5 years

News, happenings
Message
Author
User avatar
puppyluvr
Posts: 3470
Joined: Sun 06 Jan 2008, 23:14
Location: Chickasha Oklahoma
Contact:

#31 Post by puppyluvr »

:? Hello,
read my post
Well, I guess I must be blind, cause it doesnt explain WHAT SFS?
I made a pet, which if installed, works fine..
Maybe I am being an idiot here, and would gladly admit so..
But first explain how you ran a pet from an heretofore unknown SFS, and expected it to work as if installed as intended.
Maybe my hours of testing were in vain, IDK..
I INSTALLED THE PET to Jwm/Openbox/Icewm/E-16 ect and it WORKED as expected.. on Pups from 2.15 to 5.25.. But I am able to accept I may have missed something.. So INSTALL IT AS INTENDED.. Then test it and tell me where it fails for you, and I will correct, if possible, your issues..
Perhaps it cant find "/initrd/pup_ro5" or wherever the SFS is mounted, IDK..
What I do know is it works for everyone who installs it as intended...

So, in essence, I respond to your criticism with an attempt to resolve your issues with my app..
So please forgo the SFS and try it as intended, and tell me if it gives you trouble then, and I will be happy to help, if I can..
Close the Windows, and open your eyes, to a whole new world
I am Lead Dog of the
Puppy Linux Users Group on Facebook
Join us!

Puppy since 2.15CE...

User avatar
puppyluvr
Posts: 3470
Joined: Sun 06 Jan 2008, 23:14
Location: Chickasha Oklahoma
Contact:

#32 Post by puppyluvr »

:D Hello,
To be fair, I realize some new Pup`s have no /root/my-roxapps so a mkdir would be in order, but as I made it 2 years ago, that is to be expected.. :wink:
Close the Windows, and open your eyes, to a whole new world
I am Lead Dog of the
Puppy Linux Users Group on Facebook
Join us!

Puppy since 2.15CE...

jpeps
Posts: 3179
Joined: Sat 31 May 2008, 19:00

#33 Post by jpeps »

puppyluvr wrote::D Hello,
To be fair, I realize some new Pup`s have no /root/my-roxapps so a mkdir would be in order, but as I made it 2 years ago, that is to be expected.. :wink:
The directories are all present, which is how I ran the scripts and know that they toggle on/off instead of loading new instances. As I noted, the fishtank worked as expected. The issue apparently is related to gtkdialog when loaded by an SFS, if it's working as a pet. Anyway, you don't want testing, and I don't take orders, nor am I seeking help. I figured as long as I was taking up space in the thread, I might as well try your app. There's also the possibility of conflicts with other programs, etc., etc. Anyway...enough. Happy birthday.

User avatar
RSH
Posts: 2397
Joined: Mon 05 Sep 2011, 14:21
Location: Germany

#34 Post by RSH »

Hi.

I don't want to make trouble on your birthday, but I can confirm some issues, when using XToys from sfs. Can't remember in detail but I did decide then, not to use it, because I want to run all my applications from sfs. What doesn't work from sfs I don't use generally.

Happy 5th birthday (I had my 1st about four months ago :) )

RSH
[b][url=http://lazy-puppy.weebly.com]LazY Puppy[/url][/b]
[b][url=http://rshs-dna.weebly.com]RSH's DNA[/url][/b]
[url=http://murga-linux.com/puppy/viewtopic.php?t=91422][b]SARA B.[/b][/url]

jpeps
Posts: 3179
Joined: Sat 31 May 2008, 19:00

#35 Post by jpeps »

RSH wrote:Hi.

I don't want to make trouble on your birthday, but I can confirm some issues, when using XToys from sfs. Can't remember in detail but I did decide then, not to use it, because I want to run all my applications from sfs. What doesn't work from sfs I don't use generally.

Happy 5th birthday (I had my 1st about four months ago :) )

RSH
Puppyluvr doesn't understand why a tester might be reluctant to install 281 files in 72 directories to view a fish tank.

