I love sqlx and type checked queries but always found it annoying to duplicate a bunch of queries vs using the dynamic query builder (which isn't type checked). Using Option and NULLs seems to improve things though!
// Postgres version
let ids = sqlx::query_as!(
Uuid,
"SELECT id FROM users \
WHERE ($1::timestamptz IS NULL OR updated_at < $1) \
AND ($2::timestamptz IS NULL OR updated_at > $2) \
AND ($3::boolean IS NULL OR is_guest = $3)",
updated_before_option,
updated_after_option,
is_guest_option,
)
.fetch_all(&pool)
.await;
In a couple cases I've seen my feed reader "catch up" lots of posts at once, where I thought I've missed items or things were backdated and found again. This has mostly happened with folks who seem to use Ghost (Molly White's micro posts and Support Human)
The data flow I'm using is [Site RSS] -> [Miniflux] -> [Reeder]
Getting w2z setup to support lists of strings in the CMS. First time using _hyperscript to add/remove items in a list. Working really well so far!
New version of Reeder now out. I've been using Reeder (now Reeder Classic) since ~2012 (v2.5.4 from what I can tell in old email receipts).
This new version is a redesign that doesn't seem relevant for me at least. The single timeline view isn't how I want to view feeds but I can see how that is what many people might be used to coming from social media.
It looks like Github branch rulesets allow setting a bypass for specific app integrations! This should allow my Github app to avoid making a branch, PR, and auto-merging... which would be nice eventually!
First time giving rulesets a try
I'm exploring using Github Apps for w2z instead of fine-grained personal access tokens (PATs). Replacing PATs every 90 days is a bit tedious. Eventually the app flow should give a better experience.
you should know that if the algorithm chooses you it has nothing to do with the quality or value of your work. And I mean literally nothing. The algorithm is nothing more than a capitalist predator, seeking to consume what it can, monetize it quickly, then toss aside. If you make the algorithm your audience, you get very good at creating for an audience of machines rather than humans. Creating for humans is harder, it may get you ignored by the algorithm, but your work will be better for it, and it will find an audience in time.
- Brian Koberlei via P&B
Saw Daybreak via Simon Clark and would love to give it a go at some point! The collaborative nature of the game and the reality of climate action seems really appealing.
Don't really need any new espresso cups, but maybe if I break any I should give these a try.
Based on a Github issue I was able to get a collection with Zola taxonomies writing the proper TOML-frontmatter format.
collections:
- ...
fields:
- label: "Taxonomies"
name: "taxonomies"
widget: "object"
fields:
- { label: "Tags", name: "tags", widget: "list", allow_add: true}
This is written out into the frontmatter as
+++
[taxonomies]
tags = ["tag"]
+++
Anyone using DecapCMS? I gave it a try a while ago and looks like there is now Nested Collections which should now match how I have my site setup.
I guess I'll look forward to PostgreSQL 17, between better upserts and some label improvements.
The MERGE seems to take more code than I'd like, I wish ON CONFLICT didn't bloat and could have an option for ALWAYS RETURNING that would return the row even if not modified. I'd deal with the bloat if it were simple code and always returned the row.
Cool looking Rust crate registry, Cratery. OAuth login, S3 storage. Guess could setup Litestream for DB storage as well.
Last spring we found it slightly too cold to go camping when we otherwise had the chance. Maybe time to pick up a warmer sleeping bag.
I was confused when I didn't see a Vegan label on a Silk carton. I'm not really less confused after reading the FAQ.
I was able to get tracecontext passthrough to OTel working! I was missing the opentelemetry_sdk::trace::TracerProvider which pulls in the infomation to help build the OTel data that gets sent to collectors.
A new sqlparser release pushed out my changes for supporting declare-schema! I released 0.0.1 but haven't gone through and updated my services yet.
In a month or so of using the dev version of declare-schema in my services I've had a good time! Since I wrote it with my mental model for services in mind, it works as I want it to! The library itself would probably wind up with significant data loss for others.
‹ Previous Next ›