alternative puppy build system
here is what i propose
1. build a minimal puppy with jamesbonds woof-ce next
2. put new pets in the woof-ce next repository
made from components of other systems
this can be by just adding them to the folder in woof-ce next
3. run woof-ce next and it will build a new puppy with the new stuff in it
you will have a new unique puppy in minutes
and anyone will be able to do it
wanderer
1. build a minimal puppy with jamesbonds woof-ce next
2. put new pets in the woof-ce next repository
made from components of other systems
this can be by just adding them to the folder in woof-ce next
3. run woof-ce next and it will build a new puppy with the new stuff in it
you will have a new unique puppy in minutes
and anyone will be able to do it
wanderer
Re: pkg or puppy build
One can do that but a given distro might not even have the utilities that a post install script expects. Debian for instance uses a lot of very Debian specific tools in its package management system.foxpup wrote:We can already install .deb .txz .rpm ... to Puppy. Or is this not what you mean with 'components of any system'?wanderer wrote:hi a243a
yes just take any component of any system
convert it into a pet and put it in
the jamesbond woof-ce next repository
then have his system load it into a puppy
that way you can make the puppy a combination of all systems
and you wouldnt have to worry about
keeping up with the changes of the big distros
that is what i am trying to figure out how to do now
wanderer
Actually what we could have is a repository of post install scripts and then when doing the conversion incorporate the replacement post install script.
We might also want to consider moving the binaries around (or alternatively automatically creating symlinks) to avoid too long of a LD_LIBRARY_PATH. mistfires upgrade of the package manager might fix this symlinks issue
This would be preferable but more time consuming. It also risks breaking binary compatibility with a given distribution.Should we not consider to compile in the base Puppy, instead of having a repository?
It would be good if the in the tool there was an option to select the package manager. We would probably have a different repo of post install scripts for each package manager.In that case petbuilds in woof CE could be interesting.
But even more so Pkg from sc0ttman, one of the best things created in Puppyland on the last years:
Pkg - CLI package manager
Just my idea.
Also note that regardless of whether we compile or use existing binaries, we still need to develop the post install scripts, when a package requires altering a configuration file of another package.
hi all
another thing i am going to do
is to move the init file out of the initrd.gz
and put it in the main iso
i will put a symlink in the initrd.gz
so that after it loads it will look for the file init.tx in the iso
all the instructions --- i repeat all the instructions
for loading and setting up puppy will be in init.txt
so that when you want to change it
you will only need to take isomaster
and replace init.txt in the iso
you will not need to open the cpio initrd.gz
this will save an annoying step
and make changes quicker and easier
wanderer
another thing i am going to do
is to move the init file out of the initrd.gz
and put it in the main iso
i will put a symlink in the initrd.gz
so that after it loads it will look for the file init.tx in the iso
all the instructions --- i repeat all the instructions
for loading and setting up puppy will be in init.txt
so that when you want to change it
you will only need to take isomaster
and replace init.txt in the iso
you will not need to open the cpio initrd.gz
this will save an annoying step
and make changes quicker and easier
wanderer
hi all
if you have been following this thread
you may know that this project consists of 2 sub projects
1. my puppy modular build system
2. and jamesbonds woof-ce next
we have already spoken about trying to develop jamesbonds woof-ce next
with a more limited target to make things more manageable
but this will require quite a learning curve
my puppy modular build system
which i intend to develop to completion
is far simpler than woof-ce next
anyone that is able to use some simple commands
like cp mv ln -s cpio and mksquashfs will be able to maintain the code
it simply consists of breaking puppy into its components
and making modules of them
and then rearranging the modules as desired
the base of this 1st attempt will be upupbb32
but it will be generalizable to any puppy
and will be able to make unique puppies
that being said i am changing the names of the isos
because i dont like complex names
since there is only 3 main components of puppy
we can use letters rather than numbers
so we dont have dashes
iso 1 will be named a1 a2 a3 etc
iso 2 will be named b1 b2 b3 etc
iso 3 will be named c1 c2 c3 etc
i am now working on the next b iso b2.iso
which will be with an empty main.sfs file
then i will make b3.iso which will have a busybox system in it
then b4.iso which will have a basic x and jwm system in it
etc
remember
to keep things organized
all the important stuff will be in the first post
always look there
for a complete overview
and the latest news
wanderer
if you have been following this thread
you may know that this project consists of 2 sub projects
1. my puppy modular build system
2. and jamesbonds woof-ce next
we have already spoken about trying to develop jamesbonds woof-ce next
with a more limited target to make things more manageable
but this will require quite a learning curve
my puppy modular build system
which i intend to develop to completion
is far simpler than woof-ce next
anyone that is able to use some simple commands
like cp mv ln -s cpio and mksquashfs will be able to maintain the code
it simply consists of breaking puppy into its components
and making modules of them
and then rearranging the modules as desired
the base of this 1st attempt will be upupbb32
but it will be generalizable to any puppy
and will be able to make unique puppies
that being said i am changing the names of the isos
because i dont like complex names
since there is only 3 main components of puppy
we can use letters rather than numbers
so we dont have dashes
iso 1 will be named a1 a2 a3 etc
iso 2 will be named b1 b2 b3 etc
iso 3 will be named c1 c2 c3 etc
i am now working on the next b iso b2.iso
which will be with an empty main.sfs file
then i will make b3.iso which will have a busybox system in it
then b4.iso which will have a basic x and jwm system in it
etc
remember
to keep things organized
all the important stuff will be in the first post
always look there
for a complete overview
and the latest news
wanderer
woof-ce next, a try
On a dual core machine with wired eth0 I tried to build a slacko with woof-ce next..