User avatar
puppyluvr
Posts: 3470
Joined: Sun 06 Jan 2008, 23:14
Location: Chickasha Oklahoma
Contact:

#36 Post by puppyluvr »

:D Hello,
Puppyluvr doesn't understand why a tester might be reluctant to install 281 files in 72 directories to view a fish tank.
Well geez, when you put it like that.. :wink:
Look, all I am saying is if you use something in a way it wasnt intended to be used, and it doesnt perform properly, then the "weakest link" isnt my GUI, its gtkdialog and SFS related..
I`m not arguing the merits of 5.6mb (unpacked) of silly toys..
I never made an SFS of it because it is so small, and petget will remove it easily.. However, I do stand by my stuff, so tonight I will make a working SFS of it, if I can.
So I`ve gone from celebrating 5 years on this forum, to defending a silly little app I made 2 years ago.. :roll:
Maybe we got off on the wrong foot, maybe my defending GUI`s annoyed you, IDK. Much of my stuff could use updating, but as a 48 year old single parent with 3 jobs, I have not had the time...
Still, it is what it is. It does work as I intended it to. If even 1 user enjoys it, then I am still happy.... :D
Close the Windows, and open your eyes, to a whole new world
I am Lead Dog of the
Puppy Linux Users Group on Facebook
Join us!

Puppy since 2.15CE...

jpeps
Posts: 3179
Joined: Sat 31 May 2008, 19:00

#37 Post by jpeps »

puppyluvr, welcome to the world of developing :)

(and it isn't worth losing a marriage over)

gcmartin

#38 Post by gcmartin »

The discussion that are being raised here are kinda aimed wrongly.

Let me see if I can help.

It is my understanding that PUPPY Linux is a distro which comes with a PPM. Its the OFFICIAL way to bring needed subystems and application into your public distro's playground.

Several years ago, outside of Puppy Linux AND WHEN INTERNET SPEEDS WERE SLOW even on Internet Backbones, a group began with another methods of compression packaging and layering named SFS. In that design, it was anticipated to take advantage of, not just compression, but also a layering approach.

Fast forward: Over the years several distro packagers has chosen to take using SFS: some for compression and others for layering.

Puppy, seeing this trend has added SFS to its arsenal as SFS moved from a skeleton to what it is now.

Here's what I understand as OFFICIAL. PPM will and continues to process PETs. ANY PUPPY DEVELOPER who creates a PET will have their PET processed correctly by the PPM subsystem such that it can be interrogated for installation and uninstalled if necessary along with all the other things PPM does to make user lives simple, straightforward, and smooth.

Officially, JAMESBOND, Playdayz, Pemasu, and 01Micko has all shared from time to time that its a Developer's CHOICE (maybe I should say "RIGHT" to determine what vehicle to use for the distribution of his product.

This recent posts seems to be making attempts to deny or change the developers offering method to the Puppy distro user community.

As users, we sometimes forget this very fact. Most, if not all, distro developers are working in community to do something that is good for the users based upon technology that they glean useful.

Further, a user who is non-sophisticated will ALWAYS use PPM to add functionality. But, a sophisticated Linux user will draw upon their great knowledge to add stuff to their desktop experience.

To those sophisticated users, I think we can approach this in a different manner.

Here's an option which "MAY" satisfy the SFS issue for those who would prefer that as a deliverable method.
  • Help PuppyLuvr by posting an SFS of his subsystem
  • Request that after your test finding and the method of installation you used on the distro you have so that others can see your finding too.
This level of cooperation is a much better approach for both the developer and users, as well as those who would like to choose their method of installation; namely in or out of the PPM process used by MOST currently available PUPs.

There is much work being done at the community level in SFS and some distro developers are taking advantage of add SFS processing into PPM.

BUT TO BE SAFE, PET PROCESSING CONTINUES TO WORK.

Lastly on the note about the number of programs in a PET, I not so sure that the comment has merit given the fact that most of us use JAVA, FLASH, and all sort of apps that used object oriented development. So,pointing to the fact that a PET might contain more bytes than one would like is NOT a good measure in pushing any developer to SFS.

I will share that some of the development community will upon request make a PPM PET from the SFSs they offer. In some cases, the PET has been actually a wee-bit smaller and in others, the PET has been a wee-bit larger. So the idea of compression might NOT be item in consideration anymore. Oh well.

Maybe a helpful approach is in order.

Here to help
PS. Please remember, I am only trying to help. I have no interest in taking sides. Just only to eyes open enough to help anyone I can.

jpeps
Posts: 3179
Joined: Sat 31 May 2008, 19:00

#39 Post by jpeps »

gcmartin wrote:The discussion that are being raised here are kinda aimed wrongly.

Let me see if I can help.
No offense, but, as usual, you're not understanding some basics. There is nobody monitoring the quality of pets. Users just click on things. If an SFS is breaking down, there's a strong likelihood that something isn't quite right , even if the pet works. There's the little issue of data corruption, for example. In the present case, how many of those 281 files are actually being used? Is it necessary for them to be scattered around in 72 different directories? Am I asking the system to do ungodly tasks for a seemingly very simple process (eg, loading a fishtank app).

