I remember reading the docs last week or so and discover a sentence saying that using the audio recording function without having an external usb inserted will record to the sdcard and pose the risk of destroying it.
I wanted to ask here immediately if that is still the case, but forgot about it, and right now i dont find that page now anymore. When I search the wiki for “recording” I find tis page, but it’s just redirecting me twice elsewhere Zynthian UI Users Guide - ZynthianWiki
Is that still the case that recording to sdcard is problematic?
Yes. It’s still dangerous, specially if power fails suddenly. Anyway, I use it all the time and very few times (1 or 2 in 7 years) I got a corrupted SD. I would never use it in a live performance and always back up important snapshots or recordings. Up to you, but if you got a corrupted SD, don’t complaint about it. Using a USB pendrive solves the issue and it’s pretty inexpensive.
but at the same time i wondered why it’s so easy to do something potentially dangerous an there is no confirmation dialog to prevent it from happening.
As a software developer i would happily volunteer to change that - unless it’s intentionally as it is and a patch is likely not to be accepted.
in the end i got zynthian to have a hackable device and that seems like a good entry point maybe.
Well… it’s not really dangerous. You are not going to break your zynthian, not even the SD-card. Simply it will get corrupted and you probably need to re-burn the system. Anyway, under the right circumstances, the probability is so low, that the convenience gains and some (myself included) decide to do it and asume the risk. Freedom is sometime risky, my friend
I dont care about “breaking” an SDcard physically, but as a user of any kind of software i very well care about losing data.
I dont think the risk of losing data is an interesting kind of freedom any software user is looking ahead for.
But that wasn’t the question.
I asked if it’s likely that a patch adding a warning and confirmation dialog when a user tries to use this potentially harmfull function without a usb stick inserted would be accepted or if you would not accept such a contribution. They would be asked “are you sure” and can anser yes or no. So the freedom is still there, but people are warned that what they do could harm their data.
Like in most text editors, there is a warning when closing that there are unsaved changes.
I’m now guessing the answer is no, as you now said it’s a good thing that people have the freedom to corrupt their sdcard, right?
I don’t think adding a confirmation for each press of the record button is advantageous. You often want to start recording immediately and a confirmation would just get pressed OK every time. Knowing the risk is better than warning- especially when the alternative is not being able to do what you want.
If people never read it somewhere, they dont know the risk.
A dialog can also have an option “not ask again” - that way the extra step would be necessary only once. And in any case people could do what they want, just with one more press and in a risk-reduced way.
But I understand it is not desired, thanks for the explanation.
I don’t think it’s needed. Using a pendrive is the obvious way if you want to avoid problems. By not using it, you assume the risk. Anyway, implementing a one-time dialog that set a flag, stored in the config file, would not be complicated and it’s an aceptable (and good!) improvement.
Go forward with it or open a feature request, or both, if you want to follow best practice.
From personal experience the more disturbing likelihood is switching into record and it filling the disk which produces a zynthian that wont start up properly cos it has no spare space on the SSD drive to build it’s required structures.
The code where the usb/sdcard decision is made was easy to find, but it will l take a while to get around all the rest - implement a dialog, write config, tests etc - looks like and interesting task
Yes I can open a feature request and a pull request to close that issue when merged.
Thanks
