Clippy is returning just under 300 warnings/errors for this project.
I would totally accept such changes, at least in pijul and pijul-macros. Lilbpijul is trickier, and still needs some work, before we can run clippy on it.
It doesn’t compile anymore.
I agree with using
s.len() >= 1 + BLAKE3_BYTES loses the meaning of that comparison, and makes the code less readable (the
1 + constant is almost certainly optimised away by the compiler).
Edit: the other changes are good.
i just accidentally pushed this to main. i’m a little terrified that was possible!
It doesn’t compile anymore.
the latest version isn’t compiling for me without these changes
i’m no longer able to push to this discussion (insufficient permissions)
Indeed, sorry about that. I got scared when you said you could push to main, and now I’m confused: can you really push to main and see your changes there?
i did not push to main.
in actual fact it must have been a completely legitimately merged change at the top of that list ;)
i take it you’ve locked me out entirely as a precaution?
still just scratching the surface ;)
is there an equivalent of a
git merge --squash workflow to squash a channel into a single change?
if that’s not desirable, how would one handle this workflow using pijul?
ie. the changes are individually small, and it’s handy to review them separately. They also represent chipping away gradually at a larger problem. But in terms of merging them, they are closely related and represent a single ‘idea’
There isn’t really any convenient way to squash yet, hopefully we’ll get one soon. There are many possible ways to implement this, the most naive one being unrecording and recording again. Actually, that shouldn’t be very hard to do. One only needs to:
We could do this in
pijul record, or as a subcommand of
what would be your preference? I don’t want to mess up your nice neat repo history
The problem is not so much the repo history as the dependencies. I actually have little experience with Pijul (nobody has), so I don’t know the consequences of having lots of dependencies, rather than a large one. I’ll try to implement a squash today.
I know it looks like this is stalled, but I am reviewing these changes. I hope to be able to push them to main by tomorrow.
Done, thanks again! Good job. I’ve had to solve a few conflicts, but it’s done.
having issues pushing changes to this discussion. is this because it’s ‘closed’?
This is strange, nothing changed. You should be able to push to closed discussions. This could be related to a new version of Pijul. I did publish (and yank) a version that had SSH issues, maybe you’re on that one?
Hi pmeunier, i’m still seeing this issue with alpha.31. Any time i try to push or pull using SSH I am asked for a password (which I don’t have!)
Is there a way I can get some more debug output? Are there any restrictions on the types of SSH keys (especially if that’s changed recently!)?
i managed to be able to push and pull using SSH. I had to delete and re-add my public ssh key (the same one). Seems to work now
when pushing the changes to this channel a whole bunch of changes were included.
if i removed these changes and saved the document it would just regenerate. I was stuck in a loop where the only solutions were to a. delete everything and not push any changes b. push everything, including all the changes that aren’t mine
anything else was just an infinite loop
buried in this list of changes are two new changes
Am i perhaps doing this wrong? what i did was
Alright, there’s been a datacentre fire and loads of changes since these patches. I believe these changes are fixed in the current version.
In an attempt to triage the remaining bugs before the beta release, I’ll close this discussion, but if you are still interested in helping with these lints, feel free to open a new one.