Coding is an art; there's good code and a lot of bad code. It's too easy to put buttons on a group of other peoples binaries and scripts, and make a pet. Therein lies the problem.

re: installing/uninstalling: Again, a user just clicks on a few buttons without understanding what's happening. I have my own scripts that check what the files are, what is getting overwritten, etc, and my own uninstaller that gets rid of all the open directories and all the files. Our uninstaller, if you've noticed, is very conservative...meaning lots gets left behind (all the open directories, etc). I'm not worried about uninstalling overwritten files, because I don't load them to begin with.

User avatar
RSH
Posts: 2397
Joined: Mon 05 Sep 2011, 14:21
Location: Germany

#40 Post by RSH »

This recent posts seems to be making attempts to deny or change the developers offering method to the Puppy distro user community.
I don't know if I did get this right.

I don't want to deny anything but I would like to see a BIG change in the developers offering method for programs!

Generally in Linux!

Why?

So, in my humble opinion Linux, especially Puppy Linux has only one BIG issue: the f... symbolic links in the directories of the libraries!
I would like to see hanging the guy who has invented the method to use symbolic links, exactly named as another, for the use of linking it to different libraries in different programs.

I do know a lot of Linux applications from windows. Each application comes as portable application and comes with its own qt, gtk, libs etc.

Why isn't this possible in Linux, especially Puppy Linux?

If Puppy Linux could arrive a stage wherein each application would call its real libraries (and not just symbolic links, which are used for different libraries by different programs), almost all problems relating to PET and/or SFS will be gone!

And also almost all discussion on that.

Then the user would be REALLY free to choose the OS, the programs and the way how to use/install, which would be a great benefit to advise new users of Puppy Linux.

So, just three things to do:

1. Stop talking 8)
2. Kill the guy :wink:
3. Kick out the f... symbolic links from /lib and /usr/lib etc :evil:

A small step for the dog, but a giant leap for the puppy... :lol:

RSH

EDIT:

Just to remind:

Each an every year, tons of GB are wasted for the use of new blinking stuff around the same old applications.

But to have all three different versions of a library included in the OS is too much waste of working and/or saving space?

Naahhh......!

Such behavior, such look on computers and its programs and interfaces has to change - immediately!
[b][url=http://lazy-puppy.weebly.com]LazY Puppy[/url][/b]
[b][url=http://rshs-dna.weebly.com]RSH's DNA[/url][/b]
[url=http://murga-linux.com/puppy/viewtopic.php?t=91422][b]SARA B.[/b][/url]

User avatar
01micko
Posts: 8741
Joined: Sat 11 Oct 2008, 13:39
Location: qld
Contact:

with great power comes great responsibility

#41 Post by 01micko »

Who'd be a developer?

I can sympathise with both of you, puppyluver, jpeps.

Jay, I know the amount of time it takes to learn, then develop stuff. It can be detrimental to certain things therefore balance is needed. There is a fine line between "hobby" and "addiction". No need to drum that in I guess!

I started out with eye candy in pwidgets. I wrote some terrible code. Learned from that and now my code isn't so terrible, not saying it's good but it is a lot simpler and doesn't try to exceed my capabilities. That boundary is gradually pushed up. As for eye candy now, I generally steer clear of it in my personal setups. Sure I, supply compiz for slacko (sfs, with bold warnings) and that is directly based on dinky's work, you mention in the OP.

jpeps, I can fully understand your point. Especially since we can't be fully sure what puppy any application was developed on or for. These days, there are many library incompatibilities between versions. While diversity is great it has a price. I just wonder if we could make a new kind of "sandbox" type thing, say a ~32~64M disposable savefile mounted on top of the existing save file, it may well be possible with the layered filesystem, bit beyond me though. Still, that would be useless to the many who prefer "full" installs.

RSH, as for libs and symlinks, don't hold your breath for your wish! It's a complex eco-system that I don't fully understand, but I believe that it's to do with different versions of libs needing some common ground, if that makes sense. Often development of versions proceeds at great speed, take firefox for example and it's libs, nspr, nss etc. Sometimes backward compatibility is preserved, other times not, ffmpeg is a classic example of not!

I agree that it is way too easy for anyone to create a pet. I'm guilty of distributing rubbish pets, no data loss ones that I know of but I do know of better coders than me that have released pets that have wiped entire partitions. We use at own risk.

