https://archive.org/download/jammypup64 ... jammypup64

and have been checking it out.

I have noted a very considerable number of jammypup64 shortcomings;
but those must all wait until later ...
because jammypup64 seems to suffer from what seems to be a serious problem ... viz:
when starting up, it seems unable to retrieve any save-file which is on an NTFS partition.
jammypup64 can happily =create= a save-file on an NTFS partition when shutting down ...
but it seems unable to retrieve same when restarting.

I checked out various older puppy_linux editions to find out at what stage the problem crept in.

https://distro.ibiblio.org/puppylinux/p ... 64-9.5.iso ... seems to have no such problem.
https://archive.org/download/Puppy_Linu ... p/Imppup64 9.6.1.iso ... seems to have no such problem.
https://archive.org/download/jammypup64 ... 4-9.7a.iso - ahhh, there's the problem.
https://archive.org/download/jammypup64 ... 4-9.7b.iso - again, the problem.
https://archive.org/download/jammypup64 ... 64-9.8.iso - again, the problem.

So it looks like the step from imppup64 to jammypup64 which introduces the problem.

Intriguingly, though, the corresponding step from peabee's imppup32 to jammypup32 [and even beyond!] seems clean.

https://sourceforge.net/projects/zestyp ... o/download ... seems to have no such problem.
https://sourceforge.net/projects/zestyp ... o/download ... seems to have no such problem.
https://sourceforge.net/projects/zestyp ... o/download ... seems to have no such problem.

As to the problem itself:
my tracing indicates that an ntfs-3g return code of 21 is received by jammypup64's init script,
twice per call to /sbin/mountpartition,
when the latter [/sbin/mountpartition] makes the following two calls to ntfs-3g:
line 43: ntfs-3g $MNT_DEV $MNT_DIR -o umask=0,no_def_opts,noatime,rw,silent
line 47 [retrying with force option]: ntfs-3g $MNT_DEV $MNT_DIR -o umask=0,no_def_opts,noatime,rw,force,silent

Now, a return code of 21 from ntfs-3g is documented as "meaning" [hey, I use that term extremely loosely]
Unclassified FUSE error
[see https://www.systutorials.com/docs/linux ... -3g.probe/]
Not really very helpful.

The call stack as these return codes of 21 are received is as follows:

init line 1086: #ensure that save partition is mounted
init line 1089: [ "$SAVEPART" ] && ensure_save_mounted
>>>
init line 293: ensure_save_mounted() {
init line 295: ensure_mounted "$SAVEPART" "$SAVE_MP"
>>>
init line 250: ensure_mounted() {
init line 263: /sbin/mountpartition /dev/${1} $ONE_MP $ONE_FS
init line 266 [again after 3-sec sleep in case usb optical drive]: /sbin/mountpartition /dev/${1} $ONE_MP $ONE_FS

However: once jammypup64 initialization is complete,
it is then not a problem to have the "troublesome" NTFS partition mounted on request --
the return code from ntfs-3g is then 0, and the partition can be accessed completely normally
[though of course it is far too late by then for it to be set as the save-file partition].
The "return code of 21 from ntfs-3g" thing is only during the jammypup64 initialization phase
[which explains how jammypup64 can happily create a save-file on an NTFS partition when shutting down].

With that ... over to you!
[I have quite a lot of other information available, if need be -- but that's the heart of it.

[Grey]

NTFS partition.

viewtopic.php?p=75524#p75524

[Gnimmelf]

imppup works like a breeze!

up and running with only around 290 mb ram in use!

why isnt it developed on? it runs on a newer kernel than jammypup an uses lesser ram?

is it getting outdated?

kindly Gnimmelf