I've used slacko6.9.9.9, a build from norgo.
I did not know in what script and on what line I had to change /bin/dash to /bin/sh,
so I changed it in all scripts in workdir where I could find /bin/dash.
I also commented out firefox in /workdir/basesfs,
since I always use firefox as portable/linked and have it available like that on my machines.
I used the first available kernel, but it would be nice to have a more recent one as well.
I have another machine which requires at least k4.1.
Building seemed to go well, though I have not kept nor read the output.
I also made the iso and the devx.
I installed the new puppy frugally on the harddisc and added an entry to menu.lst and booted.
I goes to X and a lot works, some things do not.
examples:
- I cannot get azerty/be-latin1 keyboard lay-out
- the menu looks almost 'abandonned'
- shutdown or reboot from the menu does not work or can even hang the computer, I have to shutdown/reboot from terminal
- .desktop files or defaults are not well configured
- I could not get gslapt to work
- I could not get firefox to work, but QtWeb version 384 (not the latest 385) works
- but I have no internet connection
etc
Not sure if I should continue this.
It is probably a lot of work and a steep learning curve for me.
And I do not have much time for it
Starting from a working and well built Puppy (like upupbb) seems to be the more effective way,
though maybe not the most interesting one.
I've used slacko6.9.9.9, a build from norgo.
I did not know in what script and on what line I had to change /bin/dash to /bin/sh,
so I changed it in all scripts in workdir where I could find /bin/dash.
I also commented out firefox in /workdir/basesfs,
since I always use firefox as portable/linked and have it available like that on my machines.
I used the first available kernel, but it would be nice to have a more recent one as well.
I have another machine which requires at least k4.1.
Building seemed to go well, though I have not kept nor read the output.
I also made the iso and the devx.
I installed the new puppy frugally on the harddisc and added an entry to menu.lst and booted.
I goes to X and a lot works, some things do not.
examples:
- I cannot get azerty/be-latin1 keyboard lay-out
- the menu looks almost 'abandonned'
- shutdown or reboot from the menu does not work or can even hang the computer, I have to shutdown/reboot from terminal
- .desktop files or defaults are not well configured
- I could not get gslapt to work
- I could not get firefox to work, but QtWeb version 384 (not the latest 385) works
- but I have no internet connection
etc
Not sure if I should continue this.
It is probably a lot of work and a steep learning curve for me.
And I do not have much time for it
Starting from a working and well built Puppy (like upupbb) seems to be the more effective way,
though maybe not the most interesting one.
hi foxpup
yes i agree
its too steep a learning curve
and it also has other problems for us
it will need to be updated before it can be used
it will take a guru to maintain it and develop it
and we need something that --- anyone --- can use and develop
i will play with woof-ce next to learn about woof-ce
but i will work on my simple system for the time being
read the prior post and this will explain a little of how simple it is
wanderer
yes i agree
its too steep a learning curve
and it also has other problems for us
it will need to be updated before it can be used
it will take a guru to maintain it and develop it
and we need something that --- anyone --- can use and develop
i will play with woof-ce next to learn about woof-ce
but i will work on my simple system for the time being
read the prior post and this will explain a little of how simple it is
wanderer
@foxpup.
Congratulations for getting to this point!
As for the rest, it's up to you. Getting over the remaining hurdles can require a lot of
study and testing.
Unfortunately, your example is just another example of the devs at Woof-CE releasing
an "immature proposal". They should test and test and re-test it before publishing.
I personally like to learn new things, but not all Puppyists have the time or the inclination.
The devs at woof-CE have the knowledge: they would readily know what to correct in
their product, while we common mortals will hunt for a solution to their bugs for days
on end. That is not ok.
Our founder BarryK never did that. When he published a Pup it was usable OOTB.
With maybe a few details to adjust, but that was it.