One criticism I have of Puppy as a learning environment for Linux is the lack of multiuser support. It breeds bad habits, such as hard coding to root. I try to avoid this but sometimes it's not always possible. If not then I comment that code block. That is just a personal opinion, not on Puppy running as root, but as a general learning tool for Linux,
Puppy Linux Blog - contact me for access

bark_bark_bark
Posts: 1885
Joined: Tue 05 Jun 2012, 12:17
Location: Wisconsin USA

Re: with great power comes great responsibility

#42 Post by bark_bark_bark »

01micko wrote:...One criticism I have of Puppy as a learning environment for Linux is the lack of multiuser support. It breeds bad habits, such as hard coding to root. I try to avoid this but sometimes it's not always possible. If not then I comment that code block. That is just a personal opinion, not on Puppy running as root, but as a general learning tool for Linux,
Agreed using root for everyday use is a bad habit that is formed by puppy. Also having multiuser definitely would save hard drive space and time.

-Hard Drive space would be saved because less save files would have to be made. (no-multiuser=1 save for each user, multiuser=1 save for ALL users)
-Time would be saved because you don't have to go bot in RAM mode and configure it for the new user and wait for the save to be created and all that.

Also as I said ALWAYS being root is a bad habit to form, espically if you visit a lot of websites and have a lot of spam on your email. root + web/email = HUGE security risks.
....

User avatar
01micko
Posts: 8741
Joined: Sat 11 Oct 2008, 13:39
Location: qld
Contact:

Re: with great power comes great responsibility

#43 Post by 01micko »

bark_bark_bark wrote:
01micko wrote:...One criticism I have of Puppy as a learning environment for Linux is the lack of multiuser support. It breeds bad habits, such as hard coding to root. I try to avoid this but sometimes it's not always possible. If not then I comment that code block. That is just a personal opinion, not on Puppy running as root, but as a general learning tool for Linux,
Agreed using root for everyday use is a bad habit that is formed by puppy. Also having multiuser definitely would save hard drive space and time
Ah... be careful here, it's not our place to start this silly debate again. As you see I chose my words with care. Puppy runs as root by design. I'm only saying that it can cause bad habits for novice developers.

There are other threads devoted to user v root around the forums, that is where your discussion belongs. No offence intended.
Puppy Linux Blog - contact me for access

User avatar
greengeek
Posts: 5789
Joined: Tue 20 Jul 2010, 09:34
Location: Republic of Novo Zelande

#44 Post by greengeek »

Happy pupthday puppyluvr! Reading your list of accomplishments is an inspiration to keep on learning and fiddling.

And many happy returns to all the devs and assorted other helpers....

bark_bark_bark
Posts: 1885
Joined: Tue 05 Jun 2012, 12:17
Location: Wisconsin USA

#45 Post by bark_bark_bark »

sorry if my above post offended anyone, that wasn't my intensions.
....

User avatar
puppyluvr
Posts: 3470
Joined: Sun 06 Jan 2008, 23:14
Location: Chickasha Oklahoma
Contact:

#46 Post by puppyluvr »

