0.12.2 was released Jul 31, 2019
Would be cool to make 1.0 release Jul 31, 2020 exactly 1 year after. Or at least have an RC release at that memorial date.
Hello
Are there any news?
2 Likes
I see… so something is happening somewhere, but no updates.
Hi! Is there any status update about Pijul 1.0 somewhere? Or, to ask more directly: Is there any upper bound for a release date? 
I’m eager to try out Pijul but would like to wait until the 1.0 release, so that I don’t have to revise the knowledge I’ve acquired about Pijul.
1 Like
Good question. My life has become a little bit chaotic lately, but I’m hopefuly to finalise Pijul 1.0 in the next week or so. The main challenge is to make sure that the foreseeable future improvements won’t need a major change in the data representation.
In particular, Pijul 1.0 uses very little disk space (compared to 0.12), and patches can be downloaded in a very efficient way in many cases, allowing large binary files to be transferred more efficiently. But this also means that the network protocol is more complicated, since there are now two ways to transfer a patch, a “preliminary” one containing a small prefix of a patch, and a “full” one containing the whole thing.
12 Likes
How up to date is the documentation at https://pijul.org/manual/? If it isn’t up to date, will it be updated in conjunction with the release of 1.0; or will there be a period between the release of 1.0 and when users get information on how to user Pijul?
Excited about the release!
1 Like
TL;DR: You probably mean well, but I don’t think you are actually helping with this kind of messages.
Even if, to some extend, I can easily agree that I would have liked to see the development of pijul take another direction, one with less uncertainty and more transparency, I don’t think copy/pasting URL from a blogpost without any additional word is a meaningful way to improve the situation.
There is I think a reason why, despite the number of folks interested by pijul, only @pmeunier is the only one to make changes to the core algorithms of pijul. Even when the development was made in the open, he basically had to fix all bugs in libpijul.
At the moment (and maybe despite what was wrongly claimed, idk), pijul is not “ready.” Its main author has decided that the best way to improve the state of things in a manner that is healthy for him is to do what he does in a private set-up. I can understand why, and since I cannot be of any meaningful help wrt. what he is currently trying to achieve (i.e., the core algorithms of libpijul), I don’t see why I should ask him to change its modus operandi.
Now, if you believe you can actually try and help him better, I guess the best path forward is to reach to him with a more precise proposal, rather than just throwing a passive/aggressive message like that.
9 Likes
Thanks @lthms. Just to give a little complement, I feel that many people don’t really realise what has to be done here. The main things I’ve been doing recently were 99% hard-core algorithm design and debugging, and about 1% other stuff. In the few years where I was trying to maintain everything at the same time (Nest, Thrussh, Sanakirja, Pijul) in the open, I received invaluable help and contributions on the interface (which is why it’s now down to 1%), but very little on the algorithms, for reasons that are easy to understand.
About a year ago, when I understood what had to be done to complete this project, I decided to stop using Pijul for itself (also, I was changing all the formats all the time), and moved back to Darcs, which unfortunately doesn’t have a very useful public hosting service, but is still easier and faster to use than Git.
Now, I totally understand that people are eager to use Pijul and hack on it, but it isn’t completely ready yet, and I don’t think it could have been ready even with all the cool contributions that could have happened, had it been more public.
3 Likes
hub.darcs.net is a pretty neat interface if all you want is to provide a way for people to have a look at the code. I haven’t tried to use it for more (e.g., organizing collaborative work, but this is not what you are maybe looking for).
Seems to me they were just sharing an interesting and related essay on open source software development. You really read way too much into a post with no comment and not enough into the essay that was actually posted. I know you mean well too, but it is in fact you that came off as passive/aggressive to me.
Anyway, I’m excited for Pijul as well. I’ve been refreshing the blog constantly since pmeunier’s post last week.
I don’t think my post was passive-aggressive, but I can understand it can feel aggressive (which is not better by any means, let me be clear); I am sorry about that. I could have taken more time to word it in a more welcoming way indeed.
For a different perspective on project management, I recommend “The Mythical Man-Month”. It was very interesting.
1 Like
And remains relevant, because it’s wisdom!
By the way, the main things that remain to be done before release are:
- make Git import incremental, at the moment it’s a one-off thing that doesn’t properly save its intermediate states.
- rebalance a few things between Pijul and Libpijul. Florent is taking care of this one.
- update the Nest for the new formats and new protocol.
- write documentation.
In other words, we’re really close to a release.
6 Likes
Good to hear! Out of curiosity, were you able to keep our previous work on the CLI-side of pijul (e.g., writing a patch description from an editor, a hook, etc.)?
Sort of. This is essentially a rewrite from scratch, but it does reuse some of the ideas. The CLI-interactive thing has disappeared completely, and every user choice is now done from an editor (including records, pushes and pulls).
1 Like
That’s an interesting take; I guess I will have to try it to decide whether or not I like the idea! At the very least, I am eager to give it a try!
If the text format is simple, one could easily use as ”editor“ an interactive filter which interfaces with e.g. the newly discovered odour language of skunks. So that’s quite universal.
Certainly the lang wasn’t discovered in the wild, only in bewildering fantasy.
I’m really eager to try pijul 1.0 and, when I have time, try to find out whether sourcehut integration is feasible. The latter would require at least an email integration of pijul (full featured textual diffs to edit and discuss), as sourcehut is centered around and federated by email, and its mission is to make vcs-email workflows accesible and convenient, also via web interfaces.
1 Like
I didn’t understand what you meant by this, but I looked up the site @jorsn mentioned, and reading there, I found https://git-rebase.io/ which explains how git does things interactively through the editor when using certain parameters. I didn’t know about these features. Is the new pijul like that?
1 Like