antiX live usb issue

[BitJam]

Okay, the trick is that fuseiso gives crazy file sizes so if I am only able to mount with fuseiso then I can't use the reported sizes to see if a file is too large. I don't want to use the overall size of the iso file because that could easily get large with remasters, persistence, and other stuff without one file being 4Gig. So I am relying on the quirk (always a mistake) that fuseiso reports oversized files more than once. Someone might want to check that this works for files exactly one 4 Gig and one byte more and one byte less.

@rs55, Welcome to the wonderful world of Linux! You've had one heck of an introduction and I'm a little surprised (but delighted) that you stuck it out.

In the new code, in addition to trying 'mount" first before "fuseiso", if fuseiso was used then I look for duplicate file names. If I find any then I error out saying there are files too large for fuseiso to handle. This is probably not a great message but oh well for now.

In case you want to test this I attached rs55-test-LUM-05.tgz. It is the old game where you:

Code: Select all

tar xzf rs55-test-LUM-05.tgz.
cd rs55-test-LUM-05/LUM 
  ./live-usb-maker --version

You should get:

Code: Select all

      live-usb-maker version 2.32.00 (Thu Mar 28 21:55:53 MDT 2019)
     cli-shell-utils version 2.32.00 (Thu Mar 28 21:57:51 MDT 2019)

Then to test it on one of your large iso files, use the --force=fuse option (I'll bet you never imagined you would be using that). This will cause it to only try fuseiso and when it does it should error out. So this test is quick and painless compared to the mountain of previous testing you were doing. It seemed to work here.

I'm not planning to push these changes tonight because I am very tired and I may not be thinking clearly. If it still seems like a good idea in the morning then I will push the changes tomorrow. It is unfortunate that this is happening just hours after anticapitalista released eight new iso files. Really bad timing seems to be this year's theme. We do the best we can. I suspect that even now fehlix is trying to figure out how to get it to fuseiso mounting work anyway. That would be really something.

I'm really grateful to you rs55 for hanging in and helping us track down these problems even when I was doubting you. It is very strange that one person ran into both fuseiso issues and also ran into the lz4 issue for good measure. How likely was that? Thank you once again.

Never tell me the odds!

[redsparro]

Any ETA on when this fix might be pushed out? Just encountered the same "Error: mount: mounting /dev/loop0 on /live/linux failed: Invalid argument" on a 4.6GB iso. Similarly to rs55, i am using a Thinkpad X220 (don't know if that's just a coincidence or not?). Happy to wait it out until the fix comes through, but just wanted to confirm whether the solution was indeed on its way to the regular update channel? Thanks

[BitJam]

Any ETA on when this fix might be pushed out? Just encountered the same "Error: mount: mounting /dev/loop0 on /live/linux failed: Invalid argument" on a 4.6GB iso. Similarly to rs55, i am using a Thinkpad X220 (don't know if that's just a coincidence or not?). Happy to wait it out until the fix comes through, but just wanted to confirm whether the solution was indeed on its way to the regular update channel? Thanks

New debs should be released today. Or so.

In the meantime you can use the command line version of live-usb-maker to get around this problem with the --force=nofuse option. Try using:

Code: Select all

sudo live-usb-maker --force=nofuse --from $YOUR_ISO_FILE

Thanks to the great work by fehlix, we were able to actually solve the problem even for people who need to use fuseiso. In addition, we built in an md5 check to the live-usb-maker programs so if there is a problem then you will be able to catch it when live-usb-maker runs. We originally wanted to release a fix a few days ago but there were a couple of snags, then we added the fehlix fix and then we added the md5 check. We fixed the problem (we think) and then we added a check just in case there is some other snag we did not expect.

I'm sorry for the delay. It took longer than we expected and we didn't want to release fixes in dribs and drabs.

[Adrian]

I added a "Don't use fuseiso to mount iso files" option in Advanced Option that runs the script with "--force=nofuse" in our GUI program, it should show up shortly in the repo.

[BitJam]

I added a "Don't use fuseiso to mount iso files" option in Advanced Option that runs the script with "--force=nofuse" in our GUI program, it should show up shortly in the repo.

When the new lib and LUM arrive then that option should no longer be needed. We now try "mount" first and only fall back to "fuseiso" if mount fails. In addition, due to the wizardry of fehlix, we are now able to to handle large files even if we are forced to mount with fuseiso.

So with the new code, --force=nofuse is not really needed, we left it in for compatibility and it will cause the program to error out rather than use fuseiso when it is needed. Again, I'm sorry our changes were a few days late.

[Adrian]

Oh, sorry, I misunderstood, should I pull the change then?

[BitJam]

Oh, sorry, I misunderstood, should I pull the change then?

Please wait for the new LUM and new lib. ATM that is a great option to have.

[redsparro]

I'm sorry for the delay. It took longer than we expected and we didn't want to release fixes in dribs and drabs.

No need to apologise, I completely understand. In fact, reading through the topic I'm impressed at how rapidly you resolved it!

Thanks for responding, and the heads up on the fix!

[rs55]

Oh, sorry, I misunderstood, should I pull the change then?

Please wait for the new LUM and new lib. ATM that is a great option to have.

I noticed a new LUM in the MX updater ( LUM 19.4). Is this the final version or Adrian's interim version?

[BitJam]

I noticed a new LUM in the MX updater ( LUM 19.4). Is this the final version or Adrian's interim version?

I'm not sure. The timing is right for either. Here is a quick test:

Code: Select all

$ cd
$ cp /usr/local/bin/live-usb-maker .
$ ./live-usb-maker --version
      live-usb-maker version 2.23.04 (Mon Apr  1 01:00:41 MDT 2019)
     cli-shell-utils version 2.33.05 (Mon Apr  1 10:34:42 MDT 2019)

(change to your home directory copy live-usb-maker from /usr/local/bin and then print out the versions using this local copy of live-usb-maker.)

This new version should work with your large files even if you use --force=fuse (thanks to the wizardry of fehlix). You may also notice that we are now checking the md5 files within LUM. If you are still over-riding the library then you will need to disable that to get access to the new library that contains the fixes:

Code: Select all

cd /usr/local/
sudo mv cli-shell-utils/ cli-shell-utils.old/

The new version should also be being called from the GUI versions (once you disable the over-ride). We try "mount" first and only fall back to "fuseiso" if mount fails. This is because I believe you, rs55, have already found all the problems with fuseiso and we were able to fix them. But if not, the error should show up in the md5 check while LUM is running which should save us all some grief.

This is not the final version. This has the fixes backported to the version from antiX-17.4. A newer version is on the way for antiX-19.