ngrilly

42 ms·

571.
by ngrilly 6y ago
If you can lose the last few seconds then yes that's fine. But for most applications I've been working on, we didn't have that flexibility (committed means durable). I don't see any operational complexity with Litestream
572.
by ngrilly 6y ago
You don’t want to maintain your own database server, even managed by GCP, but with SQLite you have to maintain state on GCP Persistent Disks and backups to S3 using Litestream. Why do you think this is easier?
573.
by ngrilly 6y ago
I don’t like the click bait title, but the conclusion is interesting.
574.
by ngrilly 6y ago
That’s a very nice library. BEAM + Elixir + OTP supervision trees + Obama for background jobs + PostgreSQL for persistent storage seems a killer combination. And like you said you can scale horizontally naturally with the Erlang plateform w
575.
by ngrilly 6y ago
How do you keep track of background jobs? Is the queue persisted on disk somewhere? If it's not and a background worker crashes for some reason, then is the job lost?
576.
by ngrilly 6y ago
Not in France. Everything is publicly accessible on Legifrance.gouv.fr.
577.
by ngrilly 6y ago
It’s not just an issue for hobbyist programmers. That’s also an issue for most startups and small companies. You just can’t buy all the documents you need when you’re designing hardware. That stuff is insane. And it’s not just about the cos
578.
by ngrilly 6y ago
Your padding + deduplication solution is nice. > Edit: If you used the ID of the last item instead of an offset, then you could get errors if your last item is in fact the one that was deleted. If the items are sorted by ID, then we use
579.
by ngrilly 6y ago
> just provide the ability to get an arbitrarily large set atomically It's difficult to do in most HTTP deployments where we try to limit HTTP response duration. > if the deleted item is the one your cursor is sitting on, then yo
580.
by ngrilly 6y ago
Amazing feedback. Very much appreciated.
581.
by ngrilly 6y ago
> rust is far easier That's an unusual claim :) Could you elaborate? How is the learning curve, especially for fresh graduates who have been trained only on languages like C, Java and Python? > The fact that you can just drop a f
582.
by ngrilly 6y ago
That solves the problem of items added concurrently (by deduplicating them) but that doesn't solve the problem of removed items.
583.
by ngrilly 6y ago
But then you need to provide some kind of query ID in the the HTTP request? And if you many concurrent clients, that can be expensive in terms of RAM? If we don't care about added items (they can be deduped client-side), and only care
584.
by ngrilly 6y ago
> If items get returned twice while iterating, then I know something got added. Fair point. I agree, that's a nice side effect. But what if you're looking at page 1, and an item is removed from that page? Then you'll never
585.
by ngrilly 6y ago
How Go disappointed you beyond the case in the linked article? Is there a few cons in using Rust compared to Go? In particular, what about language complexity/learning curve and compilation times?
586.
by ngrilly 6y ago
Do not use a page argument for pagination. If you have another process/client concurrently adding/removing items, then some items will be returned twice, and others will never be returned. It is better to use, for example, the ID
587.
by ngrilly 6y ago
There's a way around it which is actively worked on apparently: https://www.sqlite.org/src/doc/begin-concurrent/doc/begin_co...
588.
by ngrilly 6y ago
Also used to PostgreSQL, and was considering SQLite for a web service. The biggest issue is how to run database migrations without downtime (while still accepting writes). There is no CREATE INDEX CONCURRENTLY. No ALTER TABLE DROP COLUMN. I
589.
by ngrilly 6y ago
I suggest to have a look at https://github.com/postgrespro/rum if you haven’t yet. It solves the issue of slow ranking in PostgreSQL FTS.
590.
by ngrilly 6y ago
Too bad it’s only for Enterprise and not Teams.
591.
by ngrilly 6y ago
Does Zig plan to support closures?
592.
by ngrilly 6y ago
I have the exact same issue with Signal, which is sad, because it's perfect except for the lack automatic and encrypted cloud backup.
593.
by ngrilly 6y ago
Same. I used to have a Macbook Air 13" and a friend convinced me to switch a Macbook Pro 15" for the screen size. Now I couldn't see myself coming back to 13".
594.
by ngrilly 6y ago
Yes, but it's only local, not cloud, as far as I understand.
595.
by ngrilly 6y ago
I ran exactly into this problem with linear.app (which uses passwordless login with "magic" links) and Outlook / Office 365. Only solution has been to ask linear.app to disable the links for our account and keep only the one-
596.
by ngrilly 6y ago
Try linear.app and it will clarify why Jira sucks :)
597.
by ngrilly 6y ago
I’d be really curious to know why your fact-based comment is downvoted.
598.
by ngrilly 6y ago
Thanks for that detailed explanation. The fact that all main functions live in a single process probably helps a lot with CPU and RAM usage, compared to k8s where etcd, API server, scheduler and controller manager live in independent proces
599.
by ngrilly 6y ago
The big problem for me with Docker Compose is the lack of zero downtime / seamless deployments. First it stops the containers running the old version, then it starts new containers running new versions, instead of doing the reverse.
600.
by ngrilly 6y ago
Thanks. Seems like it’s worth looking into it then. Still curious about the reasons why it’s more efficient.

More ›