Install Puppy to non-root location at USB stick

[Crash]

vg1,

Yes, this is great stuff. Without Pizzasgood's original post, I would have been clueless. I had downloaded the "unleashed" versions of Puppy in the past, and could look through the code, but until now I didn't know how to do anything with the information I was gleaning. Now it seems obvious. A definite Paradigm Shift - Inflection Point - what have you. I'm sure if I wasn't so lazy and had read elsewhere on the Forum or Wikis or in Barry's notes, I would have seen what was going on earlier. But late is better than never.

I'll look into the P214R code to see what Dougal was up to. It should be another good exercise, since if I read Pizzasgood correctly, I will be mounting a filesystem rather than just unzipping a file. So that will be new for me.

The "Pup_save search" code in init is more lengthy, but should be able to be revised per your suggestion. I thought the "file search" code modification task was insurmountable until I started looking at it closely. Everything is there, ready to be worked on. And it is classic code. It would be a great machine exercise for a computer science class.

Regarding having the pup_save file in a different directory than the pup_xxx.sfs file: You are right, I'm wrong (not the first time). I looked back at the directories and they all had the pup_xxx.sfs file in them. I even made a test in a fresh directory with only a pup_save.2fs file stand-alone in it. The boot process craftily put a copy of pup_412.sfs in it! I guess the lesson here is that, at least in the current configuration, if you have multiple save files they should all be in the same directory that the corresponding pup_xxx.sfs file is in.

Anyway, this post actually has solved the original problem and probably does have to move elsewhere. It will take me some time, as usual, to digest everything that was said here before I can do further work with the init code. When I get something interesting, I will put it in a different post.

This is great fun for me. The power of Puppy is the entirety of it. Barry, the developers, the forum members, everybody. It has created renewed interest in me for computing as a hobby, and I'm sure it has for countless others as well.

/// Edited May 17

Note further activity regarding this thread topic, including discussion of the PSUBOK parameter, is continued at:

http://www.murga-linux.com/puppy/viewtopic.php?t=38030

///

[ICPUG]

Just to add my congratulations for this advance.

If I read it right the modified code only searches the subdirectory specified and not all over the disk and then filter to the subdirectory.

This will speed things up and help to keep multiple puppies in their own kennels.

Not sure where add-on SFS files are stored. Do they still have to be in root?

[vg1]

Crash,

thanks for your comments.
In my last but one para above please read correctly:
'All coexisting in the same folder, and all called from grub with a dedicated menu.lst entry with the appropriate 'PUPSAVE='.

I added your code to p400 & p411 and they all work fine.

Re Dougal's code - of course it cannot be used here as the whole init layout was quite different in p2. Hopefully you will be able to grasp the principle and use it in p412 init.

[Toxic]

Ok, now I see that code from above is included into Puppy 4.2. However there's redundant kernel parameter: PSUBOK. First, It can be evalueted in runtime (PSUBOK= [ "$PSUBDIR" ] ? "TRUE" : ""), and second, there's actually no need for it at all.
In block for path forging we may evaluate maxdepth value based on PSUBDIR presence and can always use /mnt/data/$PSUBDIR format even if this parameter is not defined since concatenation result in that case will be /mnt/data/

in pseudo-code:

maxdp = [ "$PSUBDIR" ] ? 1: 2;
FND_INITRDGZ="`find /mnt/data/$PSUBDIR -maxdepth $maxdp ....

and so on...