Fork me on GitHub
#datahike
<
2023-09-04
>
zimablue17:09:31

hey datahike people, been a while, I was on a couple of video calls with you a while ago, depending on (a manually patched to work on cljs) version of datahike for my full-time project

zimablue17:09:42

how are things

timo17:09:03

Hi @zimablue. Things are good, working on some CI right now. How is your fork working out?

zimablue17:09:32

well it was only a matter of necessity, although it was my first major cljs undertaking so mostly learning. Do you guys support cljs natively yet? I remember it was a goal last time I was on here

zimablue17:09:48

would love to throw my patches away and update

zimablue17:09:30

if not, I'm getting some baffling problems halfway through some big internal changes to my app, so it's probably a good time for me to try and reapply said patches onto your newest version, since I'm confused in the base layers anyway

timo17:09:53

Unfortunately work on CLJS stalled when a major supporting company dissolved.

zimablue17:09:02

that's unforunate

zimablue17:09:04

unfortunate

zimablue17:09:07

I'm on a very old version at this point, pre-tuples-support, so at this level of confusion maybe I'll try to reapply the changes on your latest version and in doing so debug whatever's going on with my system

zimablue17:09:35

how's business generally, iirc you were mostly two guys working as consultancy but also in academia (?)

zimablue17:09:01

the work I'm depending on is all github pinned versions from "groundedsage" who I think is "whilo"

timo17:09:47

When I recall correctly there was a working version of Datahike for node but I don't know if it was inherited into the codebase.

timo17:09:26

yeah groundedsage was the guy working on the cljs-stuff. He is not whilo and maybe he answers you if you try him.

zimablue17:09:01

that probably had to be post-when-i-forked I think, because I had to patch down to the serialization layer, that was the hardest part, because it used clojure introspection in several places, which doesn't have an equivalent in clojurescript, you need to manually construct a symbol table because all the names get munged

zimablue17:09:09

ah sorry it's been a long time

zimablue17:09:26

unless the node version that worked was in-memory-only

zimablue17:09:47

I'm forked at 90891e21cdfde6386861602d9f8231a3307da77d, thu mar 18 on the datahike main repo

timo17:09:14

haha, good luck 🙂

timo17:09:43

konserve at least has cljs-tests on it

timo17:09:00

persistent-sorted-set is cljs-compatible for sure

zimablue17:09:19

no fear I am two years stronger at clojurescript to match the two years of divergence, I also forked konserve and hitchiker-tree at the time

👍 4
zimablue17:09:47

I'm going to take a look now, if @grounded_sage is around to advise or help if I get confused would be great

timo17:09:48

superv.async seems cljs-compatible as well

zimablue17:09:23

I think that has been since forever, since I never forked it, macros are all written in clojure anyway right so unless you dip into clj-only async things you'd think it would work

timo17:09:26

It would be absolutely awesome to have this working at some point. I am giving you a power up 🚀 mario-star

zimablue18:09:06

aha my error went away when I required a future version of datalog-parser, very very strange but if I can get away with updating just that and not fixing datahike-latest for cljs I'll have to stop there, too busy unfortunately

zimablue18:09:19

ah well at least know there's 1 person who'd love cljs support, also will happily help/discuss if you get back to it

grounded_sage22:09:25

Cljs support is definitely still on the roadmap. I am working on other things these days. What do you need cljs support for @zimablue?