Clojurians Log v2

> I already had a ton of useful queries in babashka, so I wrote a script that sorta maps records across databases. You just supply relationships like "to get the ID to a foo record in the apples database, get this jsonb property from the bar column in the bananas database". This mechanism of defining "resolvers" and automatic "connecting the dots" .. sounds quite like a job for #pathom to me. (and I just realized it can be used with babashka, nice! https://pathom3.wsscode.com/docs/tutorials/babashka/) In fact we have similar situation in our job, sometimes having to jump between 2 different DBs (one is postgres, another one mssql) and bb + 2 sql pods already helped us to create a few bb scripts which need to interact on both dbs at once. If I think about it, bb, nice tuis, sql pods and perhaps even few pathom resolvers .. that can be very nice data drilling thing. btw @jhane by any chance, did you hit any issue with rendering emojis in that tui?

thanks for all your time today, I'm so happy to have it working! The script I wanted it for is FLYING now

@jhane What was actually your use case for concurrent requests: querying different databases concurrently?

at work we have a very microservice-ey architecture, so related data is distributed across a whole bunch of databases. So when you get a bug report/ticket with a client ID, or see a message ID in a log, it can be a big hassle to get the whole context. You have to go into your database GUI and write a whole bunch of queries across multiple databases just to get basic info like "what user created this record?". Sometimes it's hard to even tell what it's an ID for. It just shows up in an error log among a bunch of JSON. I already had a ton of useful queries in babashka, so I wrote a script that sorta maps records across databases. You just supply relationships like "to get the ID to a foo record in the apples database, get this jsonb property from the bar column in the bananas database". When you supply it an ID • looks at all the records you supplied in the relationship map, and checks if any of them contain a record with that ID (only searching indexed properties of course, usually primary keys) • once it determines what type of record it is, it walks through all relations, and builds a lil graph of data. So instead of fiddling around in datagrip for 15 minutes, you just slap an ID in the script and have everything you need. The concurrency was a huge improvement because now I can actually reach out to all databases at once (even multiple tables within the same database) instead of waiting for em one by one.

right, this makes a ton of sense. I asked because using the same connection (object) against the same database usually doesn't work when you bang at the db with multipl threads, so wanted to double check if this was actually for querying multiple dbs :)

oh that's good to know! I hadn't really checked if multiple threads against the same DB was working or not. But the multiple DBs is def the most important part

multipls conns against the same db should work, but not the same conn

or at least, postgres should handle it. sqlite doesn't I think

oooooh ok that makes sense, that's excellent

also @timok I made a TUI for this script with charm, and it's really slick. Being able to make TUIs with bb is just the best

Let me try Emacs 30 with clojure-ts-mode

what do I have to change to use it

> By default, clojure-ts-mode assumes command over all buffers and file extensions previously associated with clojure-mode (and derived major modes like clojurescript-mode). I guess nothing then :)

hmm. when using clojure-ts-mode and pressing backspace here (+ 1 2 3)| it actually removes the paren instead of leaving the balanced parens and just removing 3. I guess I do need some config?

I had to add this:

;; clojure-ts-mode: use instead of clojure-mode
(add-to-list 'major-mode-remap-alist '(clojure-mode . clojure-ts-mode))
(add-to-list 'major-mode-remap-alist '(clojurescript-mode . clojure-ts-clojurescript-mode))
(add-to-list 'major-mode-remap-alist '(clojurec-mode . clojure-ts-clojurec-mode))

;; clojure-ts-mode: enable smartparens + LSP (same as clojure-mode)
(add-hook 'clojure-ts-mode-hook (lambda () (run-hooks 'prelude-lisp-coding-hook)))
(add-hook 'clojure-ts-mode-hook 'lsp)

Hmm, I remember we had some auto remappings in clojure-ts-mode.