:D Hello,
Many interesting points here..
jpeps,
I agree that it is too easy for anyone to make and release a pet.
And while I do test to death anything I release, I am surely not a 'guru'.
I have gone back and found major flaws in things that, with my limited skill at the time, I thought were fine.. (Nothing that would wipe a partition, thank God, but still flawed.) But that is how I learn, from my mistakes.. :wink:
As a musician, I learned that everyone makes mistakes, its how we recover from, and learn from them, that sets one apart..
RSH,
What you describe is a modular self inclusive system. Like Tinycore, which I also like a lot.. However, my attempts to build Puppy from TC taught me a few things.. I was successful at building a Pup from TC packages, however, it was rather basic, and still weighed in at around 370mb.. The point of symlinks, as I see it, is to avoid having the same libs in several different places at the same time. However, I agree that some housekeeping is definitely in order, as it has turned, in some Pups, into a 'spaghetti nightmare' of cross linked libs, and links to libs that simply are not there..
I never realized that till I tried to build an Archpup.. Since there was no blueprint, no example to go by, I started from scratch, and quickly found that Puppy symlinks damn near everything.. (IE busybox). Maybe a standardized set of base libs between Pups would help, IDK..
I stand here a relative novice amongst masters. My 'skill set' doesnt compare to most of you, and I have always been fully aware of that.
That is why I have 17 frugal installs to test on, hoping to cover all the bases.. But my lack of knowledge has led me to some unique answers, although maybe not the 'best' answers.. (IE my use of 'filemnt' in MultiPup which other, smarter coders have provided an improved fix for.)
bark_bark_bark,
We are all friends here, hence the jabs.. We all have a common goal, but maybe different ideas on the road there.. Hell, Ive gotten in a lot deeper than this on many occasions.. But I can admit I am just a hack.. :P
So I normally dont let it get to me.. I am a midget amongst giants.. :wink:
Even I am surprised that I got offended for a minute.. :oops: We speak freely here, and that can sometimes be abrasive. But you would have to really try hard to truly offend in this forum.. (IE Ecomoney is still here... :P)
I think Micko`s point was that beating that dead horse will open up doors we cannot close.. I personally cannot run as anything but root, and yes, it is Puppy`s fault.. But I find the task of elevating oneself to superuser, or SU/SUDO, to be so easy as to render the user/superuser debate invalid..
So, as my Aussie friends might say, its all good mate..
jpeps,
Perhaps a pet testing sandbox before release? A thread to post them to, for the 'guru`s' to test first, before they are allowed into the general forum?
I can think of a few I have tried that did some damage, IE a GUI for wbar that overwrote my xinitrc, addending every line.. And maybe a way for the pet spec file to specify which pups the pet has been tested on, as many things will work, for example on Lupu that wont work on Slacko, IE, my MultiPup...

Certainly Puppy has some inherent flaws, because it is unlike any other Linux distro out there, and being cutting edge sometimes means being 'non-compliant'. Also, there is a constant struggle between being Dev friendly, and being User friendly.. Not to mention that many here struggle between being Users and being Dev`s... Maybe it is good to be easy to make a pet, but should be a bit harder to release to the public..

I was at first thrown by the turn my thread had taken, but now I am glad to see it is sparking ideas that may serve the community in general..
Please continue... :)
Close the Windows, and open your eyes, to a whole new world
I am Lead Dog of the
Puppy Linux Users Group on Facebook
Join us!

Puppy since 2.15CE...

User avatar
sunburnt
Posts: 5090
Joined: Wed 08 Jun 2005, 23:11
Location: Arizona, U.S.A.

#47 Post by sunburnt »

BusyBox uses links to access the exec. code inside it.
Barry has to replace some of the links as they don`t work,
this is because BusyBox was not compiled with-for Puppy.

I`ve always thought that there`s no point in using BusyBox for a
Squash file based O.S. like Puppy, it has a useful purpose in a full-install.

gcmartin

#48 Post by gcmartin »

This thread is tit'd "5 years". Some of us have been here that long while others have not.

Even then, some of us understand its history, both in and out of the forum; others do not.

There has been much progress and lots of good innovative things have surfaced through the discussion that we have raised in the community. It has expanded everyone involved in those discussion to greater levels of understanding.

WE have much ground to cover in 2013, so lets begin by not missing the forest for the trees.

Here to help

jpeps
Posts: 3179
Joined: Sat 31 May 2008, 19:00

#49 Post by jpeps »

puppyluvr wrote::
Perhaps a pet testing sandbox before release? A thread to post them to, for the 'guru`s' to test first, before they are allowed into the general forum?
It's too easy to go that route, and my experience says that's NOT the way to go. Sure, life offers risks, but the more freedom the better...it's what inspires learning. I think that's what the phrase "if you meet the Buddha, kill him" was getting at. I believe I still have a little chess app over in testing at DSL......

User avatar
greengeek
Posts: 5789
Joined: Tue 20 Jul 2010, 09:34
Location: Republic of Novo Zelande

#50 Post by greengeek »

puppyluvr wrote:I was successful at building a Pup from TC packages, however, it was rather basic, and still weighed in at around 370mb..
Is the overall size of a puppy a problem though? Yes if you want it all in memory at once (and don't have much ram), but not a problem if you only want a small number of modules in ram at any one time surely?

I'm trying to achieve a lightweight puppy that has a small ram footprint, but the ability to load "units" or "modules" into ram one by one, and each module then removed when the job is done.

How well did your TC run? Was it able to load into a machine with low ram? Or was it choked by it's overall size?

I'd be happy to remove any network related modules from ram in order to be able to run a word processor at reasonable speed, and then kill the word proc and reload the network modules if and when I needed them. On my low power machines I don't necessarily need a puppy that multiprocesses. One job at a time would be fine.

Was TC better than puppy at that?

Post Reply