[Discussion] live-boot development

[solk]

English is not my native language, so i apologize in advance if some text is hard to understand.

Recently i came up with an issue with live-boot on Debian 12 system that raised a (sort of philosophical) question.

Seems that code that was submitted 9 years ago and got fixed 2 years ago , is now live in Debian 12.

Bug that existed 7 years enabled to use persistent storage on partition that also contained life rootfs.

Comment (from 9 years ago) on that code states:
"The devices that are hosting the actual live rootfs should not be used for persistence storage since otherwise you might mount a parent directory on top of a sub-directory of the same filesystem in one union together."

Although that makes sense if such thing happens (from technical point of view), it will create kind of "childcare" situation, where developer is trying to limit functionality so that user may not create "bad scenario".
But since this bug has previously existed 7 years, it may be logical to assume that this "bad scenario" has not emerged (at least i did not find anything on forum search), or emerged very rarely.

For about last 6 years or so, i have used pre-built Debian live OS that is running on thousands of machines, where persistent storage is located on same partition as rootfs without any problems.
This may be because i used correct labels and path parameters for live-boot.
What i'm trying to say is, if you do things correctly, that issue does not happen.

So that rises a question: how to approach this issue?

Should i create a bug report about missing functionality?
(You cannot use same partition for persistent storage and live rootfs)
Or should i create bug report about missing documentation?
(Only place where this limitation is mentioned is comment on code)

And that "philosophical" question would be: which is the "Debian way", to provide functionality to users and risk that they might broke something or try to protect users from doing destructive things by limiting functionality?

[bw123]

Should i create a bug report about missing functionality?

I think it's important enough to followup on in some way, but it was obviously a bug in the code and it needed to be fixed.

And that "philosophical" question would be: which is the "Debian way", to provide functionality to users and risk that they might broke something or try to protect users from doing destructive things by limiting functionality?

I think i'd just report the loss of a 'feature' that resulted from the long-standing bug, and see if maybe that could be reworked in another way. People might want to talk philosophy, but when it comes to code, I bet they're just being practical. If it's important and not too complicated, I think it's a reasonable thing to ask for. Have you managed to work around the new bug-fixed code?

https://lists.debian.org/debian-live/ is the list, I'm sure it will make an interesting thread, I'd like to hear more about it myself...

Or should i create bug report about missing documentation?

I have always found the live system too complicated and underdocumented, so I wouldn't pursue that. Focus on the loss of functionality, that's a good fight to fight.

Good luck!

[solk]

Have you managed to work around the new bug-fixed code?

I just made that function find_persistence_media () useless by changing that variable back to what it was before fixing before rebuilding initramfs.
So basically, i restored that bug.

https://lists.debian.org/debian-live/ is the list,...

Thank you, i will create thread there, although i have no good feelings about it, since this function that got fixed was planned as feature and i think that there is very slim chance that it will be removed.