jamulussoftware / jamuluswebsite

This is the GitHub Pages repository for the Jamulus main website. For the jamulus application source code, please visit jamulussoftware/jamulus.
https://jamulus.io
GNU Lesser General Public License v2.1
16 stars 79 forks source link

welcome path gotcha tip #927

Closed mcfnord closed 1 year ago

mcfnord commented 1 year ago

Help user locate path without providing specific guidance.

mcfnord commented 1 year ago

re: https://github.com/jamulussoftware/jamulus/issues/3072

mcfnord commented 1 year ago

This isn't good enough, because the default security prohibits access to /home/*. Even if /home/jamulus exists?

ann0see commented 1 year ago

Before this gets merged, please squash the commits or squash-merge them.

gilgongo commented 1 year ago

Would the addition of this note apply not only to -w but also to --jsonrpcsecretfile, --directoryfile and other commands that read files?

pljones commented 1 year ago

Would the addition of this note apply not to only to -w but also to --jsonrpcsecretfile, --directoryfile and other commands that read files?

Yes, it's fundamental to supplying a file that a program needs to read (and potentially write to). So it applies not only to Jamulus but to any program run from the command line with filename arguments.

gilgongo commented 1 year ago

Would the addition of this note apply not to only to -w but also to --jsonrpcsecretfile, --directoryfile and other commands that read files?

Yes, it's fundamental to supplying a file that a program needs to read (and potentially write to). So it applies not only to Jamulus but to any program run from the command line with filename arguments.

OK so this another case of deciding whether nix users can use nix. Either we add this wording to all those (and any future) command docs, or we assume a certain level of knowledge.

ann0see commented 1 year ago

I mean jsonrpcsecretfile (?) would give an error if the file is not readable - at least that's what I remember.

mcfnord commented 1 year ago

Squash has defeated me. Can I, a git user, use git? No.

pljones commented 1 year ago

Squash has defeated me. Can I, a git user, use git? No.

Where main is set to the appropriate value (main for jamulussoftware/jamulus and next-release for jamulussoftware/jamuluswebsite):

git fetch upstream ${main} ;# pull upstream project
git push origin ${main} ;# make sure we're up to date ... if this doesn't work, manual fix ups may be needed...
git status ;# manually check you're on the right branch and have commited
git rebase -i ${main} ;# interactive rebase - easiest way to squash cleanly

Then, in your chosen editor, you'll see your branch commits labelled pick. Leave the first one, change the others to s. (!}s/^pick/s in vi).

ann0see commented 1 year ago

@gilgongo @mcfnord would you be ok with me squashing it? I'd still prefer if the original author did it.

Git is confusing - I agree. https://github.com/jamulussoftware/jamulus/blob/main/CONTRIBUTING.md#submitting-code-and-getting-started is a link to a short paragraph for contributors helping to get more information about git. There are some links which might be good? I believe something like https://medium.com/swlh/squash-and-rebase-git-basics-5cb1be1e0dac for the concept and https://medium.com/@slamflipstrom/a-beginners-guide-to-squashing-commits-with-git-rebase-8185cf6e62ec would help me.

mcfnord commented 1 year ago

I tried squashing but failed. I don't care enough about this PR to try again, but do recommend publishing it.

gilgongo commented 1 year ago

@gilgongo @mcfnord would you be ok with me squashing it? I'd still prefer if the original author did it.

There is still the question of whether, if we explain the issue of permissions in one place, we need to also explain it in all other places where someone might also not understand it. https://github.com/jamulussoftware/jamuluswebsite/pull/927#issuecomment-1570973777

And if we explain that when a program needs read/write access to a file it also needs to be readable/writable by that program's user, then what else do we need to explain?

mcfnord commented 1 year ago

Philosophically, improving one thing doesn't require parallel improvements elsewhere. And philosophically, while we might say more about something, saying half often resolves 90% of confusion, leaving 10% for the other half.

I look for failure patterns. If I haven't witnessed evidence of failure, it's not a priority to me to improve guidance. Whereas with the welcome file, I experienced a cascade of failures without error guidance due to the access restrictions and the unique text-or-path architecture of -w.

ann0see commented 1 year ago

I think I'll leave the decision, further actions and merge up to @gilgongo as I feel I'd block here.

We squash merge next-release to release so I believe we can merge this without a squash.

ann0see commented 1 year ago

This means: I'll unsubscribe from this PR. If you need me, please ping me.