BFN.
Congratulations for getting to this point!
As for the rest, it's up to you. Getting over the remaining hurdles can require a lot of
study and testing.
Unfortunately, your example is just another example of the devs at Woof-CE releasing
an "immature proposal". They should test and test and re-test it before publishing.
I personally like to learn new things, but not all Puppyists have the time or the inclination.
The devs at woof-CE have the knowledge: they would readily know what to correct in
their product, while we common mortals will hunt for a solution to their bugs for days
on end. That is not ok.
Our founder BarryK never did that. When he published a Pup it was usable OOTB.
With maybe a few details to adjust, but that was it.
BFN.
musher0
~~~~~~~~~~
"You want it darker? We kill the flame." (L. Cohen)
~~~~~~~~~~
"You want it darker? We kill the flame." (L. Cohen)
My understanding is that woof-next was working and that the reason it is not working now is due to changes in the compatible distributions. Anyway, I'm glad the code is published because someone might want to continue it's development. If a given project doesn't have a certain level of polish then the developers can say so and the less adventurous people can avoid it if they choose.musher0 wrote:@foxpup.
Congratulations for getting to this point!
As for the rest, it's up to you. Getting over the remaining hurdles can require a lot of
study and testing.
Unfortunately, your example is just another example of the devs at Woof-CE releasing
an "immature proposal". They should test and test and re-test it before publishing.
I personally like to learn new things, but not all Puppyists have the time or the inclination.
The devs at woof-CE have the knowledge: they would readily know what to correct in
their product, while we common mortals will hunt for a solution to their bugs for days
on end. That is not ok.
Our founder BarryK never did that. When he published a Pup it was usable OOTB.
With maybe a few details to adjust, but that was it.
BFN.
Hi s243a.s243a wrote:My understanding is that woof-next was working and that the reason it is not working now is due to changes in the compatible distributions. Anyway, I'm glad the code is published because someone might want to continue it's development. If a given project doesn't have a certain level of polish then the developers can say so and the less adventurous people can avoid it if they choose.musher0 wrote:@foxpup.
Congratulations for getting to this point!
As for the rest, it's up to you. Getting over the remaining hurdles can require a lot of
study and testing.
Unfortunately, your example is just another example of the devs at Woof-CE releasing
an "immature proposal". They should test and test and re-test it before publishing.
I personally like to learn new things, but not all Puppyists have the time or the inclination.
The devs at woof-CE have the knowledge: they would readily know what to correct in
their product, while we common mortals will hunt for a solution to their bugs for days
on end. That is not ok.
Our founder BarryK never did that. When he published a Pup it was usable OOTB.
With maybe a few details to adjust, but that was it.
BFN.
Comment on your 1st sentence:
It is unfortunately very human to blame the guy upstream...
Comment on your 2nd sentence:
how many woof-CE devs have mentioned or do mention: "this one is experimental"?
Why I am sticking with PuppyLinux, I don't know...
Best regards.
musher0
~~~~~~~~~~
"You want it darker? We kill the flame." (L. Cohen)
~~~~~~~~~~
"You want it darker? We kill the flame." (L. Cohen)
resume
Jamesbond has stated very clearly that the project was retired.
So it is perfectly normal that not everything works anymore after 4 years of development in woof-CE and in Slackware.
It is not that the 'devs' did not warn us. No need for blaming.
I agree with musher0 that the devs
It still builds very fast and easily, so the base is good.
I should try to read and understand what the build scripts actually do.
So I wll be back with questions next time.
So it is perfectly normal that not everything works anymore after 4 years of development in woof-CE and in Slackware.
It is not that the 'devs' did not warn us. No need for blaming.
I agree with musher0 that the devs
So it would help if the devs could point us in the right direction.would readily know what to correct in their product, while we common mortals will hunt for a solution [left this out ] for days on end.
It still builds very fast and easily, so the base is good.
I should try to read and understand what the build scripts actually do.
So I wll be back with questions next time.
good luck foxpup
yes take a look at the woof-ce next scripts
i think if we play with it
we will eventually figure it out
it works and builds an iso
so maybe just adjusting where it points
will solve the problem
i wonder if just replacing the pets will work
i will try that soon
maybe jamesbond will give us some advice
to get us started
i also will continue to work on my system
which is so simple there is virtually no learning curve
thanks for helping
post if you learn anything
wanderer
yes take a look at the woof-ce next scripts
i think if we play with it
we will eventually figure it out
it works and builds an iso
so maybe just adjusting where it points
will solve the problem
i wonder if just replacing the pets will work
i will try that soon
maybe jamesbond will give us some advice
to get us started
i also will continue to work on my system
which is so simple there is virtually no learning curve
thanks for helping
post if you learn anything
wanderer