Clojurians Log v2

Awesome to see so much activity here lately 🥳. Especially @lee and @corasaurus-hex teaming up to tackle all these issues 💪👏. I've been working a bit on the SQLite->Postgres migration and could use some feedback here before going further. I just opened a WIP PR for this: https://github.com/cljdoc/cljdoc/pull/587

I shooooould be able to look at this later

very excited to be using postgres

RèWRïTÉ -> rewrite-clj

love it

Gonna go live with my changes to libraries search (which includes popularity ranking). Did lotsa testing, but still, a rather large change. 🤞

Looking sweet to me! Lemme know if you notice any library search issues.

I'd really like to understand why we're using the IStorage protocol

oh, I suppose it's in the ns docstring

sometimes it feels like apps, as they develop over time, are less designed as much as evolved. like the recurrent laryngeal nerve in giraffe's there's sometimes things so puzzling that only make sense in light of the app's past

Before my time, but the IStorage rationale seems to make enough sense to me.

But yeah, when I work on a thing, I notice another several things that sometimes confuse me.

probably handy, even, for the switch we're making

hmmm

❯ npm run dev

> dev
> parcel watch resources/public/cljdoc.html --dist-dir resources-compiled/public/out

🚨 Build failed.

Error: ENOENT: no such file or directory, stat '/Users/cora/Code/cljdoc/resources/public/opensearch.xml'

  Error: ENOENT: no such file or directory, stat '/Users/cora/Code/cljdoc/resources/public/opensearch.xml'

Try deleting your .parcel-cache

kk

aaaand there we go

thanks!

I really haven't used parcel much

yet another mysterious js build system

🙃

it's nice that it kind of all Just Works™

It might be great and powerful, but parcel.js is also very mysterious to me. Like figuring out how to have it ignore opensearch.xml was… well.. not obvious. I waded my way through several git issues where some folks did not seem to see the point of that kind of thing. Fair enough, I guess. Finally found a “resolver” that did the trick!

webpack has resolvers, too

which are kind of cool in that you can have it perform things for certain kinds of imports. like when importing an image it can make sure that it makes optimized versions if they're not on disk

but also strange magick

JavaScript world = overly-confusing to my wee brain. simple_smile

I feel the same about a lot of worlds. at least clojure doesn't change all that quickly

It is a salve for my addled mind!

I'm so glad malli exists now. it gives me the sort of structural descriptions/types/validations I've wanted in clojure for so long. it definitely fills a big gap for me in the language

it was the one part of clojure that really hurt to be without, for me. spec tried to do it but I find it hard to reason about the actual structures that are described via spec

it's particularly painful for large data structures

but in malli it's a snap https://github.com/cljdoc/cljdoc/pull/572/files#diff-2f11335bac8814e71695183f661ca9e4dae45a5d38d4de33921e41d45740c4b3

https://github.com/cljdoc/cljdoc/pull/572/files#diff-bd8c036e4a606b5ba8bd797a5a7ef1659e66595a24b496ba7cbdb93ac9eda742

I'm really excited to use this on more things

Yeah, me too. I’m no spec expert, and I might be totally missing some of its advantages, but I find malli very approachable. And those malli specs are much easier for me to parse at-a-glance.

ugh, pretty sure I'm going to need to use flexsearch and just not do highlighting on matches

for now