antiX live usb issue

[fehlix]

The problem is in fuseiso9660.

The issue with fuseiso9660 is reproducable and results in lot of errors during the run of live-usb-maker.
The main reason is: fuseiso9660 converts all file names into uppercase filenames,
hence those filenames like "isolinux.cfg" in lower case are not found anymore
as linux is case sensitive with filenames. So BitJam is right and identified the culprit correctly.

[anticapitalista]

The problem is in fuseiso9660.

The issue with fuseiso9660 is reproducable and results in lot of errors during the run of live-usb-maker.
The main reason is: fuseiso9660 converts all file names into uppercase filenames,
hence those filenames like "isolinux.cfg" in lower case are not found anymore
as linux is case sensitive with filenames. So BitJam is right and identified the culprit correctly.

So why does it work in MX with the old lum and MX uses fuseiso9660?

[rs55]

Glad you have figured this out. Will your changes make it into updates for the basic "stable" release? At the moment, my system is working fine with the gui that anticapitalista provided and fuseiso9660. So I will leave that alone until a new LUM is released.
Thanks for including me in your process.

I need your help rs55. There was nothing wrong with my code. The problem is in fuseiso9660. Please uninstall fuseiso9660 and reinstall fuseiso and try my new code and see if it fixes your fragmentation problem.

The reason anticapitalista's "solution" worked is he was calling my old code that doesn't use fuseiso9660. The fuseiso9660 program is broken and does not work. It was causing all of those strange problems you and anticapitialista and d.o. had when using my new code. One change I made in the new code is I used fuseiso9660 if it was available. I will certainly fix this but we still need to see if my new code fixed your fragmentation problem. The reason my new code worked for me and for fehlix is we both knew fuseiso works so neither one of us tried changing to fuseiso9660. I was able to reproduce the plethora of strange problems when I used fuseiso9660 instead of fuseiso.

Since yours is the only system we know of that has the fragmentation problem please uninstall fuseiso9660, reinstall fuseiso and try my new code again. All of the new problems were caused by fuseiso9660.

SUCCESS !! I downloaded you latest version : LUM 2.22.03 . Uninstalled fuseiso9660, Installed fuseiso and ran it. And it made the usb and the usb booted up properly !
Congratulations!
Question: What is the process by which this fix will make it into the "Stable" updates channel? When can we expect to see that?

[fehlix]

So why does it work in MX with the old lum and MX uses fuseiso9660?

The new code is different, which includes some new filehandling.

[rs55]

So why does it work in MX with the old lum and MX uses fuseiso9660?

The new code is different, which includes some new filehandling.

My understanding is old LUM calls fuseiso. MX has fuseiso9660 installed. So - no fuse.. is called at all. Thats why it works.
The new code from BitJam calls fuseiso, but he has made some fixes. So with fuseiso installed it works.

[BitJam]

So why does it work in MX with the old lum and MX uses fuseiso9660?

My old code doesn't try to use fuseiso9660. It only tries fuseiso. The fragmentation problem only occurs if fuseiso* is used. So if you uninstall fuseiso then the old code will work because fuse mounting is not used. It is possible I knew there was a problem with fuseiso9660 when we first added fuseiso. That would explain why my old code doesn't even try fuseiso9660.

As fehlix figured out, the problem with fuseiso9660 is it converts all filenames to uppercase. This is why it created so many strange problems and why my program couldn't find files it was looking for.

You were right to include fuseiso and not fuseiso9660 in antiX-17.4. The only problem was that rs55 had a system where we needed to work harder to defrag vmlinuz. But the fuseiso9660 red herring caused you and rs55 and d.o. to have a bunch of mysterious problems while fehlix and I had no problems because we knew fuseiso worked and we didn't bother installing fuseiso9660.

@rs55 thank you for persevering through the chaos and trying my fix once more!

I think we are all back on track. The new code works. The fragmentation problem that rs55 had is fixed. I will have my code only use fuseiso and not fuseiso9660. AND we have learned that we don't want to have fuesiso9660 on our systems but we do want fuseiso.

I sympathize with you anti. I imagine you thought "*I* did something wrong by including fuseiso instead of fuseiso9660". I got caught in this same trap. I though "I did something wrong in my code that caused these problems". We both blamed ourselves first.

[rs55]

So - now, what is the best way to get your new code to be called by the antix control center gui?

[BitJam]

So - now, what is the best way to get your new code to be called by the antix control center gui?

Please give me a day (or more). For now you could manually copy two files from the tarball, but I want to change cli-shiell-utils so it DOES NOT try to use fuseiso9660. If you want to do it now then change to the new-LUM directory and run:

Code: Select all

sudo cp -i live-usb-maker /usr/local/bin
sudo cp -i ../cli-shell-utils/cli-shell-utils.bash /usr/local/lib/cli-shell-utils/

I added -i to the cp commands so it will warn you if you want to overwrite a file. You do want to overwrite but I use this warning to make sure I'm copying to the right place. These versions will get clobbered if cli-shell-utils or live-usb-maker get updated.

An alternative is to edit /usr/local/bin/antixcc.sh (the control centre) and change it so the LUM entry points to the version of LUM in your new-LUM/ directory. That location is important because LUM will look for ../cli-shell-utils/cli-shell-utils.bash (relative to the location of LUM, not your current directory). If it exists then LUM will use that version of the library instead of the version from the standard location. The key thing is the new library.

I *hope* we can get debs of the new versions into a buster or testing repo but that may take a week or so. We are trying to get a bunch of new code ready for antiX-19 (and then MX-19) so we are all very busy now. This problem put me off schedule but I am glad you reported it and worked with us so we could fix it.

[anticapitalista]

So - now, what is the best way to get your new code to be called by the antix control center gui?

For now, I would suggest you stick to the working antiX lum-gui that got recently updated and the cli-shell-utils in the repo.

Bitjam's new fixes will require new cli-shell-utils, live-usb-maker-gui-antix (and possibly live-usb-maker) debs.

[rs55]

So - now, what is the best way to get your new code to be called by the antix control center gui?

For now, I would suggest you stick to the working antiX lum-gui that got recently updated and the cli-shell-utils in the repo.

Bitjam's new fixes will require new cli-shell-utils, live-usb-maker-gui-antix (and possibly live-usb-maker) debs.

Yes , that makes sense, thats what I will do. Thanks for providing that gui.