Post

How to get my org on the ERSC hype train 👀

am i allowed to pronounce it ersk

how large can repos go? what tricks are used to manage disk size and performance - sparse checkouts? VFS? how are large files handled? how do they work with the git compat layer? "I've used GitHub / Gerrit / Phabricator, how is Changeset different?"

deviations from git and the reasoning behind them

🫡 one important thing we’re doing here is that we have git compatibility, so you can Just Use Git up until you need the truly good stuff

YES and that’s great but it does feel already relatively established in conversations I have about jj and the bricks for the next steps into switching into jj only would be really goos to see, especially from *you* specifically

Is the PR the correct unit anymore? What are ERSC's hot takes on code review?

will blog for sure but comfortable saying PRs are bad and you want per-commit review not per-branch.

I guess this is getting into being a feature question, but how about interdiff? Because I know I do want commit review + very much want to stop having to mentally chronologically sort comments that are sorted by file + line number, split across multiple tabs.

How does CI work in ERSC? What is different? What we can do better, e.g. what can we get rid of in this context?

My number one question: what is ERSC doing differently from everyone else? What will make them stand out in our thoughts amongst all the GH alternatives coming together right now

Appreciate it! This is the number one question for sure. Don’t want to spoil anything just yet so I’ll write that down and be vague in this reply 😅

pretty simply what makes ersc different to the end user than the hubs and labs of the world

The wheels you’ve chosen to re-invent VS the boring technology you’ve chosen to adopt

are you using ersc to build ersc yet?

for ‘enthusiast engineers’, there’s a lot of resentment towards github and “what if a vcs forge was good” is a pretty straightforward pitch. but to execs who actually buy software, github is comfortable and familiar, with lots of features and audit logs and permissions and stuff …

so i’d love to see ersc publish some material along the lines of “here’s why we think it’s worth it to rewrite all of your ci from github actions yaml and move all of your integrations over to our new platform”

Yeah this is clutch. One thing we want to make sure we do is like, support those integrations, so you’re not actually losing out on stuff

Why build ESRC now?

And why do we keep getting your acronym wrong reflexively? (And will you consider changing it)

East Source River Control

jj is cool but what I want to know is what are the plans for ersc *around* software development (issue tracking, planning, discussions etc.) and how those can be integrated in. I would also like to know how jj would improve those things. jj makes stacked prs feasible (insert a new commit here) 1/2

and what I would like to know is what else the technology enables when we have a new backing storage and how this allows ersc to distinguish itself from systems like gerrit etc. 2/2

Cool cool, appreciate it! Got some good stuff coming there for sure…

When you push a commit, and you make a change to the jj change id and re-push (is that the right lingo for ersc??) is there a way to comment on or see what changed? On GH your stuck re-reviewing everything. I just want to re-review the diff to the last commit.

a tailscale-"how nat traversal works"-style 'why shouldn't i roll something like this myself?' rundown of like actual non-obvious challenges that present themselves when you start being a source control (& accessories) host

Plumbing. I know its a nice new house, but I wanna hear about the plumbing.

The plumbing is proprietary but that doesn’t mean we won’t want to show off some floor plans, for sure

why choose proprietary? (potentially for the blog, not here)

📯

I didn’t even realize that was an emoji

Can I build another wall of little green squares?

i know you're joking but, we are building a per-company product rather than a global social network and so the answer is no :(

i know you have communicated this bit loud and clear many times, but it might be worth a post communicating what you're doing and NOT doing so that it's more settled in people's minds.

What about companies that *have* to own their infra and for example, run a local gitlab deployment on their own servers? Also, given jj’s (current) lack of support for something like git lfs and stuff like Lore vcs, have you considered anything in that direction?

does ersc support woke 3 or is it still stuck on woke 2.1

The amount of trust that will be required by ERSC users. Is there credible exit or data backups, are issues tied to the platform, etc. Looking forward to the post!